SIENA European Roadmap on Grid and Cloud Standards for e-Science and Beyond
SIENA European Roadmap on Grid and Cloud Standards for e ...
SIENA European Roadmap on Grid and Cloud Standards for e ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
www.sienainitiative.eu | info@sienainitiative.eu<br />
<str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g><br />
<strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards<br />
<strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
Use Cases & Positi<strong>on</strong> Papers<br />
<str<strong>on</strong>g>SIENA</str<strong>on</strong>g> is a Specific Support Acti<strong>on</strong> funded by the GÉANT & e-Infrastructure Unit, DG In<strong>for</strong>mati<strong>on</strong> Society & Media, <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong>
Index<br />
Executive Summary................................................................................................................................................3<br />
Introducti<strong>on</strong>: Forces Driving Change............................................................................................................. 5<br />
Future <str<strong>on</strong>g>European</str<strong>on</strong>g> e-Infrastructure...................................................................................................................6<br />
e-Infrastructure Requirements.........................................................................................................................8<br />
e-Infrastructure Technology.............................................................................................................................9<br />
Enabling St<strong>and</strong>ards.............................................................................................................................................. 10<br />
Internati<strong>on</strong>al Co-ordinati<strong>on</strong>.............................................................................................................................12<br />
<strong>Cloud</strong>s St<strong>and</strong>ards Coordinati<strong>on</strong>.....................................................................................................................13<br />
C<strong>on</strong>clusi<strong>on</strong>s/Recommendati<strong>on</strong>s/Future Directi<strong>on</strong>s........................................................................... 14<br />
Target Audience....................................................................................................................................................15<br />
Timeline....................................................................................................................................................................15<br />
Scope.........................................................................................................................................................................15<br />
<str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> Editorial Board (REB) Member List........................................................................................... 16<br />
<str<strong>on</strong>g>SIENA</str<strong>on</strong>g> Project Descripti<strong>on</strong> – www.sienainitiative.eu............................................................................ 18<br />
<strong>Cloud</strong>scape III Use Cases & Positi<strong>on</strong> Papers............................................................................................. 19<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
1
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
2<br />
Disclaimer<br />
The views expressed in this roadmap are those of the authors <strong>and</strong> do not necessarily reflect<br />
the official <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong>’s view <strong>on</strong> the subject.
<str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong><br />
<strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
Executive Summary<br />
The future <str<strong>on</strong>g>European</str<strong>on</strong>g> electr<strong>on</strong>ic infrastructure <strong>for</strong> research (e-infrastructure) needs to<br />
integrate federated <strong>and</strong> virtualised technologies based <strong>on</strong> geographically distributed<br />
in<strong>for</strong>mati<strong>on</strong> <strong>and</strong> communicati<strong>on</strong>s technology (ICT) resources in a secure <strong>and</strong> interoperable<br />
way. Such ICT resources will be provided by both the public sector <strong>and</strong> commercial vendors<br />
<strong>and</strong> be dynamically <strong>and</strong> flexibly accessed <strong>on</strong> dem<strong>and</strong> to provide a set of comm<strong>on</strong> services<br />
<strong>for</strong> the communities they serve.<br />
A driving <strong>for</strong>ce <strong>for</strong> e-infrastructures in Europe is data intensive science exemplified in<br />
Europe by existing research projects at nati<strong>on</strong>al <strong>and</strong> <str<strong>on</strong>g>European</str<strong>on</strong>g> levels 1 , <strong>and</strong> future projects<br />
such as those described in the <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> of the <str<strong>on</strong>g>European</str<strong>on</strong>g> Strategy Forum <strong>on</strong> Research<br />
Infrastructures, comm<strong>on</strong>ly referred to as the ESFRI projects 2 . Our focus is to identify<br />
the core comm<strong>on</strong> requirements relating to the provisi<strong>on</strong> of e-infrastructure that the<br />
communities have rather than the specific functi<strong>on</strong>ality used by particular communities. A<br />
high-level descripti<strong>on</strong> of these requirements, <strong>and</strong> especially those that are comm<strong>on</strong> to all<br />
or most projects, is c<strong>on</strong>tained in the report of the <str<strong>on</strong>g>European</str<strong>on</strong>g> E-Infrastructure Forum 3 . Other<br />
relevant documents describing e-infrastructure requirements have been produced by the<br />
e-Infrastructure Reflecti<strong>on</strong> Group (e-IRG) 4 <strong>and</strong> the High Level Expert Group <strong>on</strong> Scientific<br />
Data 5 .<br />
An overarching <strong>and</strong> fundamentally important characteristic of an e-infrastructure is the<br />
interoperability of its comp<strong>on</strong>ent technologies. Failure to achieve interoperability can have<br />
powerful negative c<strong>on</strong>sequences <strong>for</strong> cost <strong>and</strong> efficiency of operati<strong>on</strong>, <strong>and</strong> <strong>for</strong> the research<br />
productivity of user communities of an e-infrastructure. Interoperability is best achieved<br />
through adherence to a set of open st<strong>and</strong>ards <strong>and</strong> agreed principles. Work to establish such<br />
a set of st<strong>and</strong>ards is <strong>on</strong>going <strong>for</strong> the e-infrastructure comp<strong>on</strong>ents, the services, <strong>and</strong> the<br />
metadata, <strong>and</strong> will c<strong>on</strong>tinue <strong>for</strong> the <strong>for</strong>eseeable future. Agreed principles are important to<br />
achieve interoperability as a temporary measure while an agreed set of open st<strong>and</strong>ards is<br />
being developed.<br />
Due to the highly diverse, domain specific requirements of different user communities,<br />
there is a risk of fragmentati<strong>on</strong> in the development of e-infrastructure. The fact that<br />
funding <strong>for</strong> public infrastructure comes primarily from the independent Member States<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
3<br />
1] See, <strong>for</strong> example, the book edited by Hey <strong>and</strong> Gray<br />
research.microsoft.com/en-us/collaborati<strong>on</strong>/fourthparadigm/c<strong>on</strong>tents.aspx<br />
2] ec.europa.eu/research/infrastructures/index_en.cfm?pg=esfri-roadmap<br />
3] https://documents.egi.eu/public/ShowDocument?docid=12<br />
4] www.e-irg.eu/<br />
5] cordis.europa.eu/fp7/ict/e-infrastructure/high-level-group_en.html
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
of the <str<strong>on</strong>g>European</str<strong>on</strong>g> Community also represents a risk <strong>for</strong> fragmentati<strong>on</strong> due to nati<strong>on</strong>al<br />
objectives (e.g. budgetary) possibly being misaligned with <str<strong>on</strong>g>European</str<strong>on</strong>g> level needs. These risks<br />
apply equally to research e-infrastructure <strong>and</strong> to e-government infrastructure, the use of<br />
ICTs in public sector activities. The most important recommendati<strong>on</strong> of this roadmap is to<br />
undertake determined <strong>and</strong> targeted ef<strong>for</strong>ts to discourage fragmentati<strong>on</strong>, <strong>and</strong> to encourage<br />
<strong>and</strong> participate in the development of an adequate set of structures - both organisati<strong>on</strong>al<br />
(e.g. governance, single sign <strong>on</strong>, etc.) <strong>and</strong> technical (e.g. open st<strong>and</strong>ards, security, software,<br />
etc.) to ensure the interoperability of future <str<strong>on</strong>g>European</str<strong>on</strong>g> e-infrastructures <strong>for</strong> research <strong>and</strong><br />
e-government.<br />
4
Introducti<strong>on</strong>: Forces Driving Change<br />
Powerful ec<strong>on</strong>omic <strong>and</strong> envir<strong>on</strong>mental <strong>for</strong>ces are driving a major evoluti<strong>on</strong> in the way<br />
in<strong>for</strong>mati<strong>on</strong> <strong>and</strong> communicati<strong>on</strong>s technology (ICT) is provisi<strong>on</strong>ed <strong>for</strong> user communities<br />
in industry <strong>and</strong> the public sector. Ec<strong>on</strong>omies of scale are driving c<strong>on</strong>solidati<strong>on</strong> of IT resources<br />
into a smaller number of ever larger data centers. Data centers with hundreds of thous<strong>and</strong>s<br />
of computati<strong>on</strong>al <strong>and</strong> storage units are no l<strong>on</strong>ger uncomm<strong>on</strong>. C<strong>on</strong>siderati<strong>on</strong>s of the cost<br />
of powering <strong>and</strong> cooling such large c<strong>on</strong>centrati<strong>on</strong>s of electr<strong>on</strong>ic equipment, together with<br />
envir<strong>on</strong>mental c<strong>on</strong>cerns, drive the placing of such data centers in geographic locati<strong>on</strong>s<br />
where power is plentiful <strong>and</strong> inexpensive. As communities become more dependent <strong>on</strong> ICT<br />
resources, the desire to assert their ownership of their data, legal c<strong>on</strong>cerns <strong>on</strong> the locality<br />
of the data, <strong>and</strong> the need <strong>for</strong> geographical redundancy may lead to a diffusi<strong>on</strong> of data<br />
centres. The <strong>for</strong>ces now driving change within ICT are many <strong>and</strong> potentially c<strong>on</strong>tradictory,<br />
leading to different soluti<strong>on</strong>s that optimise the needs of different communities <strong>and</strong> their<br />
use cases.<br />
These <strong>for</strong>ces <strong>and</strong> their c<strong>on</strong>sequences simultaneously enable <strong>and</strong> drive the move towards a<br />
utility model of ICT. The current manifestati<strong>on</strong> of this model is cloud computing through<br />
the commoditisati<strong>on</strong> of the underlying virtualisati<strong>on</strong> technology <strong>and</strong> the globalisati<strong>on</strong> of<br />
service provisi<strong>on</strong>. The dynamic flexibility <strong>and</strong> reduced cost of accessing ICT resources in the<br />
cloud are beginning to overwhelm most other c<strong>on</strong>siderati<strong>on</strong>s <strong>on</strong> provisi<strong>on</strong>ing ICT resources.<br />
Such a fundamental shift poses numerous challenges to user communities. For example<br />
the Integrated Sustainable Pan-<str<strong>on</strong>g>European</str<strong>on</strong>g> Infrastructure <strong>for</strong> Researchers in Europe (EGI-<br />
InSPIRE) project partially funded by the EC is resp<strong>on</strong>ding to the dem<strong>and</strong>s from its user<br />
communities by exploring aspects of cloud computing, notably flexible <strong>and</strong> elastic<br />
provisi<strong>on</strong>ing, within its grid of federated resource providers. This document addresses a<br />
number of these challenges, with a primary focus <strong>on</strong> st<strong>and</strong>ardizati<strong>on</strong> <strong>and</strong> interoperability<br />
of the infrastructures built around the utility model.<br />
Finally, market <strong>for</strong>ces may be working against st<strong>and</strong>ardizati<strong>on</strong> in cloud computing 6 .<br />
The differing requirements of diverse customer communities lead naturally to market<br />
segmentati<strong>on</strong>. These differing requirements also enable vendor differentiati<strong>on</strong> through<br />
the development of different cloud architectures to address different market segments.<br />
Competiti<strong>on</strong> am<strong>on</strong>g vendors can then lead to locking customers into distinct cloud<br />
offerings.<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
5<br />
6] See article “<strong>Cloud</strong> Computing St<strong>and</strong>ards – Not This Year”, by John C<strong>on</strong>sidine, January 2011 at<br />
cloudcomputing.sys-c<strong>on</strong>.com/node/1691805
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
6<br />
Future <str<strong>on</strong>g>European</str<strong>on</strong>g> e-Infrastructure<br />
Electr<strong>on</strong>ic infrastructures at a <str<strong>on</strong>g>European</str<strong>on</strong>g> level are becoming fundamental resources <strong>for</strong><br />
supporting activities across the public sector - primarily e-research, e-government <strong>and</strong><br />
e-health - as society attempts to exploit the data deluge it is facing from the numerous<br />
existing <strong>and</strong> future digital data sources. Obtaining knowledge from this data to benefit<br />
many areas of society requires c<strong>on</strong>vergence at three main levels:<br />
»»<br />
The provisi<strong>on</strong> of a cost-effective, flexible, adaptable <strong>and</strong> reliable e-infrastructure that is<br />
able to support different user groups <strong>and</strong> use cases;<br />
»»<br />
Access to persistently identifiable data sources - open access <strong>for</strong> public data <strong>and</strong><br />
restricted access <strong>for</strong> c<strong>on</strong>fidential data;<br />
»»<br />
The development of appropriate applicati<strong>on</strong>s, algorithms <strong>and</strong> envir<strong>on</strong>ments that use the<br />
e-Infrastructure to extract knowledge from the data sources.<br />
Tackling these issues cuts across many of the areas identified within the Digital Agenda <strong>for</strong><br />
Europe 7 as being critical <strong>for</strong> Europe’s c<strong>on</strong>tinued growth towards a smart society: reducing<br />
the fragmentati<strong>on</strong> of services, improving their interoperability, providing secure access to<br />
valuable data <strong>and</strong> resources, driving innovati<strong>on</strong> <strong>and</strong> development in these services, <strong>and</strong><br />
educating a generati<strong>on</strong> of users <strong>and</strong> developers in the benefit of such technologies.<br />
Europe has already built up significant knowledge <strong>and</strong> momentum in <strong>on</strong>e public sector<br />
area - e-research - after over a decade of investment through the <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong>’s<br />
Framework Programmes <strong>and</strong> nati<strong>on</strong>al funding sources. A successi<strong>on</strong> of projects has resulted<br />
in capacity building across Europe <strong>and</strong> its regi<strong>on</strong>al partners in both grids of high throughput<br />
computing (e.g. EGEE 8 , EGI-InSPIRE 9 ) <strong>and</strong> high per<strong>for</strong>mance computing (e.g. DEISA 10 , PRACE 11<br />
that are linked by the pan-<str<strong>on</strong>g>European</str<strong>on</strong>g> networking infrastructure GÉANT 12 . Al<strong>on</strong>gside the<br />
establishment of this e-infrastructure, innovative scalable middleware 13 has been developed<br />
<strong>and</strong> deployed into operati<strong>on</strong> to meet the needs of researchers across many disciplines<br />
investigating such scientific <strong>and</strong> societal challenges as particle physics, the human genome,<br />
or climate modeling.<br />
The e-research community comprises researchers in such domains as high-energy physics,<br />
astr<strong>on</strong>omy <strong>and</strong> astrophysics, energy research, <strong>and</strong> the earth, material, biological <strong>and</strong> life<br />
sciences. For this e-research community, the next decade will see <str<strong>on</strong>g>European</str<strong>on</strong>g> e-infrastructure<br />
being used as a foundati<strong>on</strong> <strong>for</strong> establishing multi-nati<strong>on</strong>al multi-disciplinary research<br />
infrastructures such as those described in the ESFRI roadmap. Although the maturity of<br />
these individual projects varies, together they have comm<strong>on</strong> needs that if provided<br />
c<strong>on</strong>sistently across the sector will promote many aspects of the Digital Agenda <strong>for</strong> Europe<br />
<strong>and</strong> provide cost-effective return <strong>on</strong> investment.<br />
Central to meeting these different use cases across the public sector is to provide a best<br />
7] ec.europa.eu/in<strong>for</strong>mati<strong>on</strong>_society/digital-agenda/index_en.htm<br />
8] www.eu-egee.org<br />
9] www.egi.eu/projects/egi-inspire/<br />
10] www.deisa.eu/<br />
11] www.prace-project.eu/<br />
12] www.geant.net<br />
13] en.wikipedia.org/wiki/Middleware
of breed e-infrastructure that brings together public <strong>and</strong> commercial providers to deliver<br />
a series of increasingly sophisticated plat<strong>for</strong>ms that are tuned to the particular needs of<br />
these communities. At the heart of this visi<strong>on</strong> is the provisi<strong>on</strong> of a federated, virtualised<br />
e-infrastructure:<br />
»»<br />
Federated: Bringing commercial <strong>and</strong> public sector providers from different countries that<br />
are able to inter-operate with each other - ultimately through the adopti<strong>on</strong> of open<br />
st<strong>and</strong>ards;<br />
»»<br />
Virtualised: Using new <strong>and</strong> emerging software to flexibly partiti<strong>on</strong> these resources <strong>on</strong><br />
dem<strong>and</strong> to meet the needs of various user communities dynamically;<br />
»»<br />
e-infrastructure: Having a set of comm<strong>on</strong> services (e.g. identity management, accounting,<br />
provisi<strong>on</strong>ing, data access, etc) that provides a plat<strong>for</strong>m <strong>for</strong> adopti<strong>on</strong>, portability <strong>and</strong> reuse<br />
across different communities.<br />
The visi<strong>on</strong> presented in this document is by no means guaranteed. The investment that has<br />
been committed by nati<strong>on</strong>al governments <strong>and</strong> the <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong> in GÉANT, EGI <strong>and</strong><br />
PRACE provides vital structural building blocks in the e-infrastructure community, but in<br />
moving from core e-infrastructure to higher-level comp<strong>on</strong>ents the priorities <strong>for</strong> investment<br />
begin to diverge across Europe <strong>and</strong> between communities. The need <strong>for</strong> software to<br />
manage, deploy <strong>and</strong> run in the federated virtualized envir<strong>on</strong>ments remains. To avoid a single<br />
m<strong>on</strong>olithic software deployment across Europe the development <strong>and</strong> implementati<strong>on</strong> of<br />
st<strong>and</strong>ards remains essential if individual sectors are not to fragment into using their own<br />
bespoke <strong>and</strong> n<strong>on</strong>-interoperable software soluti<strong>on</strong>s.<br />
While the Infrastructure as a Service (IaaS) model is at the heart of this visi<strong>on</strong> <strong>for</strong> Europe as<br />
a whole, it will be used as a basis <strong>for</strong> deploying plat<strong>for</strong>ms (Plat<strong>for</strong>ms as a Service - PaaS) <strong>and</strong><br />
software, notably applicati<strong>on</strong> software (Software as a Service - SaaS) that are developed to<br />
meet the needs of particular communities.<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
7
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
8<br />
e-Infrastructure Requirements<br />
Different communities will have different needs from the future <str<strong>on</strong>g>European</str<strong>on</strong>g><br />
e-infrastructure. Our focus is to identify the core comm<strong>on</strong> requirements relating<br />
to the provisi<strong>on</strong> of e-infrastructure that the communities have rather than the specific<br />
functi<strong>on</strong>ality used by particular communities.<br />
»»<br />
Single Sign-On: Inter-domain access to services from different communities dem<strong>and</strong>s<br />
secure, portable, electr<strong>on</strong>ic identity that can be used across different service providers.<br />
The federated identity providers that are being established in Europe present <strong>on</strong>e<br />
possible soluti<strong>on</strong> to this requirement.<br />
»»<br />
Security: Supporting secure <strong>and</strong> dynamic resource (including data, knowledge, <strong>and</strong> services)<br />
sharing <strong>and</strong> collaborati<strong>on</strong>s across instituti<strong>on</strong>al <strong>and</strong> nati<strong>on</strong>al boundaries is an essential part<br />
of achieving the visi<strong>on</strong> of an e-infrastructure. Robust electr<strong>on</strong>ic authenticati<strong>on</strong> capable of<br />
reliably identifying remote users (human beings or software comp<strong>on</strong>ents) with a certain<br />
level of assurance in authenticati<strong>on</strong> strength is an important pre-requisite to facilitate<br />
effective user authorisati<strong>on</strong> <strong>and</strong> fine-grained access c<strong>on</strong>trol to distributed services 14 .<br />
»»<br />
Group Management: Managing individual access to resources across Europe is not feasible<br />
c<strong>on</strong>sidering the number of users <strong>and</strong> resources. Using group based access c<strong>on</strong>trol, such<br />
as the virtual organisati<strong>on</strong> models used in grids, the project model used in HPC <strong>and</strong> the<br />
attributes model used in federated identities, provides a more scalable access c<strong>on</strong>trol<br />
model.<br />
»»<br />
Persistent Data Identifiers: The ability to uniquely identify a data set, <strong>and</strong> from that data<br />
set identify its ownership, access rights, privacy attributes provenance, life-time, stored<br />
locati<strong>on</strong>s, etc. is vital <strong>for</strong> systematic reuse of data across communities.<br />
»»<br />
User Support: Support is needed <strong>for</strong> all types of users (end-users, system administrators,<br />
developers, etc.) across the complete life-cycle of e-infrastructure adopti<strong>on</strong>. This<br />
includes training <strong>on</strong> the deployed technologies, c<strong>on</strong>sultancy <strong>on</strong> their use <strong>and</strong> problem<br />
solving when something goes wr<strong>on</strong>g. This is needed both <strong>for</strong> the core infrastructure <strong>and</strong><br />
any domain specific software that is deployed <strong>on</strong> top of it.<br />
»»<br />
Virtualisati<strong>on</strong>: Communities need to deploy their own services, potentially co-located<br />
with particular data sets, <strong>on</strong> sites across Europe <strong>on</strong> dem<strong>and</strong>. Such activity can then be<br />
decoupled from the deployment activities of other communities.<br />
»»<br />
High Throughput Data Analysis: Such communities need to be able to move large datasets<br />
to where the computing resources are available, <strong>and</strong> to move the results from such<br />
analysis to where l<strong>on</strong>g-term storage capacity is available. In additi<strong>on</strong> to the previous<br />
requirements this requires a high-per<strong>for</strong>mance pan-<str<strong>on</strong>g>European</str<strong>on</strong>g> networking infrastructure<br />
closely coupled to data-centres with large computing <strong>and</strong> storage capabilities as<br />
supported through the EGI-InSPIRE project.<br />
»»<br />
High Per<strong>for</strong>mance Computing: Peta-scale computing resources are essential <strong>for</strong> the small<br />
proporti<strong>on</strong> of researchers solving science’s most dem<strong>and</strong>ing problem through projects<br />
such as PRACE. Efficient access to the small number of peta-scale machines in Europe is<br />
facilitated through high-per<strong>for</strong>mance networking links.<br />
14] See E-infrastructure Security: Levels of Assurance Final Report:<br />
www.jisc.ac.uk/media/documents/programmes/einfrastructure/finalreport.pdf
e-Infrastructure Technology<br />
e<br />
-infrastructure in Europe has reached a producti<strong>on</strong> status over the last decade by<br />
driving innovati<strong>on</strong> in middleware <strong>and</strong> networking technology. This innovati<strong>on</strong> needs to<br />
c<strong>on</strong>tinue over the next decade in areas such as:<br />
» » Virtualisati<strong>on</strong>: High-quality hypervisors that underpin virtualisati<strong>on</strong> in modern datacentres<br />
are becoming comm<strong>on</strong>place. Commercial soluti<strong>on</strong>s provide integrati<strong>on</strong> with<br />
data centre operati<strong>on</strong>s. Open-source soluti<strong>on</strong>s, such as the OpenNebula envir<strong>on</strong>ment,<br />
are being used as powerful tools <strong>for</strong> innovati<strong>on</strong> <strong>and</strong> interoperability in the research<br />
community, <strong>and</strong> as plat<strong>for</strong>ms to implement new st<strong>and</strong>ards in cloud computing.”<br />
» » Networking: Driven by the worldwide growth of the Internet commercial networking<br />
soluti<strong>on</strong>s are available <strong>for</strong> deployment to support public service activities. A focus <strong>on</strong><br />
<strong>on</strong>-dem<strong>and</strong> cross-domain provisi<strong>on</strong>ing of high-speed data transfer links (light paths) with<br />
defined service level agreements is an area which needs c<strong>on</strong>tinuing investment.<br />
» » Software: The software plat<strong>for</strong>ms <strong>and</strong> services necessary to federate the virtualised<br />
resources to provide seamless access <strong>and</strong> to run within the virtualised envir<strong>on</strong>ments<br />
c<strong>on</strong>tinue to need investment. Increasingly, investment needs to take place through<br />
acquisiti<strong>on</strong> of commercially provided software soluti<strong>on</strong>s where they exist <strong>and</strong> allowing<br />
the research community to innovate through open-source software in areas where they<br />
can add unique value bey<strong>on</strong>d the scope of commercial soluti<strong>on</strong>s.<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
9
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
Enabling St<strong>and</strong>ards<br />
St<strong>and</strong>ardisati<strong>on</strong> <strong>and</strong> interoperability are invaluable characteristics to a successful<br />
applicati<strong>on</strong> of distributed computing.<br />
The importance of the need <strong>for</strong> open st<strong>and</strong>ards to support interoperability goals is now<br />
well documented in the e-business world. Of particular relevance to the e-research <strong>and</strong><br />
e-government communities are the statements made in the EICTA Interoperability White<br />
Paper of 2004 15 , the ETSI White Paper No. 3. “Achieving Technical Interoperability” 16 <strong>and</strong><br />
the EC’s <str<strong>on</strong>g>European</str<strong>on</strong>g> Interoperability Strategy (EIS) 17 <strong>and</strong> Interoperability Framework (EIFv2) 18<br />
documents of 2010.<br />
Given a policy of using open st<strong>and</strong>ards to achieve interoperability, the next questi<strong>on</strong> is<br />
which st<strong>and</strong>ards? At present this is not easy to answer. There are many initiatives to define<br />
the optimum set of st<strong>and</strong>ards to support all aspects of cloud computing 19 , but as yet the full<br />
set does not exist. Putting in place the necessary <strong>on</strong>-going procedures <strong>for</strong> tracking emerging<br />
st<strong>and</strong>ards <strong>and</strong> technologies in order to a) set up <strong>and</strong> maintain a central agreed list of open<br />
st<strong>and</strong>ards, <strong>and</strong> b) provide best practice advice to e-infrastructure projects, is a significant<br />
task, <strong>and</strong> will require future investments. In an ef<strong>for</strong>t to align the needs of both the research<br />
<strong>and</strong> e-government communities it may be beneficial to take into c<strong>on</strong>siderati<strong>on</strong> current EC<br />
work <strong>on</strong> Project CAMSS 20 <strong>and</strong> SEMIC.eu 21 .<br />
However the following questi<strong>on</strong>s will persist <strong>for</strong> some time to come:<br />
1. How does <strong>on</strong>e proceed with interoperability if sufficient st<strong>and</strong>ards do not yet exist?<br />
2. What happens if a large market develops <strong>for</strong> commercial offerings without open st<strong>and</strong>ard<br />
specificati<strong>on</strong>s?<br />
3. What if relevant open st<strong>and</strong>ard specificati<strong>on</strong>s exist but are not, or not yet, supported by industry?<br />
10<br />
The EIS/EIF provides the following pragmatic guidance <strong>on</strong> these questi<strong>on</strong>s which should be<br />
equally applicable to the research communities:<br />
»»<br />
Public administrati<strong>on</strong>s may decide to use less open specificati<strong>on</strong>s, if open specificati<strong>on</strong>s<br />
do not exist or do not meet functi<strong>on</strong>al interoperability needs.<br />
»»<br />
In some cases, public administrati<strong>on</strong>s may find that no suitable <strong>for</strong>malised specificati<strong>on</strong> is<br />
available <strong>for</strong> a specific need in a specific area. If new specificati<strong>on</strong>s have to be developed,<br />
15] EICTA Interoperability white paper www.digitaleurope.org/fileadmin/user_upload/document/document1166548285.pdf<br />
. In March 2009 EICTA was rebr<strong>and</strong>ed DIGITALEUROPE.<br />
16] ETSI White Paper No. 3 Achieving Technical Interoperability - the ETSI Approach. By Hans van der Veer (Alcatel-<br />
Lucent), Anth<strong>on</strong>y Wiles (ETSI Secretariat). 3rd editi<strong>on</strong>, April 2008.<br />
www.etsi.org/WebSite/document/whitepapers/IOP%20whitepaper%20Editi<strong>on</strong>%203%20final.pdf<br />
17] COM(2010) 744 final, Annex 1 ec.europa.eu/isa/strategy/doc/annex_i_eis_en.pdf<br />
18] COM(2010) 744 final, Annex 2 ec.europa.eu/isa/strategy/doc/annex_ii_eif_en.pdf<br />
19] See, <strong>for</strong> example <strong>for</strong>ge.grid<strong>for</strong>um.org/sf/go/doc15990<br />
20] ec.europa.eu/isa/workprogramme/doc/detail_descripti<strong>on</strong>_of_acti<strong>on</strong>s.pdf . CAMSS, an initiative of the <str<strong>on</strong>g>European</str<strong>on</strong>g><br />
Commissi<strong>on</strong>’s IDABC programme, aims to initiate, support <strong>and</strong> coordinate the collaborati<strong>on</strong> between<br />
volunteer Member States in defining a “Comm<strong>on</strong> Assessment Method <strong>for</strong> St<strong>and</strong>ards <strong>and</strong> Specificati<strong>on</strong>s” <strong>and</strong> to<br />
share the assessment study results <strong>for</strong> the development of eGovernment services.<br />
21] www.semic.eu/semic/view/snav/shared-development.xhtml . SEMIC.EU is a participatory plat<strong>for</strong>m <strong>and</strong> a service<br />
by the <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong> that supports the sharing of assets of interoperability to be used in public administrati<strong>on</strong><br />
<strong>and</strong> eGovernment.
public administrati<strong>on</strong>s may either develop the specificati<strong>on</strong>s themselves <strong>and</strong> put <strong>for</strong>ward<br />
the result <strong>for</strong> st<strong>and</strong>ardizati<strong>on</strong>, or request a new <strong>for</strong>malised specificati<strong>on</strong> to be developed<br />
by st<strong>and</strong>ards developing organisati<strong>on</strong>s.<br />
» » Even where existing <strong>for</strong>malised specificati<strong>on</strong>s are available, they evolve over time<br />
<strong>and</strong> experience shows that revisi<strong>on</strong>s often take a l<strong>on</strong>g time to be completed. Active<br />
government participati<strong>on</strong> in the st<strong>and</strong>ardizati<strong>on</strong> process mitigates c<strong>on</strong>cerns about<br />
delays, improves alignment of the <strong>for</strong>malised specificati<strong>on</strong>s with public sector needs <strong>and</strong><br />
can help governments keep pace with technology innovati<strong>on</strong>.<br />
In the c<strong>on</strong>text of the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g>, it is essential that the research communities<br />
who need e-infrastructures <strong>for</strong> their work define their requirements of the relevant<br />
e-infrastructures. Without such definiti<strong>on</strong>s <strong>and</strong> c<strong>on</strong><strong>for</strong>mance, little can be d<strong>on</strong>e to furnish<br />
st<strong>and</strong>ards-compliant soluti<strong>on</strong>s that meet any community requirements. They should also<br />
support <strong>and</strong> c<strong>on</strong>tribute to the current st<strong>and</strong>ardizati<strong>on</strong> initiatives <strong>and</strong> not seek to re-invent<br />
wheels. As an interim measure they should c<strong>on</strong>sider building adaptors to fill gaps in the<br />
st<strong>and</strong>ards l<strong>and</strong>scape, but adapters should not be seen as the l<strong>on</strong>g term soluti<strong>on</strong> to achieve<br />
interoperability.<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
11
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
Internati<strong>on</strong>al Co-ordinati<strong>on</strong><br />
Work <strong>on</strong> the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> roadmap complements that of the far larger US Nati<strong>on</strong>al Institute<br />
of St<strong>and</strong>ards <strong>and</strong> Technology (NIST) <strong>Cloud</strong> Computing Program 22 . A US Federal <strong>Cloud</strong><br />
Computing Strategy document has been released which outlines the Federal Government’s<br />
approaches to <strong>Cloud</strong> Computing 23 . The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> project is c<strong>on</strong>cerned with e-infrastructure <strong>for</strong><br />
research including grids <strong>and</strong> clouds. The NIST program is c<strong>on</strong>cerned with government use<br />
of cloud computing. The NIST SAJACC initiative 24 develops cloud system use cases to drive<br />
the <strong>for</strong>mati<strong>on</strong> of cloud computing st<strong>and</strong>ards.<br />
Cross communicati<strong>on</strong> between <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <strong>and</strong> the NIST program is proving beneficial. A number<br />
of members of the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> REB are also participants in the NIST cloud computing expert<br />
group.<br />
Similar work is going <strong>on</strong> in Japan 25 China 26 <strong>and</strong> other countries. The NIST program in the<br />
US, GICTF in Japan, <strong>and</strong> CESI in China are all potential partners in evaluating potential cloud<br />
st<strong>and</strong>ards relevant <strong>for</strong> <str<strong>on</strong>g>European</str<strong>on</strong>g> e-infrastructure.<br />
12<br />
22] www.nist.gov/itl/cloud/index.cfm, collaborate.nist.gov/twiki-cloud-computing/bin/view/<strong>Cloud</strong>Computing/<br />
WebHome<br />
23] Federal <strong>Cloud</strong> Computing Strategy - Vivek Kundra U.S. Chief In<strong>for</strong>mati<strong>on</strong> Officer, February 8th 2011. www.nist.<br />
gov/itl/cloud/<br />
24] www.nist.gov/itl/cloud/sajacc.cfm<br />
25] See www.gictf.jp/index_e.html <strong>and</strong> the presentati<strong>on</strong> “Smart <strong>Cloud</strong> Strategy in Japan” by Yasu Taniwaki, Divisi<strong>on</strong><br />
Director, ICT Strategy Divisi<strong>on</strong>, Japanese Ministry of Internal Affairs <strong>and</strong> Communicati<strong>on</strong>s, November 2010<br />
items-int.eu/IMG/pdf/1011_Smart_<strong>Cloud</strong>_Strategy_Global_Forum_.pdf<br />
26] www.en.cesi.cn
<strong>Cloud</strong>s St<strong>and</strong>ards Coordinati<strong>on</strong><br />
<strong>Cloud</strong> st<strong>and</strong>ardisati<strong>on</strong> ef<strong>for</strong>ts led by the Distributed Management Task Force (DMTF),<br />
the Storage Networking Industry Associati<strong>on</strong> (SNIA) <strong>and</strong> the Open <strong>Grid</strong> Forum<br />
(OGF) are frequently cited as being enablers that could have a major impact <strong>on</strong> compute<br />
infrastructure in the future. Work <strong>on</strong> additi<strong>on</strong>al st<strong>and</strong>ards <strong>for</strong> various aspects of cloudbased<br />
services is underway in the Organisati<strong>on</strong> <strong>for</strong> Advancement of Structured In<strong>for</strong>mati<strong>on</strong><br />
St<strong>and</strong>ards (OASIS) <strong>and</strong> the Internet Engineering Task Force (IETF). At the same time, market<br />
adopti<strong>on</strong> of some of these st<strong>and</strong>ards is mixed, <strong>and</strong> different regi<strong>on</strong>s (US, China, Japan) are<br />
still evaluating their approaches to cloud st<strong>and</strong>ards, so it is difficult to predict whether<br />
c<strong>on</strong>sensus will emerge in the near term. The st<strong>and</strong>ards listed below that have emerged from<br />
analysis of use cases collected to date are being coordinated through an alliance between<br />
the OGF <strong>and</strong> SNIA as well as through a cross-SDO cloud st<strong>and</strong>ards collaborati<strong>on</strong> group 27 :<br />
»»<br />
Open Virtualizati<strong>on</strong> Format (OVF) 28 developed by DMTF. OVF is a packaging st<strong>and</strong>ard<br />
designed to address the portability <strong>and</strong> deployment of virtual appliances. This is<br />
recognised as a DMTF, ANSI st<strong>and</strong>ard categorized under IaaS, Interoperability. There are<br />
firms who provide tools <strong>for</strong> c<strong>on</strong>versi<strong>on</strong> between various appliance <strong>for</strong>mats, including<br />
OVF <strong>for</strong>mat to Amaz<strong>on</strong> Machine Image (AMI) <strong>for</strong>mat. 29<br />
»»<br />
The Open <strong>Cloud</strong> Computing Interface (OCCI) 30 developed by the OGF. OCCI describes<br />
applicati<strong>on</strong> programming interfaces (APIs) that enable cloud providers to expose their<br />
services. It focuses <strong>on</strong> “IaaS” based clouds <strong>and</strong> allows the deployment, m<strong>on</strong>itoring<br />
<strong>and</strong> management of virtual workloads (like virtual machines), but is applicable to any<br />
interacti<strong>on</strong> with a virtual cloud resource through defined http(s) header fields <strong>and</strong><br />
extensi<strong>on</strong>s. While there are several open-source implementati<strong>on</strong>s, OCCI has not yet been<br />
widely adopted in commercial plat<strong>for</strong>ms. OCCI is also an input to the DMTF st<strong>and</strong>ard <strong>for</strong><br />
cloud management.<br />
»»<br />
The <strong>Cloud</strong> Data Management Interface (CDMI) 31 developed by SNIA. CDMI defines the<br />
functi<strong>on</strong>al interface that applicati<strong>on</strong>s use to create, retrieve, update <strong>and</strong> delete data<br />
elements from the <strong>Cloud</strong>. CDMI is not yet widely implemented in commercial plat<strong>for</strong>ms.<br />
Other st<strong>and</strong>ards may emerge that enable interoperability between clouds <strong>and</strong> grids. For<br />
example, the OGF GLUE 32 st<strong>and</strong>ard provides <strong>on</strong>e in<strong>for</strong>mati<strong>on</strong> model <strong>for</strong> describing grid<br />
<strong>and</strong> cloud entities while the CIM model from DMTF 33 provides an alternative model used<br />
frequently in industry.<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
13<br />
27] See the summary at www.ogf.org/st<strong>and</strong>ards/; the <strong>Cloud</strong> St<strong>and</strong>ards Wiki is available at<br />
cloud-st<strong>and</strong>ards.org<br />
28] A descripti<strong>on</strong> is available at dmtf.org/st<strong>and</strong>ards/ovf<br />
29] aws.amaz<strong>on</strong>.com/amis/<br />
30] occi-wg.org/<br />
31] www.snia.org/tech_activities/st<strong>and</strong>ards/curr_st<strong>and</strong>ards/cdmi/<br />
32] GLUE Specificati<strong>on</strong> v. 2.0, by S. Andreozzi (INFN); S. Burke (RAL); F. Ehm (CERN); L. Field (CERN); G. Galang (ARCS);<br />
B. K<strong>on</strong>ya (Lund University); M. Litmaath (CERN); P. Millar (DESY); JP Navarro (ANL). March 2009<br />
www.ogf.org/documents/GFD.147.pdf<br />
33] www.dmtf.org/st<strong>and</strong>ards/cim
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
C<strong>on</strong>clusi<strong>on</strong>s Recommendati<strong>on</strong>s Future Directi<strong>on</strong>s<br />
The most important recommendati<strong>on</strong> of this roadmap is to:<br />
Undertake determined <strong>and</strong> targeted ef<strong>for</strong>ts to discourage fragmentati<strong>on</strong>, while at<br />
the same time preserving innovati<strong>on</strong> in the development of e-infrastructure.<br />
In support of this recommendati<strong>on</strong> we believe the following acti<strong>on</strong>s are necessary by all<br />
stakeholders to achieve the desired outcomes:<br />
Fund participati<strong>on</strong> in the l<strong>on</strong>g-term development of an adequate set of open st<strong>and</strong>ards<br />
to ensure the interoperability of future <str<strong>on</strong>g>European</str<strong>on</strong>g> infrastructures <strong>for</strong> research <strong>and</strong><br />
e-government.<br />
Public sector <strong>and</strong> commercial providers should engage more to explore shared st<strong>and</strong>ards<br />
requirements.<br />
An <strong>on</strong>going process is needed to track emerging st<strong>and</strong>ards, technologies, <strong>and</strong> best practices<br />
in order to create <strong>and</strong> maintain a structured repository of open st<strong>and</strong>ards (from various<br />
SDOs) <strong>for</strong> grids <strong>and</strong> clouds, <strong>and</strong> provide updated guidance to <str<strong>on</strong>g>European</str<strong>on</strong>g> e-infrastructure<br />
projects. This activity will benefit from interacti<strong>on</strong> with worldwide initiatives <strong>and</strong> other<br />
<str<strong>on</strong>g>European</str<strong>on</strong>g> projects (e.g. NIST, GICTF, CESI, CAMSS 34 , SEMIC.eu 35 , etc.).<br />
Encourage <strong>and</strong> fund the definiti<strong>on</strong> of sound security policies c<strong>on</strong>cerning the access, use<br />
<strong>and</strong> provisi<strong>on</strong>ing of services within distributed infrastructures.<br />
Introduce guidelines <strong>for</strong> dealing with data privacy, l<strong>on</strong>g term data curati<strong>on</strong>, liability <strong>and</strong><br />
taxati<strong>on</strong> issues in clouds <strong>and</strong> grids <strong>for</strong> work across legislative boundaries.<br />
14<br />
Fund procurement of open source or commercially provided software soluti<strong>on</strong>s<br />
allowing the research community to innovate in areas where they can add unique value<br />
bey<strong>on</strong>d the scope of commercial soluti<strong>on</strong>s.<br />
Fund <strong>on</strong>-dem<strong>and</strong> cross-domain provisi<strong>on</strong>ing of high-speed data transfer links (light<br />
paths) with defined service level agreements.<br />
Involve <str<strong>on</strong>g>European</str<strong>on</strong>g>s citizens in e-science through volunteer computing (using, e.g.,<br />
desktop grids <strong>and</strong> clouds).<br />
34] ec.europa.eu/idabc/en/document/7407.html. See also footnote n. 20.<br />
35] www.semic.eu/semic/. See also footnote n. 21.
Target Audience<br />
This initial draft document is <strong>for</strong> circulati<strong>on</strong> to the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> Editorial Board (REB),<br />
Industry Expert Group (IEG), Special Liais<strong>on</strong> Group (SLG) <strong>and</strong> the <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong>.<br />
Timeline<br />
Since October 2010, REB members have been c<strong>on</strong>tributing material to the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> Wiki. The<br />
material is structured according to a table of c<strong>on</strong>tents <strong>for</strong> a final document. This initial draft<br />
has been prepared as a <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> deliverable to the EC. The REB has developed a publishable<br />
versi<strong>on</strong> circulated at <strong>Cloud</strong>scape-III (Brussels, 15-16/03/2011). The REB will then integrate<br />
further elements, namely the use cases presented at <strong>Cloud</strong>scape III from <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <strong>and</strong> NIST.<br />
Scope<br />
This document addresses requirements, technologies, <strong>and</strong> interoperability <strong>and</strong> st<strong>and</strong>ards<br />
<strong>for</strong> e-infrastructure to support existing, <strong>on</strong>going, <strong>and</strong> future research in the <str<strong>on</strong>g>European</str<strong>on</strong>g><br />
Research Area. The term e-infrastructure encompasses the distributed in<strong>for</strong>mati<strong>on</strong> <strong>and</strong><br />
communicati<strong>on</strong>s technologies (ICTs), together with federating software, that together<br />
provide services <strong>and</strong> access to resources needed to support public sectors such as research<br />
in the natural <strong>and</strong> social sciences <strong>and</strong> humanities. While not a focus of this specific<br />
document, some c<strong>on</strong>siderati<strong>on</strong> is given to aspects of e-infrastructure that apply also to<br />
e-government. The most recent <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong> call under Framework Programme 7<br />
<strong>for</strong> proposals relevant to e-infrastructure can be found in the <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong> Work<br />
Programme 2011 Capacities Part 1 Research Infrastructures 36 .<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
15<br />
36] cordis.europa.eu/fp7/wp-2011_en.html
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
<str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> Editorial Board (REB) Member List<br />
REB Member Role & Organisati<strong>on</strong> Country<br />
John Borras Independent C<strong>on</strong>sultant & OASIS United Kingdom<br />
Goetz-Philip Brasche<br />
Mark Carls<strong>on</strong><br />
Guy Coates<br />
Juan Cáceres<br />
Program Director <strong>Cloud</strong> Computing EMIC<br />
& Venus-C representative<br />
Senior Architect, Oracle & SNIA & DMTF<br />
representative<br />
Group leader, In<strong>for</strong>matics systems group<br />
at Wellcome Trust Sanger Institute<br />
Middleware Technologies Specialist,<br />
Telefónica I+D & StratusLab<br />
representative<br />
Germany<br />
United States<br />
United Kingdom<br />
Spain<br />
Michel Drescher EGI.eu Technical Manager The Netherl<strong>and</strong>s<br />
Åke Edlund<br />
Mike Fisher<br />
Patrick Guillemin<br />
KTH project manager <strong>and</strong> researcher &<br />
ECEE representative<br />
Distributed Computing Research<br />
Group Leader BT & Chair of Technical<br />
Committee, ETSI<br />
ETSI Secretariat, Strategy & New<br />
Initiatives<br />
Sweden<br />
United Kingdom<br />
France<br />
Jenny Huang AT&T , OMG representative United States<br />
Gersh<strong>on</strong> Janssen<br />
Independent C<strong>on</strong>sultant & OASIS<br />
St<strong>and</strong>ards Group representative<br />
The Netherl<strong>and</strong>s<br />
16<br />
Craig Lee The Aerospace Corporati<strong>on</strong> United States<br />
Bob Marcus ET-Strategies United States<br />
Ignacio Martin Llorente<br />
Complutense University of Madrid &<br />
OpenNebula representative<br />
Spain<br />
Steven Newhouse EGI.eu Director & EGI-InSPIRE Director The Netherl<strong>and</strong>s<br />
Alex<strong>and</strong>er Papaspyrou<br />
Morris Riedel<br />
Alan Sill<br />
Technische Univ. Dortmund & IGE<br />
representative<br />
Jülich Supercomputing Centre & EMI<br />
representative<br />
VP of St<strong>and</strong>ards, OGF & Senior Scientist,<br />
Texas Tech University<br />
Germany<br />
Germany<br />
United States<br />
Etienne Urbah LAL, Univ Paris-Sud & EDGI representative France<br />
Martin Ant<strong>on</strong>y Walker Independent C<strong>on</strong>sultant & REB Chair France
<str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> c<strong>on</strong>tent has been c<strong>on</strong>tributed by members of the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> Editorial Board<br />
(REB) <strong>and</strong> Industry Expert <strong>and</strong> Special Liais<strong>on</strong> Groups (IEG <strong>and</strong> SLG), who also c<strong>on</strong>tributed<br />
to the editing process. <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> c<strong>on</strong>tent structuring, producti<strong>on</strong>, <strong>and</strong> final editing were<br />
d<strong>on</strong>e by Martin Ant<strong>on</strong>y Walker, REB chair, John Borras, co-chair, <strong>and</strong> Steven Newhouse,<br />
Director of EGI.eu <strong>and</strong> EGI-InSPIRE, with c<strong>on</strong>tributi<strong>on</strong>s by Silvana Muscella, <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> technical<br />
coordinator, <strong>and</strong> James Ahtes, ATOS Origin. Organisati<strong>on</strong> <strong>and</strong> coordinati<strong>on</strong> of the REB <strong>and</strong><br />
editorial activities have been carried out by the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> c<strong>on</strong>sortium.<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
17
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong> e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
<str<strong>on</strong>g>SIENA</str<strong>on</strong>g> Project Descripti<strong>on</strong> – www.sienainitiative.<br />
eu<br />
S<br />
IENA (RI-261575) the St<strong>and</strong>ards <strong>and</strong> Interoperability <strong>for</strong> eInfrastructure Implementati<strong>on</strong><br />
Initiative (2010-2012), is a Support Acti<strong>on</strong> funded by the <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong><br />
under Framework Programme 7 (2007-2013) Research infrastructures projects. <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> will<br />
c<strong>on</strong>tribute to defining a future eInfrastructures roadmap focusing <strong>on</strong> interoperability <strong>and</strong><br />
st<strong>and</strong>ards, in close collaborati<strong>on</strong> with the <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong>, Distributed Computing<br />
Infrastructures (DCI) projects <strong>and</strong> St<strong>and</strong>ard Development Organisati<strong>on</strong>s (SDOs) to gain an<br />
in-depth underst<strong>and</strong>ing of how distributed computing technology is being developed in<br />
this c<strong>on</strong>text. The roadmap will define scenarios, identify trends, investigate the innovati<strong>on</strong><br />
<strong>and</strong> impact sparked by cloud <strong>and</strong> grid computing, <strong>and</strong> deliver insight into how st<strong>and</strong>ards<br />
<strong>and</strong> the policy framework is defining <strong>and</strong> shaping current <strong>and</strong> future development <strong>and</strong><br />
deployment in Europe <strong>and</strong> globally.<br />
18
15-16 March 2011<br />
Brussels, Belgium<br />
Use Cases &<br />
Positi<strong>on</strong> Papers
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
Index<br />
Introducti<strong>on</strong>............................................................................................................................................................21<br />
Uses <strong>and</strong> perspectives from <strong>Science</strong> <strong>and</strong> Research<br />
BiG<strong>Grid</strong> HPC <strong>Cloud</strong>.............................................................................................................................................23<br />
Biology <strong>on</strong> the <strong>Cloud</strong>.........................................................................................................................................25<br />
CONTRAIL - Open Computing Infrastructures <strong>for</strong> Elastic Services...............................................27<br />
RESERVOIR - IaaS <strong>Cloud</strong> Interoperability..................................................................................................29<br />
T<strong>Cloud</strong>s - Trustworthy <strong>Cloud</strong> Computing..................................................................................................31<br />
<str<strong>on</strong>g>European</str<strong>on</strong>g> Distributed Computing Infrastructures<br />
EDGI, DEGISCO & IDGF - <str<strong>on</strong>g>European</str<strong>on</strong>g> Desktop <strong>Grid</strong> Initiative, Desktop <strong>Grid</strong>s <strong>for</strong> Internati<strong>on</strong>al<br />
Scientific Collaborati<strong>on</strong> & Internati<strong>on</strong>al Desktop <strong>Grid</strong> Federati<strong>on</strong>................................................ 33<br />
EGI - <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Grid</strong> Infrastructure.............................................................................................................. 35<br />
EMI - <str<strong>on</strong>g>European</str<strong>on</strong>g> Middleware Initiative......................................................................................................... 37<br />
IGE - Initiative <strong>for</strong> Globus in Europe............................................................................................................39<br />
StratusLab - Enhancing <strong>Grid</strong> Infrastructures with Virtualizati<strong>on</strong> <strong>and</strong> <strong>Cloud</strong> Technologies... 41<br />
VENUS-C - Virtual Multidisciplinary Envir<strong>on</strong>ments using <strong>Cloud</strong> Infrastructures.....................43<br />
Business & Government<br />
The shift to cloud computing in government in the EU......................................................................45<br />
G-<strong>Cloud</strong> - UK Government <strong>Cloud</strong> Computing Infrastructure...........................................................48<br />
CitySourced/FreedomSpeaks citizen services plat<strong>for</strong>m....................................................................50<br />
CUSTOM - Cultural Heritage & Tourism Store <strong>on</strong> the <strong>Cloud</strong>............................................................52<br />
20<br />
St<strong>and</strong>ards & Interfaces<br />
OpenNebula - A reference open cloud stack to enable interoperable enterprise-class<br />
cloud computing plat<strong>for</strong>ms............................................................................................................................54<br />
OCCI - Open <strong>Cloud</strong> Computing Interface specificati<strong>on</strong> set..............................................................56<br />
Legal, Ec<strong>on</strong>omic, Ethical <strong>and</strong> Security Issues<br />
<strong>Cloud</strong> computing <strong>and</strong> its ethical challenges............................................................................................58<br />
VENUS-C study <strong>on</strong> ec<strong>on</strong>omic <strong>and</strong> legal implicati<strong>on</strong>s of sustainable scientific clouds......... 60<br />
The <strong>Cloud</strong>: Underst<strong>and</strong>ing security, privacy <strong>and</strong> trust challenges..................................................62
Introducti<strong>on</strong><br />
<strong>Cloud</strong>scape III use cases <strong>and</strong> Positi<strong>on</strong> Papers <strong>for</strong> the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g><br />
<str<strong>on</strong>g>European</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> <strong>Grid</strong> <strong>and</strong> <strong>Cloud</strong> St<strong>and</strong>ards <strong>for</strong><br />
e-<strong>Science</strong> <strong>and</strong> Bey<strong>on</strong>d<br />
<strong>Cloud</strong> computing is expected to play a key role in the digital ec<strong>on</strong>omy in Europe <strong>and</strong> bey<strong>on</strong>d.<br />
To ensure <str<strong>on</strong>g>European</str<strong>on</strong>g> citizens gain real benefits from the cloud, it is essential that we address<br />
legal <strong>and</strong> instituti<strong>on</strong>al barriers, as well as technical challenges such as interoperability.<br />
The <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> <strong>on</strong> grids <strong>and</strong> clouds <strong>for</strong> <str<strong>on</strong>g>European</str<strong>on</strong>g> research infrastructures <strong>and</strong> public<br />
services addresses interoperability <strong>and</strong> st<strong>and</strong>ards <strong>and</strong> in the next 15 m<strong>on</strong>ths is committed to<br />
delivering a policy framework <strong>for</strong> distributed computing that ensures fair competiti<strong>on</strong> <strong>and</strong><br />
brings to bear <str<strong>on</strong>g>European</str<strong>on</strong>g> strategic priorities.<br />
To help achieve these goals, the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> c<strong>on</strong>sortium is drawing <strong>on</strong> <strong>Cloud</strong>scape III to showcase<br />
speakers from all over the globe who will offer their pers<strong>on</strong>al insights <strong>on</strong> specific use cases<br />
or interoperability issues surrounding <strong>Cloud</strong> computing.<br />
The following use cases <strong>and</strong> positi<strong>on</strong> papers have been collected <strong>for</strong> the <strong>Cloud</strong>scape III<br />
event, serving primarily as a sample of the cloud computing l<strong>and</strong>scape. They highlight<br />
potential challenges <strong>for</strong> deliberati<strong>on</strong> at <strong>Cloud</strong>scape III <strong>and</strong> <strong>for</strong> the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> Editorial<br />
Board in the coming m<strong>on</strong>ths, with the aim of shaping future developments <strong>and</strong> the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g><br />
<str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> itself.<br />
The full collecti<strong>on</strong> of use cases <strong>and</strong> positi<strong>on</strong> papers are available at<br />
www.sienainitiative.eu/<strong>Cloud</strong>scapeIII-UseCases&Positi<strong>on</strong>Papers/<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
21
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
22
Uses <strong>and</strong> perspectives from <strong>Science</strong> <strong>and</strong> Research<br />
BiG<strong>Grid</strong> HPC <strong>Cloud</strong><br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
With the newly developed BiG<strong>Grid</strong> High Per<strong>for</strong>mance Computing (HPC) <strong>Cloud</strong> envir<strong>on</strong>ment,<br />
scientific researchers get access to their own Virtual Private HPC Cluster. It is a virtualized HPC<br />
Cluster that users can c<strong>on</strong>figure to exactly match their needs, without interfering with the needs<br />
of other users. It is flexible, offers self service <strong>and</strong> is dynamically scalable.<br />
Users can start from existing templates (images), or build their own cluster from scratch. It is<br />
even possible to make a copy from their current IT software envir<strong>on</strong>ment (<strong>for</strong> example their<br />
laptop or desktop pc) <strong>and</strong> turn that into a HPC cluster in our <strong>Cloud</strong>. In that way, there will be very<br />
little difference between their development envir<strong>on</strong>ment <strong>and</strong> their producti<strong>on</strong> envir<strong>on</strong>ment.<br />
There is no need <strong>for</strong> an (expensive) rewrite of their software, <strong>and</strong> scientific challenges can be<br />
scaled up very easily from desktop scale to High Per<strong>for</strong>mance Compute cluster scale.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The importance of interoperability<br />
For us the most important part of <strong>Cloud</strong> st<strong>and</strong>ards is that we offer infrastructure as a service,<br />
but we want to hide all the differences <strong>and</strong> little details of hardware behind an abstract<br />
interface or API. For example, it does not really matter which <strong>Cloud</strong> middleware we use <strong>and</strong><br />
which OS runs <strong>on</strong> the hosts to deploy our VMs, we use OCCI as an interface between our<br />
GUI <strong>and</strong> OpenNebula. Also, we are finalizing an implementati<strong>on</strong> of CDMI to have the same<br />
setup <strong>for</strong> storage. CDMI will hide the complexities <strong>for</strong> users of where data is located in a<br />
distributed cloud <strong>and</strong> which protocols they can use to access it. Also through CDMI users<br />
can deploy a storage volume <strong>and</strong> manage their data, including fine grained authorizati<strong>on</strong>s,<br />
without manual steps by our administrators.<br />
23<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
We are also starting to work <strong>on</strong> an API <strong>for</strong> network c<strong>on</strong>figurati<strong>on</strong>s. Our users will be able to<br />
manage many network settings by themselves, <strong>for</strong> example the creati<strong>on</strong> of a VLAN between<br />
VMs, setting firewall rules <strong>and</strong> setting up secure c<strong>on</strong>necti<strong>on</strong>s to their virtual machines.<br />
Our goal is that we fully automate the management of virtual HPC clusters. All (skilled) end<br />
users can be completely self supporting <strong>and</strong> can access <strong>and</strong> c<strong>on</strong>figure their virtual private<br />
HPC cluster in the BiG<strong>Grid</strong> HPC <strong>Cloud</strong> through a secure <strong>and</strong> functi<strong>on</strong>ally complete API.<br />
When these st<strong>and</strong>ards <strong>for</strong> compute, storage <strong>and</strong> network are complete, it can also be used<br />
between <strong>Cloud</strong> clusters/providers to (automatically) negotiate migrati<strong>on</strong> of workloads.<br />
Security c<strong>on</strong>figurati<strong>on</strong>s are especially important <strong>for</strong> this use case.
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
Finally, st<strong>and</strong>ards should be open, so everybody can benefit <strong>and</strong> end users will actually have<br />
a choice of where to deploy.<br />
Possible future cooperati<strong>on</strong><br />
ECEE – Enabling <strong>Cloud</strong>s <strong>for</strong> e<strong>Science</strong> – is an open collaborati<strong>on</strong> spot <strong>for</strong> cloud projects in<br />
Europe. The purpose with ECEE is to share experiences to find out as much as possible, as<br />
quick as possible, about how clouds can help our users in their daily work.<br />
e<strong>Science</strong> projects involved so far are NEON, Baltic<strong>Cloud</strong>, NGS, GRNET cloud, SARA cloud,<br />
UCM (OpenNebula), StratusLab, VENUS-C, SEECCI <strong>and</strong> CESGA – which together represent<br />
a fair share of the <str<strong>on</strong>g>European</str<strong>on</strong>g> cloud community. ECEE focus <strong>on</strong> interoperability-now, sharing<br />
its input <strong>and</strong> requirements with <strong>on</strong>going st<strong>and</strong>ardizati<strong>on</strong> ef<strong>for</strong>ts. Meeting twice a year since<br />
OGF28 in March 2010, the projects together share roadmaps, experiences <strong>and</strong> issues – trying<br />
to identify: a comm<strong>on</strong> roadmap over all; gap analysis; “Market analysis” – today’s users,<br />
tomorrow’s; Guidelines – best practices, quick start <strong>on</strong>e-pager, checklists <strong>and</strong> practical ‘rules<br />
of thumb’. A number of Focus Areas were identified at an early stage including: Security,<br />
Metering, Accounting, Billing, Business models, Federati<strong>on</strong> of clouds, Network <strong>and</strong> Licences,<br />
Scheduling, load balancing (resource sharing, applicati<strong>on</strong> correlati<strong>on</strong>) <strong>and</strong> in making a list of<br />
tested soluti<strong>on</strong>s, <strong>and</strong> their pros <strong>and</strong> c<strong>on</strong>s.<br />
C<strong>on</strong>tacts: Floris Sluiter Ake Edlund<br />
Organisati<strong>on</strong>; SARA HPC centre KTH Royal Institute of Technology<br />
C<strong>on</strong>tact details: floris@sara.nl edlund@nada.kth.se<br />
Web:<br />
www.cloud.sara.nl<br />
24
Biology <strong>on</strong> the <strong>Cloud</strong><br />
The <strong>Cloud</strong> provides a wide range of infrastructure <strong>and</strong> software services that can be used<br />
by the Biology user community. Indeed, experienced technical computing users are already<br />
finding ways in which to use these services to augment their existing computing resources.<br />
The greater promise of the cloud is that it can make technical computing pervasive,<br />
opening up the field to new researchers who have not been traditi<strong>on</strong>al HPC users. These<br />
researchers will be able to co-opt sophisticated cloud services provided by both academia<br />
<strong>and</strong> commercial providers to aid them in their research. In this paper I will showcase two<br />
Biology <strong>Cloud</strong> use cases which offer a number of advantages to users.<br />
IaaS: Web-services Mirrors<br />
The Ensembl project provides a variety of web services which allows researchers to visualise<br />
<strong>and</strong> data-mine genomic data (www.ensembl.org). Ensembl has a world-wide audience <strong>and</strong> is<br />
accessed 24 hours a day. Historically, the web service was hosted in a single UK datacentre.<br />
Whilst this provided fast access to users in the UK <strong>and</strong> Europe, users in Asia <strong>and</strong> the<br />
Americas found that access to the web services was slow, due the large latencies involved<br />
in serving requests across the globe. Single site hosting also made the website vulnerable<br />
to datacentre <strong>and</strong> network outages.<br />
The global, distributed nature of commercial <strong>Cloud</strong> IaaS make them a useful building block<br />
<strong>for</strong> providing world-wide availability <strong>and</strong> reach. Ensembl has used public IaaS providers to<br />
build mirrors of its web services in the United States of America <strong>and</strong> Asia. Not <strong>on</strong>ly has this<br />
massively increased the per<strong>for</strong>mance of the website <strong>for</strong> n<strong>on</strong> <str<strong>on</strong>g>European</str<strong>on</strong>g> users, but it also<br />
provides c<strong>on</strong>tinued availability of service when the UK datacentre is offline.<br />
<strong>Cloud</strong> hosting provides several advantages over hosting in a traditi<strong>on</strong>al co-locati<strong>on</strong> facility.<br />
Installing real hardware in a remote co-locati<strong>on</strong> facility requires time-c<strong>on</strong>suming <strong>and</strong> costly<br />
logistics. Hardware has to be shipped to the facility <strong>and</strong> cleared through customs, <strong>and</strong><br />
staff need to be present <strong>on</strong> site to oversee hardware installati<strong>on</strong> <strong>and</strong> initial provisi<strong>on</strong>ing.<br />
In c<strong>on</strong>trast, provisi<strong>on</strong>ing virtual hardware in a remote cloud IaaS facility can be d<strong>on</strong>e from<br />
any locati<strong>on</strong> with internet access, whilst the “<strong>on</strong>-dem<strong>and</strong>” facilities allow machines to be<br />
provisi<strong>on</strong>ed within a matter of minutes<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
25<br />
SaaS: Providing In<strong>for</strong>matics services <strong>for</strong> Next-Generati<strong>on</strong><br />
Sequencing (NGS)<br />
SasS provides new opportunities <strong>for</strong> organisati<strong>on</strong>s to provide IT services to researchers.<br />
IT service provisi<strong>on</strong> <strong>for</strong> next-generati<strong>on</strong> sequencing machines is a huge challenge. A single<br />
sequencing instrument can produce approximately a terabyte of raw data per day <strong>and</strong> a<br />
large sequencing study may end up with a total dataset of many hundreds of terabytes.<br />
Dealing with this data is a challenge <strong>for</strong> organisati<strong>on</strong>s of all sizes, whether they are a small<br />
lab with a single machine, or a large sequencing centre with many tens of machines.
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
Although sequencing manufacturers provide basic analysis software <strong>for</strong> their machines,<br />
there is a whole extended eco-system of software that researchers typically want to run <strong>on</strong><br />
their data. The large volumes of data means that labs need to integrate their instruments<br />
with a LIMS (Laboratory In<strong>for</strong>mati<strong>on</strong> Management System), in order to organise <strong>and</strong> track<br />
their data. Researchers will also want to run down-stream analysis <strong>on</strong> their data <strong>on</strong>ce it<br />
comes off the sequencers; raw sequence data is typically <strong>on</strong>ly the first stage in a scientific<br />
investigati<strong>on</strong>. Down-stream analysis software is typically complex, <strong>and</strong> requires a highper<strong>for</strong>mance<br />
computing (HPC) infrastructure.<br />
Rather than having to provide software <strong>and</strong> HPC support in house, the <strong>Cloud</strong> SaaS model<br />
allows researchers to obtain LIMs <strong>and</strong> data analysis services from specialised bio-in<strong>for</strong>matics<br />
suppliers.<br />
Using this model, researchers run a sequencing experiment in-house, <strong>and</strong> the raw data is then<br />
uploaded to the SaaS providers, who will then analyse, track <strong>and</strong> store their data. Researchers<br />
are there<strong>for</strong>e freed from having to manage their own LIMs <strong>and</strong> HPC infrastructure.<br />
Whilst most sequencing SaaS is currently provided by commercial entities (eg https://<br />
www.seqcentral.com, www.dnanexus.com) opportunities also exist <strong>for</strong> academic cloud<br />
providers. Many large scale sequencing projects are carried out by large academic c<strong>on</strong>sortia,<br />
composed of many different organisati<strong>on</strong>s with differing specialities. (eg the Internati<strong>on</strong>al<br />
Cancer Genome C<strong>on</strong>sortium www.icgc.org). Members of the c<strong>on</strong>sortium with a high level<br />
of IT expertise can provide SaaS services to the whole of the c<strong>on</strong>sortium. These services<br />
may be hosted <strong>on</strong> the c<strong>on</strong>sortium’s own infrastructure, or <strong>on</strong> cloud IaaS provided by a third<br />
party. Private cloud SaaS provisi<strong>on</strong> within a c<strong>on</strong>sortium may be especially useful when<br />
data-privacy <strong>and</strong> security policies make it impractical to host data <strong>on</strong> third-party cloud<br />
services.<br />
Challenges remain. Although research organisati<strong>on</strong>s are c<strong>on</strong>nected by high speed networks,<br />
these networks are currently not well c<strong>on</strong>nected to the commercial networks used by<br />
commercial cloud providers. In practice, transfers of large amount of data into commercial<br />
cloud providers is time c<strong>on</strong>suming, <strong>and</strong> can limit the usefulness of SaaS services <strong>for</strong><br />
sequencing applicati<strong>on</strong>s, especially <strong>for</strong> organisati<strong>on</strong>s with limited network c<strong>on</strong>nectivity.<br />
26<br />
C<strong>on</strong>tact: Guy Coates<br />
Organisati<strong>on</strong>: Wellcome Sangar Institute<br />
C<strong>on</strong>tact details: gmpc@sanger.ac.uk<br />
Web: www.ensembl.org<br />
Relevant Links: www.seqcentral.com; www.dnanexus.com; www.icgc.org
CONTRAIL – Open Computing<br />
Infrastructures <strong>for</strong> Elastic Services<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
The C<strong>on</strong>trail project will deliver federated access to cloud resources. Single registrati<strong>on</strong> <strong>and</strong><br />
account management are core features of the use cases, where “account management” also<br />
includes roles <strong>and</strong> permissi<strong>on</strong>s, billing, resource allocati<strong>on</strong>s, etc. Services are selected based <strong>on</strong><br />
published service levels <strong>and</strong> “quality of protecti<strong>on</strong>,” as well as, of course, cost <strong>and</strong> permissi<strong>on</strong>s.<br />
Federated access must be transparent, with the federati<strong>on</strong> accessing, or enabling access to,<br />
remote cloud services <strong>on</strong> behalf of the user, but of course without incurring unexpected<br />
costs. Account management will thus need to include an internal ec<strong>on</strong>omic model.<br />
Briefly, the use cases (case studies) cover geo-referenced data, processing streaming<br />
multimedia, real-time high per<strong>for</strong>mance scientific data analysis, <strong>and</strong> drug discovery. Our<br />
user communities cover both industry <strong>and</strong> academic users. (The mapping of use cases to<br />
requirements is still <strong>on</strong>going.)<br />
C<strong>on</strong>trail will provide both PaaS <strong>and</strong> IaaS. The PaaS services will be using existing comp<strong>on</strong>ents<br />
<strong>for</strong> “structured storage” – a key/value store, a database infrastructure (using SQL), as well<br />
as hosting services enabling hosting of PHP applicati<strong>on</strong>s, MapReduce-enabled storage with<br />
Hadoop, <strong>and</strong> “bag-of-tasks” services. In additi<strong>on</strong> to the native interfaces, we will need<br />
interfaces <strong>for</strong> provisi<strong>on</strong>ing <strong>and</strong> managing PaaS resources.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The importance of interoperability<br />
Interoperability is very important to C<strong>on</strong>trail. As the federati<strong>on</strong> accesses services <strong>on</strong> behalf<br />
of users, having st<strong>and</strong>ard interfaces into clouds (such as OCCI from OGF <strong>and</strong> CDMI from<br />
SNIA) will be very useful. Otherwise, we will need to code an interface <strong>for</strong> each service<br />
provider which will limit the number of service providers we can support. As we currently<br />
plan to work with OpenNebula, we will support their interfaces.<br />
The other role of st<strong>and</strong>ards is to ensure that the interface remains stable: a proprietary<br />
interface could be changed by its owner, potentially without c<strong>on</strong>sulting us, whereas a<br />
st<strong>and</strong>ard managed by a st<strong>and</strong>ards body will have processes <strong>for</strong> updating protocols. In this<br />
respect, it would be useful to focus <strong>on</strong> open st<strong>and</strong>ards bodies <strong>and</strong>/or working groups,<br />
where the participati<strong>on</strong> is open <strong>and</strong> not prohibitively expensive.<br />
Whenever possible, we try to identify existing st<strong>and</strong>ards <strong>and</strong> evaluate them to see whether<br />
they are appropriate <strong>for</strong> C<strong>on</strong>trail. If not, we c<strong>on</strong>sider working with the st<strong>and</strong>ards working<br />
groups to augment their st<strong>and</strong>ard. While we reuse whenever possible, we will also seek<br />
st<strong>and</strong>ardisati<strong>on</strong> of our own work whenever appropriate. Having learnt from other EUprojects,<br />
we will identify work <strong>for</strong> potential st<strong>and</strong>ardisati<strong>on</strong> <strong>and</strong> collaborati<strong>on</strong> in st<strong>and</strong>ards<br />
bodies at an early stage in the project, to ensure that such work has a reas<strong>on</strong>able chance of<br />
completi<strong>on</strong> during the lifetime of C<strong>on</strong>trail. We make as much use as possible of collaborati<strong>on</strong><br />
27
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
28<br />
events <strong>and</strong> are currently working <strong>on</strong> identifying peer projects <strong>for</strong> collaborati<strong>on</strong>.<br />
There are additi<strong>on</strong>al benefits to collaborating <strong>on</strong> st<strong>and</strong>ards: we avoid duplicati<strong>on</strong> of ef<strong>for</strong>t,<br />
<strong>and</strong> get more ef<strong>for</strong>t behind the work by collaborating.<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
The maturity of st<strong>and</strong>ards – <strong>and</strong> their implementati<strong>on</strong>s – is very important: a st<strong>and</strong>ard which<br />
has more than <strong>on</strong>e implementati<strong>on</strong> behind it, at least <strong>on</strong>e in C or C++ <strong>and</strong> <strong>on</strong>e in Java, where<br />
the implementati<strong>on</strong>s are robust <strong>and</strong> independent of each other, <strong>and</strong> the underlying libraries are<br />
themselves mature, will be much more useful. We could in principle use a protocol which has a<br />
single implementati<strong>on</strong> (most of our own code will be implemented in Java), but C<strong>on</strong>trail will also<br />
need to interoperate with more than itself, so mature implementati<strong>on</strong>s should be preferred.<br />
As an example, there are many security-related st<strong>and</strong>ards from IETF, W3C, OASIS, ITU which<br />
are relevant to C<strong>on</strong>trail. We note that even very mature st<strong>and</strong>ards like X.509 certificates can<br />
pose interoperati<strong>on</strong> problems, <strong>and</strong> many later st<strong>and</strong>ards (e.g. in WS-Security) have themselves<br />
taken a l<strong>on</strong>g time to mature, <strong>and</strong> not all of these are usable yet. There is also a risk with new<br />
st<strong>and</strong>ards that they <strong>on</strong>ly partially implemented the specificati<strong>on</strong>, in which case we will need<br />
to know – or learn “the hard way” – which parts of the specificati<strong>on</strong> we can use.<br />
We are still reviewing existing st<strong>and</strong>ards <strong>for</strong> suitability <strong>for</strong> C<strong>on</strong>trail, as well as related work<br />
produced by other EU-funded projects. We are following interoperati<strong>on</strong> activities in OGF<br />
(e.g. GIN, PGI, <strong>and</strong> the proposed <strong>Cloud</strong>-BP (BP=Basic Profile, analogous to HPC-BP.)<br />
We see interoperati<strong>on</strong> testing happening mainly in collaborati<strong>on</strong>s with peer projects, <strong>and</strong>/<br />
or within the scope of st<strong>and</strong>ards bodies, not usually within C<strong>on</strong>trail itself.<br />
It is possible that we can help emerging st<strong>and</strong>ards mature by using them both within C<strong>on</strong>trail<br />
<strong>and</strong> in collaborati<strong>on</strong>s, but this will require more ef<strong>for</strong>t <strong>and</strong> will extend the development time<br />
<strong>for</strong> our own comp<strong>on</strong>ents. So, all other things being equal, a mature st<strong>and</strong>ard is preferred.<br />
We are likely to use (or at the very least evaluate) the following emerging st<strong>and</strong>ards:<br />
OCCI from OGF; CDMI from SNIA; Proposed extensi<strong>on</strong>s to XACML (to bring it in line with<br />
functi<strong>on</strong>ality in POLPA): DMTF st<strong>and</strong>ards may be relevant (OVF, “OVF+”); St<strong>and</strong>ards (if any)<br />
<strong>for</strong> managing workflow: AMQP – Advanced Message Queuing Protocol (www.amqp.org).<br />
Possible future cooperati<strong>on</strong><br />
Existing projects:<br />
SLA@SOI – SLA management, service management – uses Apache TASHI, <strong>and</strong> they claim<br />
their service manager is “based <strong>on</strong> OCCI”(?); MASTER - protecti<strong>on</strong> profiles, risks, trusted<br />
infrastructure; DEPLOY – <strong>for</strong>mal methods; <strong>Cloud</strong>4SOA; RESERVOIR framework <strong>for</strong> business<br />
applicati<strong>on</strong>s – applicati<strong>on</strong>s, SLA. Use of OpenNebula; StratusLab; mOSAIC.<br />
C<strong>on</strong>tact: Dr Christine MORIN<br />
Organisati<strong>on</strong>: INRIA Rennes<br />
C<strong>on</strong>tact details: c<strong>on</strong>trail-c<strong>on</strong>tact@inria.fr<br />
Web: c<strong>on</strong>trail-project.eu<br />
Dr Jens Jensen<br />
<strong>Science</strong> <strong>and</strong> Technology Facilities Council<br />
jens.jensen@stfc.ac.uk
RESERVOIR - IaaS <strong>Cloud</strong> Interoperability<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
The RESERVOIR project is developing an IaaS cloud computing plat<strong>for</strong>m with advanced<br />
features regarding current alternatives, such as automatic scalability <strong>and</strong> site federati<strong>on</strong>.<br />
The applicati<strong>on</strong>s to which RESERVOIR is aimed are multi-tier services that are deployed <strong>and</strong><br />
managed using the RESERVOIR middleware. The services dem<strong>on</strong>strated in the project range<br />
all applicati<strong>on</strong> fields, from GRID computing, corporate services (e.g. SAP), eGovernment<br />
<strong>and</strong> the telco industry. RESERVOIR architecture provides site federati<strong>on</strong> <strong>and</strong> functi<strong>on</strong>ality<br />
is split in three different middleware layers: Service Manager (SM), which provides holistic<br />
service management; Virtual Executi<strong>on</strong> Envir<strong>on</strong>ment Management (VEEM), which manages<br />
the virtual machines that compose the service implementing the federati<strong>on</strong> capabilities;<br />
<strong>and</strong> Virtual Executi<strong>on</strong> Envir<strong>on</strong>ment Host (VEEH) which implements the virtualizati<strong>on</strong><br />
plat<strong>for</strong>m (i.e. hypervisor).<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The importance of interoperability<br />
Interoperability is key in RESERVOIR <strong>and</strong> st<strong>and</strong>ards are used in three areas. Firstly, the<br />
service packaging <strong>for</strong>mat should leverage st<strong>and</strong>ard <strong>for</strong>mats, so the same services that<br />
customers get from ISVs, deploy in their in-house IT infrastructure <strong>and</strong>/or other cloud, can<br />
also be seamlessly deployed in RESERVOIR. Sec<strong>on</strong>dly, the deployment <strong>and</strong> management<br />
API used by users to interact with RESERVOIR cloud should be st<strong>and</strong>ardized. Thirdly, as<br />
RESERVOIR is composed of three independent middleware layers (Service Manager, Virtual<br />
Executi<strong>on</strong> Envir<strong>on</strong>ment Management <strong>and</strong> Virtual Executi<strong>on</strong> Envir<strong>on</strong>ment Host) that could<br />
be developed <strong>and</strong> provided independently, st<strong>and</strong>ard APIs between them are needed.<br />
29<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
In order to package the services that are deployed in RESERVOIR cloud, the Distributed<br />
Management Task Force (DMTF)’s Open Virtualizati<strong>on</strong> Format (OVF) is used. The challenge<br />
with OVF in RESERVOIR is how to adhere to the basic st<strong>and</strong>ard, widely used am<strong>on</strong>g<br />
industry but without the advanced features in RESERVOIR (elasticity, deployment-time<br />
c<strong>on</strong>figurati<strong>on</strong>, deployment c<strong>on</strong>straints, etc), <strong>and</strong> at the same time how to introduce these<br />
features without breaking it. The key to achieving this goal is using OVF built-in extensibility.<br />
Apart from OVF, st<strong>and</strong>ard APIs are needed to allow the interacti<strong>on</strong> between users <strong>and</strong> the<br />
RESERVOIR cloud. In this area, we have found a lot of fragmentati<strong>on</strong>, due to each alternative<br />
in the IaaS management API l<strong>and</strong>scape being actually a vendor-specific API rather than a<br />
st<strong>and</strong>ard <strong>on</strong>e. However, some emerging ef<strong>for</strong>ts are being taken to define a truly st<strong>and</strong>ard<br />
IaaS management API <strong>and</strong> <strong>on</strong>e of the most outst<strong>and</strong>ing <strong>on</strong>es is the work carried out in the
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
DMTF’s <strong>Cloud</strong> Management WG. In the RESERVOIR project, T<strong>Cloud</strong> API has been defined<br />
<strong>and</strong> used as IaaS management API <strong>and</strong>, in order to get a close alignment with the final DMTF<br />
st<strong>and</strong>ard, we submitted this proposal to DMTF <strong>and</strong> actively participate in CMWG work.<br />
Regarding interoperability between RESERVOIR middleware layers, st<strong>and</strong>ard alternatives<br />
are also being explored <strong>and</strong> used: T<strong>Cloud</strong> API (being the “intra-layer” functi<strong>on</strong>ality a<br />
subset of the API exposed to cloud users) <strong>and</strong> libvirt. Once the DMTF’s CMWG API<br />
c<strong>on</strong>solidates, interoperability tests could be d<strong>on</strong>e between RESERVOIR <strong>and</strong> future vendors’<br />
implementati<strong>on</strong>s.<br />
Possible future cooperati<strong>on</strong><br />
The st<strong>and</strong>ards c<strong>on</strong>solidated in RESERVOIR (OVF <strong>and</strong> T<strong>Cloud</strong> API) will c<strong>on</strong>tinue its evoluti<strong>on</strong><br />
in other cloud-related projects participated by the same partners (such as FP7 4CaaST, FP7<br />
VISION or Spanish funded NUBA) <strong>and</strong> in the products developed by the industrial partners<br />
in those c<strong>on</strong>sortia.<br />
C<strong>on</strong>tact: Fermín Galán Márquez<br />
Organisati<strong>on</strong>: Telefónica I+D<br />
C<strong>on</strong>tact details: fermin@tid.es<br />
Web: www.reservoir-fp7.eu/<br />
30
T<strong>Cloud</strong>s – Trustworthy <strong>Cloud</strong> Computing<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
The T<strong>Cloud</strong>s project investigates two use cases:<br />
1. The Smart <strong>Grid</strong> Use Case<br />
This case is based <strong>on</strong> a smart grid applicati<strong>on</strong> that has been developed jointly by Portugal’s<br />
main energy provider EDP (www.edp.pt) <strong>and</strong> the engineering company EFACEC (www.<br />
efacec.pt). The applicati<strong>on</strong> is in a pre-commercial stage <strong>and</strong> is currently piloted with<br />
public agencies. A central element is the real-time data generati<strong>on</strong>, intelligent analysis<br />
<strong>and</strong> smart c<strong>on</strong>trol of public lightning.<br />
2. The eHealth Use Case<br />
This case is based <strong>on</strong> a patient m<strong>on</strong>itoring, medical data analysis <strong>and</strong> remote diagnosis<br />
applicati<strong>on</strong> that is being developed by Philips (www.healthcare.philips.com) <strong>and</strong> the<br />
St. Raffaele Hospital (www.sanraffaele.org) in Milan. The applicati<strong>on</strong> is in the research<br />
<strong>and</strong> development stage. Central requirements are differentiated data access according<br />
to roles such as patient, doctor, pharmacist or patient family members. Also, strict<br />
regulatory requirements need to be observed in order to protect the privacy of the<br />
treated in<strong>for</strong>mati<strong>on</strong>.<br />
T<strong>Cloud</strong>s investigates the migrati<strong>on</strong> of central elements of these applicati<strong>on</strong>s into an<br />
IaaS cloud envir<strong>on</strong>ment – in particular the scalable operati<strong>on</strong>al data storage as well as<br />
per<strong>for</strong>mance critical run-time comp<strong>on</strong>ents. In both cases specific regulatory c<strong>on</strong>diti<strong>on</strong>s<br />
apply that are derived from EU as well as nati<strong>on</strong>al law. Both cases also imply specific<br />
requirements <strong>for</strong> security <strong>and</strong> need to protect the applicati<strong>on</strong> from external as well as<br />
insider attacks from cloud provider maintenance pers<strong>on</strong>nel.<br />
T<strong>Cloud</strong>s is specifically investigating the migrati<strong>on</strong> into a cloud-of-clouds envir<strong>on</strong>ment<br />
that is composed by multiple federated IaaS providers. For this reas<strong>on</strong>, T<strong>Cloud</strong>s will set-up<br />
several test-sites as well as use commercial IaaS providers.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
31<br />
The importance of interoperability<br />
T<strong>Cloud</strong>s is to <strong>on</strong>e extent researching <strong>on</strong> technologies that can provide external security<br />
<strong>and</strong> privacy to any IaaS cloud – such as allowing computati<strong>on</strong> with encrypted data in the<br />
cloud or the automated integrity verificati<strong>on</strong> of results received from software comp<strong>on</strong>ents<br />
deployed in a cloud.<br />
However, complementary mechanisms that T<strong>Cloud</strong>s is developing will also involve<br />
interfaces <strong>and</strong> interacti<strong>on</strong> with the IaaS providers <strong>on</strong> the deployment <strong>and</strong> en<strong>for</strong>cement of<br />
security <strong>and</strong> privacy policies. This relates to the IaaS service management interface level as<br />
well as to the st<strong>and</strong>ards <strong>for</strong> deployment descripti<strong>on</strong>s <strong>and</strong> m<strong>on</strong>itoring.
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
T<strong>Cloud</strong>s is investigating two Open Source cloud plat<strong>for</strong>ms: OpenStack (www.openstack.org)<br />
<strong>and</strong> Open Nebula (opennebula.org).<br />
Tclouds also envisages the adopti<strong>on</strong> <strong>and</strong> extensi<strong>on</strong> of Open cloud st<strong>and</strong>ards. Currently, the<br />
following are examples <strong>for</strong> st<strong>and</strong>ards that are c<strong>on</strong>sidered:<br />
»»<br />
The DMTF Open Virtualizati<strong>on</strong> Format (OVF)<br />
»»<br />
The OGF Open <strong>Cloud</strong> Computing Interface (OCCI)<br />
»»<br />
The SNIA <strong>Cloud</strong> Data Management Interface (CDMI)<br />
»»<br />
The NIST <strong>Cloud</strong> St<strong>and</strong>ards <str<strong>on</strong>g>Roadmap</str<strong>on</strong>g> – e.g. SCAP / Security C<strong>on</strong>tent Automati<strong>on</strong> Protocol<br />
»»<br />
Existing security st<strong>and</strong>ards – such as <strong>for</strong> identity <strong>and</strong> access management, encrypti<strong>on</strong> <strong>and</strong><br />
key management<br />
Possible future cooperati<strong>on</strong><br />
TCLouds is collaborating with the following initiatives:<br />
»»<br />
Effectsplus – Networking of EU Security Projects<br />
»»<br />
FIA - <str<strong>on</strong>g>European</str<strong>on</strong>g> Future Internet Assembly<br />
»»<br />
NESSI – Networked <str<strong>on</strong>g>European</str<strong>on</strong>g> Software <strong>and</strong> Services ETP<br />
Relevant EU cloud projects (<strong>on</strong>ly first indicati<strong>on</strong>s):<br />
»»<br />
RESERVOIR (federated IaaS clouds)<br />
»»<br />
VISION (federated cloud storage)<br />
»»<br />
SAIL (cloud networking)<br />
C<strong>on</strong>tacts: Elmar Husmann Matthias Schunter<br />
Organisati<strong>on</strong>: IBM Strategy & Change - Innovati<strong>on</strong> IBM Research – Zurich<br />
C<strong>on</strong>tact details: huselmar@de.ibm.com mts@zurich.ibm.com<br />
Web : www.tclouds-project.eu<br />
32
<str<strong>on</strong>g>European</str<strong>on</strong>g> Distributed Computing Infrastructures<br />
EDGI, DEGISCO & IDGF<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
The EDGI (<str<strong>on</strong>g>European</str<strong>on</strong>g> Desktop <strong>Grid</strong> Initiative) <strong>and</strong> DEGISCO (Desktop <strong>Grid</strong>s <strong>for</strong> Internati<strong>on</strong>al<br />
Scientific Collaborati<strong>on</strong>) <str<strong>on</strong>g>European</str<strong>on</strong>g> projects, together with IDGF (Internati<strong>on</strong>al Desktop<br />
<strong>Grid</strong> Federati<strong>on</strong>), are exp<strong>and</strong>ing the power of e<strong>Science</strong> infrastructures such as EGI with<br />
Desktop resources (which are numerous <strong>and</strong> cheap) <strong>and</strong> <strong>Cloud</strong> resources (which provide<br />
Quality of Service) in full producti<strong>on</strong>.<br />
On the e-Infrastructures side, we interface with the computing element by presenting the<br />
collected Desktop resources as just another Batch System. On the Desktop <strong>Grid</strong> side, we<br />
interface with the Desktop <strong>Grid</strong> server by submitting jobs to it. We interface with <strong>Cloud</strong>s<br />
by using their API.<br />
Our ‘Applicati<strong>on</strong> Repository’ middleware publishes applicati<strong>on</strong>s from government, industry<br />
or academia which have been adapted <strong>and</strong> validated <strong>for</strong> secure executi<strong>on</strong> <strong>on</strong> Desktop<br />
resources.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The importance of interoperability<br />
Our projects are needed because of the current lack of interoperability between the various<br />
middleware stacks <strong>for</strong> <strong>Grid</strong>s, Desktop <strong>Grid</strong>s <strong>and</strong> <strong>Cloud</strong>s. In fact, we are providing practical<br />
interoperati<strong>on</strong> through our bridge, using ad-hoc adapters, c<strong>on</strong>verters <strong>and</strong> translators <strong>for</strong><br />
each c<strong>on</strong>nected <strong>Grid</strong> or <strong>Cloud</strong> middleware.<br />
Our work would be eased very much by comm<strong>on</strong> <strong>Grid</strong>/<strong>Cloud</strong> open st<strong>and</strong>ards which are not<br />
<strong>on</strong>ly published, but widely implemented in a really interoperable manner. We present here<br />
the relevant st<strong>and</strong>ardizati<strong>on</strong> domains by decreasing level of importance.<br />
33<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
We are currently using many of the following de facto <strong>and</strong> official st<strong>and</strong>ards <strong>and</strong> we plan to<br />
use more of them in the future:<br />
» » In<strong>for</strong>mati<strong>on</strong> publicati<strong>on</strong> <strong>and</strong> discovery is st<strong>and</strong>ardized by OGF GLUE 2.0.<br />
» » Security is covered by IGFT, RFC-3820 compliant X509 proxies, OGF VOMS, Oasis SAML<br />
<strong>and</strong> EGI SPG.<br />
» » Log records will be st<strong>and</strong>ardized by OGF Activity Instance Document Schema.<br />
» » Accounting records are st<strong>and</strong>ardized by OGF Usage Record.<br />
» » M<strong>on</strong>itoring may be per<strong>for</strong>med using the WLCG Nagios stack.<br />
» » Data management is st<strong>and</strong>ardized by OGF DFDL, OGF ByteIO, <strong>Grid</strong>FTP, SRM, DMI <strong>and</strong><br />
SNIA CDMI; Virtual image <strong>for</strong>mat <strong>and</strong> definiti<strong>on</strong> by DMTF OVF.
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
»»<br />
VM instantiati<strong>on</strong> <strong>and</strong> management by OGF OCCI.<br />
»»<br />
Job descripti<strong>on</strong> language; by OGF JSDL; Job management protocol i by OGF BES <strong>and</strong> HPC<br />
Basic Profile.<br />
Possible future cooperati<strong>on</strong><br />
IDGF <strong>and</strong> EDGI/DEGISCO are working in str<strong>on</strong>g collaborati<strong>on</strong> with EGI, EMI, Nordu<strong>Grid</strong>,<br />
UNICORE Forum <strong>and</strong> interested NGIs in order to reach the widest possible user <strong>and</strong><br />
resource provider communities.<br />
IDGF is organising desktop grid operators <strong>and</strong> applicati<strong>on</strong> developers. St<strong>and</strong>ardizati<strong>on</strong><br />
activities are carried out mainly inside OGF. EDGI is carefully following any improvements<br />
<strong>and</strong> further developments of ARC, gLite <strong>and</strong> UNICORE maintained by EMI in order to make<br />
sure that the Service <strong>Grid</strong>s to Desktop <strong>Grid</strong>s bridge middleware developed by EDGI will be<br />
compatible with any new versi<strong>on</strong>s of the ARC, gLite, UNICORE <strong>and</strong> UMD middleware stacks.<br />
IDGF <strong>and</strong> EDGI/DEGISCO will explore the integrati<strong>on</strong> in future eInfrastructures. This means<br />
possible collaborati<strong>on</strong>s with <strong>Cloud</strong> research projects such as C<strong>on</strong>trail <strong>and</strong> mOSAIC. And it<br />
will look at extending virtualizati<strong>on</strong> techniques to the Desktop <strong>Grid</strong> client.<br />
C<strong>on</strong>tact: Etienne Urbah<br />
Organisati<strong>on</strong>: LAL, Univ Paris-Sud<br />
C<strong>on</strong>tact details: urbah@lal.in2p3.fr<br />
Web: edgi-project.eu<br />
Relevant Links: desktopgridfederati<strong>on</strong>.eu<br />
34
EGI - <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Grid</strong> Infrastructure<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
EGI provides an e-infrastructure to support the data analysis <strong>and</strong> computati<strong>on</strong>al needs<br />
of its publicly funded <strong>and</strong> supported end-users from the research community within<br />
Europe. Increasingly, this community has experimented with the interfaces provided from<br />
commercial cloud providers (IaaS, PaaS & SaaS) <strong>and</strong> would like to experience similar ease of<br />
use <strong>and</strong> flexibility, but with the efficiency, data transfer rates, c<strong>on</strong>trol <strong>and</strong> cost (free at the<br />
point of use) that they have experienced within publicly funded e-Infrastructure.<br />
The main users of such an envir<strong>on</strong>ment are not <strong>for</strong>eseen to (directly) be end-users.<br />
Rather they will be experts associated with the Virtual Research Community (or Virtual<br />
Organisati<strong>on</strong>) that will manage the preparati<strong>on</strong>, deployment <strong>and</strong> operati<strong>on</strong> of the virtual<br />
machines. These experts will come either from within the community or within an NGI<br />
working <strong>on</strong> behalf of that community. These experts would decide <strong>on</strong> behalf of their<br />
community the distributi<strong>on</strong> of the services at the resource centres that they have access to,<br />
when to deploy new software updates, <strong>and</strong> even the software that they would use.<br />
Essential to this model is to federate the virtual resources located at the resource<br />
infrastructure providers (the <str<strong>on</strong>g>European</str<strong>on</strong>g> NGIs <strong>and</strong> EIROs within EGI) to provide:<br />
»»<br />
Authenticati<strong>on</strong> <strong>and</strong> authorizati<strong>on</strong> model that permits the access to virtual machine<br />
management functi<strong>on</strong>s (deploy, start, stop, inspect, etc.) located at sites in different<br />
administrative domains<br />
»»<br />
Provisi<strong>on</strong>ing <strong>and</strong> maintenance of virtualized resources driven by locality to existing data<br />
sources, data sinks, or high per<strong>for</strong>mance networking links<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The importance of interoperability<br />
Interoperability is essential to a federated virtualised infrastructure. Each resource centre<br />
(site) will wish to make its own decisi<strong>on</strong> as to the underlying virtual machine management<br />
system it uses. This capability will need to be exposed in a systematic <strong>and</strong> c<strong>on</strong>sistent way<br />
to a distributed user group which will need to access many such centres. St<strong>and</strong>ards such as<br />
OCCI <strong>and</strong> other IaaS activity are essential <strong>for</strong> this usage model.<br />
Likewise, coordinati<strong>on</strong> is a key aspect of any federated model. For a virtualised federated<br />
infrastructure, the ability to manage c<strong>on</strong>sistent access to these resources dem<strong>and</strong>s a<br />
comm<strong>on</strong> security model that scales with regards to authenticati<strong>on</strong> <strong>and</strong> authorizati<strong>on</strong>.<br />
The X.509 related technology coupled to virtual organizati<strong>on</strong> model has shown to work<br />
technically at this scale, <strong>and</strong> if its primary use is to govern access to the virtual machine<br />
management functi<strong>on</strong>s (as opposed to access to the services run inside the virtual machine)<br />
it provides a st<strong>and</strong>ards based soluti<strong>on</strong>.<br />
A key aspect of federati<strong>on</strong> is resource discovery <strong>and</strong> to report <strong>on</strong> its usage. St<strong>and</strong>ards such as<br />
GLUE2 are being used within EGI to describe resources <strong>and</strong> derivatives of the Usage Record<br />
35
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
specificati<strong>on</strong> are used to aggregate accounting records <strong>on</strong> a <str<strong>on</strong>g>European</str<strong>on</strong>g> wide basis. Much of<br />
this in<strong>for</strong>mati<strong>on</strong> flow is now being supported by messaging technologies implemented the<br />
JMS specificati<strong>on</strong>.<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
Many of the emerging st<strong>and</strong>ards/specificati<strong>on</strong>s menti<strong>on</strong>ed previously (GLUE2, Usage<br />
Record, OCCI, JMS, X.509, etc.) have multiple servers or clients <strong>and</strong> are frequently sourced<br />
from communities bey<strong>on</strong>d EGI. This not <strong>on</strong>ly gives us technical c<strong>on</strong>fidence in adopting<br />
the technologies (they are proven to work in many other areas) but gives us adopti<strong>on</strong><br />
c<strong>on</strong>fidence as there are multiple providers that already need <strong>and</strong> know that their work<br />
needs to inter-operate.<br />
Any deployment of new technology releases will go through staged rollout be<strong>for</strong>e widescale<br />
producti<strong>on</strong> deployment to ensure that the interoperability is actually achieved between<br />
the critical comp<strong>on</strong>ents where it is needed. However, having to do explicit interoperability<br />
tests with different technologies would dem<strong>on</strong>strate low c<strong>on</strong>fidence in the technical<br />
provider… <strong>and</strong> these would not be <strong>on</strong>es we would chose to work with.<br />
Possible future cooperati<strong>on</strong><br />
36<br />
The technologies emerging out of the <str<strong>on</strong>g>European</str<strong>on</strong>g> Middleware Initiative, StratusLab, Initiative<br />
<strong>for</strong> Globus in Europe could all c<strong>on</strong>tribute to this activity. The C<strong>on</strong>trail project is exploring<br />
the issue as to how different resource sites can c<strong>on</strong>tribute to a cloud as an infrastructure,<br />
as opposed to individual sites.<br />
A missing capability in the open-source area seems to be the provisi<strong>on</strong>ing aspect across<br />
multiple cloud providers. Dealing with the negotiati<strong>on</strong> of resources from each provider to<br />
match the high-level deployment plan coming from the requesting user seems to be a gap.<br />
Likewise, linking a local virtualised network topology to existing high-speed networking<br />
links between virtualised resources does not seem to have an integrated soluti<strong>on</strong> at the<br />
moment.<br />
C<strong>on</strong>tact: Steven Newhouse<br />
Organisati<strong>on</strong>: <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Grid</strong> Initiative<br />
C<strong>on</strong>tact details: steven.newhouse@egi.eu<br />
Web: www.egi.eu<br />
Relevant Links: Integrati<strong>on</strong> of <strong>Cloud</strong>s <strong>and</strong> Virtualisati<strong>on</strong> into the <str<strong>on</strong>g>European</str<strong>on</strong>g> producti<strong>on</strong><br />
infrastructure – go.egi.eu/258
EMI - <str<strong>on</strong>g>European</str<strong>on</strong>g> Middleware Initiative<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
As being primarily a ‘research middleware provider’, <str<strong>on</strong>g>European</str<strong>on</strong>g> Middleware Intiative (EMI)<br />
use cases, in the c<strong>on</strong>text of e-infrastructures, are driven by ‘complex distributed highlevel<br />
scientific workflows’ that partly span over different types of e-Infrastructures.<br />
These require the transparent access to different types of heterogeneous computati<strong>on</strong>al<br />
resources (i.e. HPC <strong>and</strong> HTC) as well as per<strong>for</strong>ming storage management <strong>and</strong> necessary<br />
data transfers between resources. Here different computati<strong>on</strong>al paradigms such as<br />
HPC <strong>and</strong> HTC are needed in order to support comm<strong>on</strong> scientific community accepted<br />
different low-level applicati<strong>on</strong> programming models (i.e. OpenMP, MPI vs. task farming).<br />
This in turn points to requirements <strong>for</strong> comm<strong>on</strong> interfaces to computing resources,<br />
storage management, <strong>and</strong> the use of comm<strong>on</strong>ly agreed interfaces <strong>for</strong> data transfer<br />
adopted by middleware services that provide access to such resources. Related to this are<br />
challenging security requirements such as enabling single-sign <strong>on</strong> across e-Infrastructure<br />
boundaries or even per<strong>for</strong>ming work <strong>on</strong> behalf of another identity than the initial<br />
middleware user itself (i.e. delegati<strong>on</strong> of rights). Although many security models (PKI,<br />
SLC-services, OpenID, etc.) <strong>and</strong> interfaces/st<strong>and</strong>ards (X.509, SAML, etc.) exist, they<br />
have been not c<strong>on</strong>sistently adopted across technology providers. More recently, cloud<br />
computing is emerging using virtualizati<strong>on</strong> technologies that <strong>for</strong>m a dynamic kind of ‘<strong>on</strong>dem<strong>and</strong><br />
e-Infrastructure’. EMI explores soluti<strong>on</strong>s to enable middleware services to take<br />
advantage of such emerging virtualized infrastructures. In this c<strong>on</strong>text, we c<strong>on</strong>sider two<br />
opti<strong>on</strong>s. EMI services that are part of virtual machine appliances <strong>and</strong> the seamless access<br />
to existing cloud infrastructures from already established <strong>and</strong> broadly used middleware<br />
services/clients.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The importance of interoperability<br />
The requirement <strong>for</strong> interoperability between existing middleware services that are<br />
deployed as part of virtual appliances is relatively well supported by available st<strong>and</strong>ards<br />
in the field that EMI is comm<strong>on</strong>ly adopting during the course of the project (i.e. compute,<br />
data, in<strong>for</strong>mati<strong>on</strong>, security area, etc.). However, end-users typically require interoperability<br />
to take advantage of middleware services with unique capabilities that specifically offer<br />
access to HPC, HTC, or storage resources across all different kinds of e-Infrastructures<br />
(e.g. PRACE, EGI, clouds). While HPC-based clouds are rather rare, we mostly experience<br />
interoperability requirements <strong>for</strong> middleware to use it seamlessly with already existing<br />
cloud-based infrastructures (<strong>and</strong> their access <strong>and</strong> management interfaces) offering HTC<br />
resources <strong>and</strong> dynamic storage capabilities. EMI will work towards the interoperability<br />
with implementati<strong>on</strong>s providing emerging st<strong>and</strong>ards-based interfaces to existing cloud<br />
infrastructures, with a particular focus <strong>on</strong> the access of computing <strong>and</strong> data resources.<br />
37
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
Scientific end-users already take advantage of comm<strong>on</strong>ly used middleware (client) tools<br />
today which require the seamless access to these infrastructures by having interoperability<br />
in the areas of security, job <strong>and</strong> data management, as well as accounting.<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
Several agreed st<strong>and</strong>ard interfaces/schemas <strong>for</strong> the interoperability between established<br />
middleware technologies are adopted <strong>and</strong> c<strong>on</strong>tinuously tested <strong>for</strong> compliance during the<br />
course of the EMI project (e.g. SRM, GLUE2, etc.). Nevertheless, from a ‘client perspective’,<br />
several middleware services are expected to be compliant with emerging st<strong>and</strong>ard<br />
interfaces of cloud-based infrastructures. At the time of writing, there is currently <strong>on</strong>e<br />
emerging st<strong>and</strong>ard named as Open <strong>Cloud</strong> Computing Interface (OCCI) that might be<br />
relevant <strong>for</strong> EMI when it offers functi<strong>on</strong>ality <strong>on</strong> the PaaS <strong>and</strong> SaaS-level rather than <strong>on</strong> the<br />
IaaS-level as today. In terms of storage, the st<strong>and</strong>ard <strong>Cloud</strong> Data Management Interface<br />
(CDMI) seems to be a promising st<strong>and</strong>ard to be adopted by EMI services as well while the<br />
st<strong>and</strong>ard still needs to prove its relevance in industry. In both cases, EMI has to be aware<br />
of the dynamics of virtual resources <strong>and</strong> at the same time make good use of them ideally<br />
through the adopti<strong>on</strong> of comm<strong>on</strong>ly agreed st<strong>and</strong>ard interfaces.<br />
Possible future cooperati<strong>on</strong><br />
»»<br />
StratusLab (Providing EMI middleware-based virtual machine appliances)<br />
»»<br />
VENUS-C (EMI clients might benefit via similar st<strong>and</strong>ard interfaces based <strong>on</strong> BES/JSDL)<br />
C<strong>on</strong>tact: Morris Riedel<br />
Organisati<strong>on</strong>: Jülich Supercomputing Centre<br />
C<strong>on</strong>tact details: m.riedel@fz-juelich.de<br />
Web: www.eu-emi.eu/en<br />
38
IGE - Initiative <strong>for</strong> Globus in Europe<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
IGE targets, as a base middleware provider, various fields of applicati<strong>on</strong>s <strong>and</strong> does not limit<br />
itself to a certain community. However, a str<strong>on</strong>g focus lies <strong>on</strong> helping scientists in their daily<br />
work, making the use of eInfrastructure as simple <strong>and</strong> seamless as possible while not trying<br />
to cover specific issues, but rather cover general services. The two general use cases IGE has<br />
collected from the user communities, <strong>and</strong> which are seen as the most important, are “<strong>Grid</strong><br />
<strong>on</strong> top of <strong>Cloud</strong>” <strong>and</strong> “<strong>Cloud</strong> <strong>on</strong> top of <strong>Grid</strong>”.<br />
While the “<strong>Grid</strong> <strong>on</strong> top of <strong>Cloud</strong>” use case covers the exercise of running <strong>Grid</strong> middleware<br />
services in an IaaS envir<strong>on</strong>ment <strong>and</strong> is basically solved by technology providers from various<br />
directi<strong>on</strong>s (the EGI roadmap, commercial IaaS vendors, open-source projects, infrastructure<br />
st<strong>and</strong>ardizati<strong>on</strong> ef<strong>for</strong>ts, etc.), it still requires significant automati<strong>on</strong> ef<strong>for</strong>ts to bring benefit<br />
to the operators of such services.<br />
The “<strong>Cloud</strong> <strong>on</strong> top of <strong>Grid</strong>” use case, in turn, requires an entirely new set of interfaces, which<br />
are yet to be defined. For example, the typical IaaS model of managing virtual machines needs<br />
to be mapped to current <strong>Grid</strong> middleware envir<strong>on</strong>ments. A starting point <strong>for</strong> this is the Globus<br />
Online ef<strong>for</strong>t, which is an integral part of the project <strong>for</strong> the <str<strong>on</strong>g>European</str<strong>on</strong>g> Research community.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The importance of interoperability<br />
For the “<strong>Grid</strong> <strong>on</strong> top of <strong>Cloud</strong>” use case, interoperability is a key issue: the deployment of <strong>Grid</strong><br />
services should work as seamless as possible <strong>for</strong> the operators, even cross-infrastructure. As<br />
such, comm<strong>on</strong> interfaces to the underlying infrastructure are crucial <strong>and</strong> should be available<br />
as broadly as possible. One c<strong>and</strong>idate <strong>for</strong> this process would be OCCI, but the area of “service<br />
templates” <strong>and</strong> deployment automati<strong>on</strong>, also with respect to instance-specific c<strong>on</strong>figurati<strong>on</strong><br />
<strong>and</strong> adaptati<strong>on</strong>, is yet to be resolved since no accepted st<strong>and</strong>ards are available here.<br />
For the “<strong>Cloud</strong> <strong>on</strong> top of <strong>Grid</strong>” use case, the capabilities as defined by the EGI roadmap are a<br />
starting point <strong>for</strong> possible st<strong>and</strong>ards. However, in this c<strong>on</strong>text, the applicati<strong>on</strong>s <strong>and</strong> plat<strong>for</strong>ms<br />
comprising the <strong>Cloud</strong> envir<strong>on</strong>ment highly influence the requirements <strong>for</strong> such st<strong>and</strong>ards. Here it<br />
would be necessary to collect <strong>Cloud</strong> applicati<strong>on</strong> use cases that are eligible to run <strong>on</strong> top of <strong>Grid</strong><br />
infrastructure <strong>and</strong> extract comm<strong>on</strong> requirements that need to be addressed by the DCI projects.<br />
39<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
At the moment, IGE evaluates the applicability of <strong>Cloud</strong> st<strong>and</strong>ards to the project goals. As<br />
said be<strong>for</strong>e, a good c<strong>and</strong>idate <strong>for</strong> the described use cases is the OCCI family of specificati<strong>on</strong>s.<br />
Interoperability tests c<strong>on</strong>ducted by IGE would largely c<strong>on</strong>sider using <strong>Cloud</strong> interfaces from<br />
the client perspective; as such, the project requirements are c<strong>on</strong>sumer-oriented regarding
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
IaaS services. From the provider perspective, the upcoming <str<strong>on</strong>g>European</str<strong>on</strong>g> deployment of a<br />
cloud-based file transfer service <strong>on</strong> top of <strong>Grid</strong> infrastructure, Globus Online, will show<br />
whether <strong>and</strong> how scalability is an issue, but is unlikely to touch interoperability issues <strong>on</strong><br />
the <strong>Cloud</strong> interface level.<br />
Possible future cooperati<strong>on</strong><br />
A main issue seems to lie in the field of usable templates in the c<strong>on</strong>text of virtualized<br />
services. Especially the post-template creati<strong>on</strong> aspects such as individual VM modificati<strong>on</strong><br />
(tailoring towards the VRC that is to be targeted) seems to be an open issue. While the EGI<br />
roadmap seems to touch this field, c<strong>on</strong>crete steps are yet to be defined.<br />
C<strong>on</strong>tact: Alex<strong>and</strong>er Papaspyrou<br />
Organisati<strong>on</strong>: Technische Universität Dortmund<br />
C<strong>on</strong>tact details: alex<strong>and</strong>er.papaspyrou@tu-dortmund.de; eglo@ige-project.eu<br />
Web: www.ige-project.eu<br />
40
StratusLab – Enhancing <strong>Grid</strong> Infrastructures<br />
with Virtualizati<strong>on</strong> <strong>and</strong> <strong>Cloud</strong> Technologies<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
The StratusLab project started in June 2010 with the purpose of investigating the impact<br />
of the emerging cloud computing paradigm in the provisi<strong>on</strong> of grid computing services.<br />
StratusLab focuses <strong>on</strong> the Infrastructure-as-a-Service (IaaS) cloud paradigm, which implies<br />
the usage of virtualizati<strong>on</strong> technologies <strong>for</strong> the provisi<strong>on</strong> of computing resources. The<br />
project is integrating a cloud distributi<strong>on</strong>, based <strong>on</strong> the OpenNebula cloud management<br />
toolkit, specifically designed with the purpose of hosting grid services. During the design<br />
phase the specific requirements <strong>and</strong>/or restricti<strong>on</strong>s of grid services are taken into account<br />
in order to provide optimized cloud envir<strong>on</strong>ments <strong>for</strong> deploying virtualized producti<strong>on</strong><br />
grid sites. The first versi<strong>on</strong> of the StratusLab distributi<strong>on</strong> was released in October 2010. The<br />
distributi<strong>on</strong> is used by the project itself to setup <strong>and</strong> provide a reference cloud service.<br />
Currently two capabilities are available to the public: a cloud IaaS service, giving users the<br />
ability to to instantiate <strong>and</strong> manage VMs <strong>and</strong> a appliance repository where the VM images<br />
are stored. This reference cloud service is used also internally by the project as a testbed<br />
<strong>for</strong> deploying grid sites <strong>and</strong> in order to investigate potential implicati<strong>on</strong>s of their operati<strong>on</strong><br />
over the cloud.<br />
The primary applicati<strong>on</strong> domains that the project is targeting are similar to those of grid<br />
computing, i.e. scientific applicati<strong>on</strong>s either in research or producti<strong>on</strong> phase. In particular<br />
the Bioin<strong>for</strong>matics group from CNRS/IBCP participates in the project offering the primary<br />
use cases <strong>for</strong> end-user applicati<strong>on</strong>s <strong>on</strong> the StratusLab infrastructure.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The importance of interoperability<br />
Interoperability plays an important role <strong>for</strong> StratusLab as with any large scale shared<br />
infrastructure envir<strong>on</strong>ment. Currently the main focus is <strong>on</strong> IaaS interfaces, access to virtual<br />
machine appliances <strong>and</strong> security. Another level of interoperability particularly important<br />
<strong>for</strong> StratusLab is the <strong>on</strong>e between grid middleware <strong>and</strong> cloud management service. In this<br />
level issues of accounting <strong>and</strong> m<strong>on</strong>itoring have been identified as a priority <strong>for</strong> investigati<strong>on</strong>.<br />
41<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
OpenNebula is in the core of StratusLab distributi<strong>on</strong> <strong>and</strong> has already adopted the OGF<br />
OCCI st<strong>and</strong>ard. The toolkit’s development team, which also participates in StratusLab, plays<br />
a central role in the st<strong>and</strong>ardizati<strong>on</strong> process of OCCI. Although OCCI support is currently<br />
not yet integrated in the StratusLab distributi<strong>on</strong>, it is scheduled <strong>for</strong> the upcoming releases<br />
of the project. For what c<strong>on</strong>cerns security <strong>and</strong> authenticati<strong>on</strong>, StratusLab has adopted X.509
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
certificates <strong>and</strong> utilizes VOMS services <strong>for</strong> VO management <strong>and</strong> end-user authenticati<strong>on</strong>.<br />
During the sec<strong>on</strong>d year of the project we plan to investigate hybrid cloud soluti<strong>on</strong>s <strong>and</strong><br />
exploitati<strong>on</strong> of commercial cloud infrastructures. In this case IaaS interoperability will<br />
become even more relevant <strong>and</strong> may re-focus the development <strong>and</strong> integrati<strong>on</strong> activities<br />
of the project.<br />
Possible future cooperati<strong>on</strong><br />
StratusLab keeps close c<strong>on</strong>tact with most of the DCI <str<strong>on</strong>g>European</str<strong>on</strong>g> projects currently under way.<br />
In particular the project is in close collaborati<strong>on</strong> with EGI-InSPIRE, EMI <strong>and</strong> EDGI projects.<br />
These collaborati<strong>on</strong>s are being <strong>for</strong>malized with respective MoUs. The project is also<br />
planning to collaborate with commercial cloud providers like ElasticHosts <strong>and</strong> Flexiscale in<br />
order to test the applicati<strong>on</strong> of the StratusLab distributi<strong>on</strong> in hybrid cloud envir<strong>on</strong>ments.<br />
C<strong>on</strong>tact: Vangelis Floros<br />
Organisati<strong>on</strong>: GRNET<br />
C<strong>on</strong>tact details: support@stratuslab.eu<br />
Web: www.stratuslab.eu<br />
42
VENUS-C – Virtual Multidisciplinary<br />
Envir<strong>on</strong>ments Using <strong>Cloud</strong> Infrastructures<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
The VENUS-C project is aimed at validating the use of cloud infrastructures to support<br />
research in seven user scenarios, plus around ten more applicati<strong>on</strong>s that will be identified<br />
through an open call. Current user scenarios include seven applicati<strong>on</strong>s across four<br />
thematic areas: civil engineering, marine biodiversity, civil protecti<strong>on</strong> <strong>and</strong> emergencies <strong>and</strong><br />
biomedicine. Specifically, applicati<strong>on</strong>s focus <strong>on</strong> 3D static <strong>and</strong> dynamic structural analysis<br />
(Universidad Politecnica de Valencia), building in<strong>for</strong>mati<strong>on</strong> management (Collaboratorio),<br />
marine biodiversity maps (Nati<strong>on</strong>al Research Council of Italy), wildfire risk predicti<strong>on</strong><br />
<strong>and</strong> fire propagati<strong>on</strong> simulati<strong>on</strong> (University of the Aegean), bioin<strong>for</strong>matics (Universidad<br />
Politecnica de Valencia), systems biology (Center <strong>for</strong> Computati<strong>on</strong>al <strong>and</strong> Systems Biology),<br />
<strong>and</strong> drug discovery (Newcastle University), covering a wide range of scientific use cases<br />
targeting <strong>on</strong> the use of intensive computing <strong>and</strong> data storage.<br />
<strong>Cloud</strong> infrastructures are envisaged as a way to access improved computing power bey<strong>on</strong>d<br />
users’ facilities (l<strong>on</strong>g-durati<strong>on</strong> earthquake simulati<strong>on</strong>s, the alignment of large-scale<br />
sequences with respect to public databases, drug discovery over large lig<strong>and</strong> databases,<br />
biological systems simulati<strong>on</strong>, <strong>and</strong> so <strong>on</strong>), by adapting computing kernels as worker roles<br />
or complete virtual appliances. These working units are orchestrated in a coordinated<br />
<strong>and</strong> reliable framework that ensures the effective executi<strong>on</strong> of the multiple parallel<br />
comp<strong>on</strong>ents. However, cloud infrastructures are also acting as enabling technologies<br />
providing computing resources <strong>for</strong> web applicati<strong>on</strong>s (as in the generati<strong>on</strong> of fire risk <strong>and</strong><br />
behavior maps, ad-hoc views of marine biodiversity maps or <strong>for</strong> rendering capabilities in<br />
building in<strong>for</strong>mati<strong>on</strong>).<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The importance of interoperability<br />
From the point of view of the user, interoperability can be understood as the ability to<br />
switch, choose or use several infrastructures simultaneously. The ability to seamlessly<br />
switch from local to external resources provided by cloud infrastructures c<strong>on</strong>stitutes an<br />
attractive usage model <strong>for</strong> research. Local resources could deal with test or planning work,<br />
whereas external producti<strong>on</strong>-quality resources can be used in large experiments. This could<br />
be the case <strong>for</strong> example of a phylogenetic annotati<strong>on</strong> or a drug discovery experiment or<br />
the dynamic simulati<strong>on</strong> of an earthquake <strong>on</strong> a building structure. There are limitati<strong>on</strong>s<br />
<strong>on</strong> binaries (which could be hidden by the use of virtual appliances) <strong>and</strong> per<strong>for</strong>mance<br />
restricti<strong>on</strong>s, but different infrastructures could even bring different opportunities (<strong>and</strong><br />
costs). Another important issue is the interoperability in data objects across infrastructures<br />
(as data science infrastructures holding public data <strong>and</strong> computing clouds dealing with it),<br />
which would also require “business” interoperability in the way costs could be charged.<br />
43
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
Al<strong>on</strong>g with the st<strong>and</strong>ards that will be adopted at the level of the VENUS-C executi<strong>on</strong> models,<br />
data access is an area in which user applicati<strong>on</strong>s could be impacted more by st<strong>and</strong>ards. In<br />
VENUS-C, the <strong>Cloud</strong> Data Management Interface (CDMI – Storage <strong>and</strong> Networking Industry<br />
Associati<strong>on</strong>) is being adopted to provide a st<strong>and</strong>ard access to local <strong>and</strong> remote data. CDMI<br />
will hide the particularities of the storage back-ends that will improve the interoperability<br />
of applicati<strong>on</strong>s when accessing data in different plat<strong>for</strong>ms. However, there is a c<strong>on</strong>cern<br />
about the effect <strong>on</strong> per<strong>for</strong>mance that could have adopting CDMI as a neutral interface <strong>for</strong><br />
accessing data. There<strong>for</strong>e, large-scale tests will be per<strong>for</strong>med using both CDMI neutral <strong>and</strong><br />
plat<strong>for</strong>m specific data access interfaces, looking <strong>for</strong> trade-offs between per<strong>for</strong>mance <strong>and</strong><br />
interoperability.<br />
Possible future cooperati<strong>on</strong><br />
CDMI, which defines the functi<strong>on</strong>al interface that applicati<strong>on</strong>s will use to create, retrieve,<br />
update <strong>and</strong> delete data elements from the <strong>Cloud</strong>. As part of this interface the client will<br />
be able to discover the capabilities of the cloud storage offering <strong>and</strong> use this interface to<br />
manage c<strong>on</strong>tainers <strong>and</strong> the data that is placed in them. In additi<strong>on</strong>, metadata can be set <strong>on</strong><br />
c<strong>on</strong>tainers <strong>and</strong> their c<strong>on</strong>tained data elements through this interface.<br />
This interface is also used by administrative <strong>and</strong> management applicati<strong>on</strong>s to manage<br />
c<strong>on</strong>tainers, accounts, security access <strong>and</strong> m<strong>on</strong>itoring/billing in<strong>for</strong>mati<strong>on</strong>, even <strong>for</strong> storage<br />
that is accessible by other protocols. The capabilities of the underlying storage <strong>and</strong> data<br />
services are exposed so that clients can underst<strong>and</strong> the offering.<br />
C<strong>on</strong>tact: Ignacio Blanquer<br />
Organisati<strong>on</strong>: Universidad Politecnica de Valencia<br />
C<strong>on</strong>tact details: iblanque@dsic.upv.es<br />
Web: www.venus-c.eu<br />
44
Business & Government<br />
The shift to cloud computing in<br />
government in the EU<br />
Government is <strong>on</strong>e of the biggest sectors <strong>for</strong> ICT spending in the EU. The factors that are<br />
driving government to use cloud are a little different from the business/private sector. The<br />
business, <strong>and</strong> particularly small business is leading the shift to cloud computing, primarily<br />
because cloud offers companies increased flexibility in their use of computing resources.<br />
This enables companies to be more efficient <strong>and</strong> operate more effectively. <strong>Cloud</strong> also has<br />
advantages over traditi<strong>on</strong>al computer deployment such as desktop, in allowing customers<br />
to save capital expenditure (switching to opex), <strong>and</strong> save property, labour <strong>and</strong> other indirect<br />
costs associated with owning <strong>and</strong> operating a traditi<strong>on</strong>al computer estate. A third party<br />
running a major data warehouse or cloud computing facility <strong>and</strong> upgrading software more<br />
regularly also has the ability to offer latest generati<strong>on</strong> of products/technology <strong>on</strong> a faster<br />
<strong>and</strong> more regular basis <strong>and</strong> can literally offer more <strong>for</strong> less given scale ec<strong>on</strong>omies.<br />
Government is less driven by operati<strong>on</strong>al efficiency, <strong>and</strong> more by the major drives to reduce cost<br />
<strong>and</strong> save m<strong>on</strong>ey to pay off debts <strong>and</strong> reduce government deficits. Programmatic change, such as<br />
government seeking ‘buy <strong>on</strong>ce’ benefits, rather than buying <strong>on</strong> a silo-ed <strong>and</strong> departmental basis<br />
<strong>and</strong> government looking to leverage its buying power, are focusing government ef<strong>for</strong>ts <strong>on</strong> cost,<br />
<strong>and</strong> cost of processing is leading purchasing managers to look at cloud computing because of its<br />
significant cost savings over traditi<strong>on</strong>al desktop soluti<strong>on</strong>s. Many governments have announced,<br />
<strong>and</strong> are in the throws of implementing, service oriented architectures (SOA), that are intended<br />
to create a technology plat<strong>for</strong>m in government enabling applicati<strong>on</strong>s to be bought ‘off the shelf’<br />
<strong>and</strong> added more quickly <strong>and</strong> cost effectively to the government’s ‘app’s store’. Virtualisati<strong>on</strong> <strong>and</strong><br />
st<strong>and</strong>ards may help, but an increasing issue is the dependence <strong>on</strong> the technology of particular<br />
vendors that are needed in getting an SOA to work, or inter-operate, with others’ technology.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
UK government: driven by cost savings<br />
In the UK, the government is pursuing a cloud computing strategy, <strong>and</strong> looking to c<strong>on</strong>solidate<br />
central government computing from 200+ data centers to about 10. This will inevitably<br />
involve virtualisati<strong>on</strong>, <strong>and</strong> may involve public cloud as well as the c<strong>on</strong>solidati<strong>on</strong> <strong>on</strong>to a<br />
private cloud plat<strong>for</strong>m. In line with the private sector, government is c<strong>on</strong>cerned about data<br />
protecti<strong>on</strong> <strong>and</strong> data security, <strong>and</strong> it is hoped that the move from a physical to a more virtual<br />
envir<strong>on</strong>ment may increase the security of data (the UK has had a history of loss of data <strong>on</strong><br />
physical items such as sticks <strong>and</strong> disks <strong>and</strong> computers, <strong>and</strong> the promise of remote storage is<br />
expected to reduce the risks of physical data loss).<br />
45
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
Legal issues<br />
The legal issues facing government <strong>and</strong> government agencies include data protecti<strong>on</strong> <strong>and</strong><br />
the need to produce data <strong>for</strong> law en<strong>for</strong>cement. These are the same issues facing the private<br />
sector more generally. These are in the process of being worked through. Technology lock-in<br />
<strong>and</strong> <strong>for</strong>wards <strong>and</strong> backwards compatibility of existing technology with cloud deployment is<br />
a major additi<strong>on</strong>al issue <strong>for</strong> government: the shift to SOA means that st<strong>and</strong>ards are needed<br />
against which purchases can be made <strong>and</strong> third parties need technical interfaces in order to<br />
run their technology with that which is already in the existing government estate. However,<br />
st<strong>and</strong>ards <strong>on</strong>ly solve the issue of technology lock-in where they exist, <strong>and</strong> they typically <strong>on</strong>ly<br />
exist when many players have already agreed to operate to a particular st<strong>and</strong>ard. Licensing of<br />
underlying intellectual property rights is often needed in such situati<strong>on</strong>s, <strong>and</strong> can be achieved<br />
<strong>on</strong> FRAND (fair reas<strong>on</strong>able <strong>and</strong> n<strong>on</strong> discriminatory, terms) which has been agreed in many cases.<br />
Also, in its previous approach, which allowed individual departments to purchase technology<br />
to meet specific needs of the department by the department, the UK government has already<br />
outsourced many of its functi<strong>on</strong>s to third parties. Now, looking at cross departmental cost<br />
savings <strong>and</strong> cross departmental technology soluti<strong>on</strong>s means addressing cross departmental<br />
needs <strong>and</strong> will cut across agreements with existing suppliers <strong>and</strong> the different technologies<br />
that have already been bought <strong>for</strong> particular departmental needs.<br />
Interoperability<br />
46<br />
Interoperability is the issue of the moment. This is the issue of how to make existing technology<br />
work with the latest generati<strong>on</strong>, often available from a cloud computing soluti<strong>on</strong>. Where<br />
situati<strong>on</strong>s of dependency <strong>on</strong> a single player’s technology arise, such as with interoperability with<br />
the IBM mainframe, (often used <strong>for</strong> tax <strong>and</strong> benefits systems by government), then st<strong>and</strong>ards<br />
are unlikely to resolve the issue, <strong>and</strong> there is a need <strong>for</strong> anti-trust laws or regulati<strong>on</strong> to achieve<br />
interoperability. The <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong> is currently investigating the lock in between legacy<br />
programs running <strong>on</strong> proprietary IBM mainframes <strong>and</strong> is looking at the issue of interoperability<br />
between applicati<strong>on</strong>s <strong>and</strong> hardware <strong>and</strong> software in relati<strong>on</strong> to a complaint brought by an<br />
open source company Turbo Hercules. This complaint involves Turbo Hercules attempts to run<br />
customer data <strong>on</strong> other hardware <strong>and</strong> software outside the mainframe envir<strong>on</strong>ment, <strong>and</strong> the<br />
way that interoperability between the customer’s established technology others’ technology.<br />
The <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong> is also examining a range of ‘measures’ that could be adopted to deal<br />
with pervasive technologies under acti<strong>on</strong> point 25 of its Digital Agenda.<br />
These issues typically arise where technology has been provided by a supplier <strong>on</strong> a vertically<br />
integrated basis: hardware <strong>and</strong> software lock-in is a well known approach of technology<br />
companies. The issue may be thought of as bundling, of m<strong>on</strong>opoly <strong>and</strong> n<strong>on</strong> m<strong>on</strong>opoly<br />
comp<strong>on</strong>ents, whether of hardware <strong>and</strong> software or of a number of software products<br />
together. The Commissi<strong>on</strong> intervened in the Microsoft case (<strong>and</strong> in other cases such as IMS<br />
health), <strong>and</strong> has adopted remedies to ensure that interface in<strong>for</strong>mati<strong>on</strong> is published <strong>and</strong><br />
that applicati<strong>on</strong> programmers obtain the in<strong>for</strong>mati<strong>on</strong> needed so that their software can run<br />
with others’ <strong>and</strong> that applicati<strong>on</strong>s can run <strong>on</strong> other companies hardware.<br />
This is a current issue in the UK. The UK’s Cabinet Office has recently (end Jan 2011) announced<br />
its preference <strong>for</strong> n<strong>on</strong>-IPR based st<strong>and</strong>ards in government purchasing. However, it is a
mystery how such an approach would resolve these issues or even how such an approach<br />
is compatible with EU <strong>and</strong> WTO obligati<strong>on</strong>s to ensure that government purchasing is evenh<strong>and</strong>ed<br />
<strong>and</strong> technologically neutral.<br />
These issues are current <strong>and</strong> require resoluti<strong>on</strong>. Can they be left to look after themselves?<br />
Can a market soluti<strong>on</strong> solve the problem? Often this is the case with applicati<strong>on</strong>s at higher<br />
levels in the technology stack where customers can buy an alternative applicati<strong>on</strong> if <strong>on</strong>e<br />
does not work. Where a customer has bought technology <strong>and</strong> has become dependent up<strong>on</strong><br />
it interoperability may be the <strong>on</strong>ly soluti<strong>on</strong>. Clearly, if existing technology is owned by an<br />
existing supplier, use of that technology will often require compensati<strong>on</strong> <strong>and</strong> intellectual<br />
property right licenses may be needed.<br />
Interventi<strong>on</strong> may be needed so that the market is not held back <strong>and</strong> the government is not<br />
held to ransom. Unlike markets <strong>for</strong> apps, these dependency situati<strong>on</strong>s are not capable of<br />
being dealt with as matters at the higher levels in the technology stack where the market<br />
can be expected to operate freely, but are issues that arise where customers are dependent<br />
<strong>on</strong> technology or technology plat<strong>for</strong>ms <strong>and</strong> where suppliers have market power. There is<br />
clearly no issue of dependency <strong>and</strong> no issue of market power where no dependency exists,<br />
however, where there is market power <strong>and</strong> dependency, then there is a major need <strong>for</strong><br />
interoperability that requires real inter-working between existing <strong>and</strong> future technology.<br />
Apart from case by case investigati<strong>on</strong> by anti-trust authorities, the shift to cloud computing<br />
can be seen as a shift toward greater intelligence being included in communicati<strong>on</strong>s<br />
infrastructure: off-premises processing is truly dependent <strong>on</strong> communicati<strong>on</strong>s at a distance,<br />
<strong>and</strong> dependent <strong>on</strong> the interoperability <strong>and</strong> access to technical in<strong>for</strong>mati<strong>on</strong>. Some aspects<br />
of the existing telecommunicati<strong>on</strong>s infrastructure will need to be upgraded in order to be<br />
able to cope with the increased needs <strong>and</strong> dem<strong>and</strong>s of cloud computing soluti<strong>on</strong>s.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
Communicati<strong>on</strong>s regulati<strong>on</strong>?<br />
These issues are not new, <strong>and</strong> in order <strong>for</strong> the many different technologies that are used in<br />
a communicati<strong>on</strong>s system to talk to another communicati<strong>on</strong>s system regulati<strong>on</strong> has existed<br />
<strong>for</strong> many years to make sure that systems can interc<strong>on</strong>nect. The regime that exists <strong>and</strong><br />
governs the use of telecommunicati<strong>on</strong>s infrastructure addresses these issues <strong>and</strong> may apply<br />
to improve the c<strong>on</strong>diti<strong>on</strong>s of access <strong>and</strong> use of the telecoms infrastructure <strong>for</strong> the new<br />
phenomena of cloud computing. Regulati<strong>on</strong> of interoperability <strong>and</strong> access has been needed<br />
<strong>and</strong> applied to telecoms companies <strong>for</strong> many years, as the shift to a new era of computing<br />
takes place with more processing in the system needing to work with computing at the<br />
edge the system of laws are faced with familiar issues.<br />
Resoluti<strong>on</strong> is taking place now <strong>and</strong> a fair balance between the needs of rights holders <strong>and</strong><br />
the efficiency gains <strong>and</strong> cost benefits needed by government will mean that the system will<br />
require adaptati<strong>on</strong> by industry <strong>and</strong> regulatory/anti-trust authorities alike.<br />
47<br />
C<strong>on</strong>tact: Tim Cowen<br />
Organisati<strong>on</strong>: Sidley Austin & Open Computing Alliance<br />
C<strong>on</strong>tact details: tcowen@Sidley.com<br />
Web: www.opencomputingalliance.org
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
G-CLOUD – UK Government <strong>Cloud</strong><br />
Computing Infrastructure<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
The G-<strong>Cloud</strong> programme is a UK Government initiative to provide a Service orientated<br />
infrastructure <strong>for</strong> delivering services to the citizen <strong>and</strong> support <strong>for</strong> business processes<br />
across Government. While the l<strong>on</strong>ger term visi<strong>on</strong> <strong>for</strong> G-<strong>Cloud</strong> <strong>for</strong>esees widespread sharing<br />
of services <strong>and</strong> scaling of applicati<strong>on</strong>s to the public cloud with appropriate security, in<br />
the short term the challenge is to harvest small amounts of infrastructure distributed<br />
am<strong>on</strong>gst Government’s several hundred Data Centres, to provide support <strong>for</strong> virtualized<br />
applicati<strong>on</strong>s to scale within the Firewall. St<strong>and</strong>ards which allow the development of a shared<br />
infrastructure <strong>and</strong> classes of operati<strong>on</strong> <strong>for</strong> IaaS which would support different software<br />
types is essential. Some <strong>for</strong>m of shared middleware <strong>for</strong> scheduling <strong>and</strong> load balancing<br />
is also required. This should recognize c<strong>on</strong>cepts of class of software supported, locality<br />
<strong>and</strong> required c<strong>on</strong>figurati<strong>on</strong>s, including security impact level. While it is not envisaged that<br />
management of shared data will be implemented in this way, this may become a requirement<br />
downstream.<br />
The importance of interoperability<br />
48<br />
Ultimately, the design of G-<strong>Cloud</strong> applicati<strong>on</strong>s should not need to take account of<br />
hardware/software envir<strong>on</strong>ments in which to operate, whether owned or rented as a service.<br />
However, in the short term there will be a requirement to assure interoperability in order<br />
to take full advantage of available capacity across the UK estate <strong>and</strong> bey<strong>on</strong>d into public<br />
cloud service providers. The ability to integrate a number of classes of cloud infrastructure<br />
<strong>and</strong> schedule freely across multiple sites would be ideal. Some proprietary vendors can<br />
offer this capability (e.g. Plat<strong>for</strong>m). Indeed, Amaz<strong>on</strong> offers a wide range of services <strong>on</strong> its<br />
Infrastructure. It also offers a range of tools <strong>for</strong> scheduling <strong>and</strong> c<strong>on</strong>figuring applicati<strong>on</strong>s in<br />
the <strong>Cloud</strong>. This is the benchmark. For administrati<strong>on</strong> purposes the UK Government requires<br />
usage accounting to be implemented across the organizati<strong>on</strong>.<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
I would expect that any proposed st<strong>and</strong>ards would need to meet Government requirements<br />
<strong>for</strong> openness <strong>and</strong> quality. I would not expect to spend government resources supporting<br />
testing <strong>for</strong> interoperati<strong>on</strong>. This m<strong>on</strong>ey would be better spent in designing applicati<strong>on</strong>s <strong>for</strong><br />
scalability <strong>and</strong> adaptability into mobile/smartph<strong>on</strong>e domains.
Possible future cooperati<strong>on</strong><br />
I believe that Amaz<strong>on</strong> Web Services is offering the best articulated path <strong>for</strong>wards towards<br />
cloud based scalable applicati<strong>on</strong> support. The specificati<strong>on</strong> of services is more important<br />
than an interoperable API. The ability to design <strong>and</strong> instantiate a c<strong>on</strong>figured applicati<strong>on</strong> is<br />
key to delivering services “<strong>on</strong> dem<strong>and</strong>”. OGF flirted with the idea of templates a couple of<br />
years ago. Chris Smith <strong>and</strong> Ian Osborne presented a paper <strong>on</strong> this topic at <strong>Cloud</strong>World in<br />
San Francisco in August 2009. The UK Government will be willing to collaborate more <strong>on</strong><br />
this subject via the Cabinet Office. However, it is worth c<strong>on</strong>sidering that most government IT<br />
activity is outsourced to major 3rd party Systems Integrators (e.g. HP/EDS, IBM, Capgemini,<br />
CSC, etc.) <strong>and</strong> as such a large measure of interest <strong>and</strong> support is required from them.<br />
C<strong>on</strong>tact: Ian Osborne<br />
Organisati<strong>on</strong>: Intellect<br />
C<strong>on</strong>tact details: ian.osborne@intellectuk.org<br />
Web: www.digitalsystemsktn.org<br />
Relevant Links: www.cabinetoffice.gov.uk/resource-library/g-cloud-programme-phase-2<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
49
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
CitySourced/FreedomSpeaks citizen<br />
services plat<strong>for</strong>m<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
FreedomSpeaks is a political social network with a missi<strong>on</strong> to facilitate governmental<br />
transparency <strong>and</strong> open communicati<strong>on</strong> between c<strong>on</strong>stituents <strong>and</strong> publicly elected officials.<br />
According to the organizati<strong>on</strong>’s missi<strong>on</strong> statement: “In<strong>for</strong>mati<strong>on</strong> has always been power,<br />
so we’ve decided to bring the power back to the people. We’re hard at work digging up our<br />
officials’ voting records, <strong>and</strong> we’d like to keep this data archived <strong>for</strong> all of you.” To fulfill this<br />
pledge, FreedomSpeaks manages an intense data mining <strong>and</strong> translati<strong>on</strong> operati<strong>on</strong>, turning<br />
HTML data files into a collecti<strong>on</strong> of in<strong>for</strong>mati<strong>on</strong> that can be accessed <strong>and</strong> acted up<strong>on</strong> by<br />
engaged citizens. Since its incepti<strong>on</strong> in 2006, FreedomSpeaks has mined data from the<br />
United States Census, Senate, <strong>and</strong> C<strong>on</strong>gressi<strong>on</strong>al records, plus hundreds of state <strong>and</strong> local<br />
government websites.<br />
In 2009, FreedomSpeaks decided to exp<strong>and</strong> its offerings to include a new mobile product,<br />
CitySourced. CitySourced provides a way <strong>for</strong> citizens to report issues in their city using their<br />
smartph<strong>on</strong>es. Residents can take a picture of almost any city issue—such as potholes or<br />
graffiti—then select a category <strong>and</strong> submit that report directly to city hall. The applicati<strong>on</strong><br />
makes use of the smartph<strong>on</strong>e’s internal global positi<strong>on</strong> system (GPS) capabilities <strong>and</strong><br />
internal compass. FreedomSpeaks needed to migrate to a cloud-based technology plat<strong>for</strong>m<br />
to support this new product offering.<br />
The importance of interoperability<br />
50<br />
Interoperability in these systems occurs between users <strong>on</strong> desktop browsers <strong>and</strong> Android,<br />
iPh<strong>on</strong>e, Windows 7 <strong>and</strong> Balckberry smartph<strong>on</strong>es <strong>and</strong> the applicati<strong>on</strong> running <strong>on</strong> the cloud<br />
service; between the applicati<strong>on</strong> running in the cloud service <strong>and</strong> multiple data sources; <strong>and</strong><br />
between the applicati<strong>on</strong> <strong>and</strong> other applicati<strong>on</strong>s run by news outlets, government officials<br />
<strong>and</strong> others.<br />
Since 2006, FreedomSpeaks has added data including legislative data, elected official data,<br />
<strong>and</strong> even geographic in<strong>for</strong>mati<strong>on</strong> systems data. This abundance of data was gathered by<br />
employing a network of spiders to create complex web crawls that execute <strong>on</strong> the cloud<br />
plat<strong>for</strong>m. This in<strong>for</strong>mati<strong>on</strong> is presented <strong>on</strong> the FreedomSpeaks website, <strong>and</strong> provided <strong>for</strong><br />
use by other parites through a rest interface.<br />
The data from the CitySourced soluti<strong>on</strong> is queued up to FreedomSpeaks server computers,<br />
<strong>and</strong> later processed—all of which takes approximately 60 sec<strong>on</strong>ds. It also requires running<br />
milli<strong>on</strong>s of geo lookups against thous<strong>and</strong>s of state <strong>and</strong> city agencies across the United<br />
States. Once processed, the in<strong>for</strong>mati<strong>on</strong> is directed to the appropriate governing body.<br />
When the governmental agency takes care of the issue, a notificati<strong>on</strong> is sent back to the<br />
citizens letting them know that the city has resp<strong>on</strong>ded. This two-way communicati<strong>on</strong>
makes people feel like they are an active part of their local government. CitySourced also<br />
presents in<strong>for</strong>mati<strong>on</strong> through REST <strong>for</strong> use by governments, news agencies <strong>and</strong> others.<br />
Use of REST interfaces in the absence of established semantic st<strong>and</strong>ards allows rapid use of<br />
the data with minimal new programming ef<strong>for</strong>t.<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
The system is based <strong>on</strong> multiple web st<strong>and</strong>ards that support REST, including HTML, XML,<br />
JSON. In additi<strong>on</strong>, the data is presented in st<strong>and</strong>ard geospatial <strong>for</strong>mats, including KML.<br />
Possible future cooperati<strong>on</strong><br />
C<strong>on</strong>necti<strong>on</strong> to additi<strong>on</strong>al government data through open interfaces <strong>and</strong> potentially<br />
decorated with semantic web in<strong>for</strong>mati<strong>on</strong> would allow expansi<strong>on</strong> of both services. For<br />
example, the San Francisco Open311 API allows in<strong>for</strong>mati<strong>on</strong> from CitySourced to be sent<br />
directly into San Francisco’s n<strong>on</strong>-emergency resp<strong>on</strong>se system.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
C<strong>on</strong>tact: Gregg Brown<br />
Organisati<strong>on</strong>: Microsoft<br />
C<strong>on</strong>tact details: GreggB@Microsoft.com<br />
Webs: www.citysourced.com; www.freedomspeaks.com<br />
51
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
CUSTOM – Cultural Heritage & Tourism<br />
Store <strong>on</strong> the <strong>Cloud</strong><br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
The CUSTOM project tries to achieve the technological innovati<strong>on</strong> that comes from the use<br />
of cloud based services <strong>and</strong> applicati<strong>on</strong> integrati<strong>on</strong> in the field of cultural-oriented services.<br />
It aims to create a cultural heritage & tourism store, a plat<strong>for</strong>m <strong>for</strong> the commercializati<strong>on</strong><br />
of integrated applicati<strong>on</strong>s <strong>and</strong> services that will be delivered over the web according to the<br />
Software-as-a-Service paradigm.<br />
Developers of web-oriented services will develop server applicati<strong>on</strong>s according to the<br />
plat<strong>for</strong>m guidelines <strong>for</strong> applicati<strong>on</strong>s interoperability <strong>and</strong> then will sell those applicati<strong>on</strong>s<br />
<strong>on</strong> the CUSTOM market. Public instituti<strong>on</strong>s <strong>and</strong> companies that work in the field of tourism<br />
<strong>and</strong> cultural promoti<strong>on</strong> will buy those applicati<strong>on</strong>s as a service.<br />
CUSTOM will provide the plat<strong>for</strong>m <strong>for</strong> hosting, deploying <strong>and</strong> integrating services<br />
automatically, allowing customers c<strong>on</strong>trol of the acquired software suite, hiding the<br />
complexities of the management of a hardware physical infrastructure. Even though the<br />
plat<strong>for</strong>m will be open to welcome new kinds of services, so far we have <strong>for</strong>eseen the<br />
implementati<strong>on</strong> of several kinds of software building blocks: CMS, GIS Server, Image Library,<br />
Digital Library, Streaming Server.<br />
The importance of interoperability<br />
52<br />
CUSTOM’s focus <strong>on</strong> interoperability will affect both the infrastructure <strong>and</strong> applicati<strong>on</strong><br />
level. As the system will c<strong>on</strong>sist of a cloud plat<strong>for</strong>m, we plan to adopt open-source<br />
soluti<strong>on</strong>s <strong>and</strong> st<strong>and</strong>ard interfaces in order to manage cloud resources. Our choice as a cloud<br />
management system will be based <strong>on</strong> OpenNebula, a plat<strong>for</strong>m which exposes the OCCI<br />
st<strong>and</strong>ard interface <strong>and</strong> partial support <strong>for</strong> the Amaz<strong>on</strong> EC2 API. The latter will be used<br />
by the middleware software that allows automatic management of resources, effectively<br />
decoupling the cloud plat<strong>for</strong>m from the applicati<strong>on</strong>s management module. This choice will<br />
potentially allow the CUSTOM middleware to be moved <strong>on</strong> top of another cloud plat<strong>for</strong>m.<br />
Although CUSTOM does not aim to create an hybrid cloud envir<strong>on</strong>ment that leverages<br />
<strong>on</strong> cloud-bursting techniques, another issue that is related to interoperability is the<br />
capacity to efficiently migrate applicati<strong>on</strong>s <strong>on</strong> top of virtual machines in an heterogeneous<br />
virtualizati<strong>on</strong> plat<strong>for</strong>m envir<strong>on</strong>ment.<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
At the infrastructure level OpenNebula provides a subset of Amaz<strong>on</strong> EC2 API <strong>and</strong> the<br />
st<strong>and</strong>ard OCCI interface. It also provides seamless integrati<strong>on</strong> with Amaz<strong>on</strong> EC2 public
cloud, allowing partial c<strong>on</strong>trol of the resources related to this cloud envir<strong>on</strong>ment, which<br />
<strong>on</strong>ly requires a working EC2/S3 account with already loaded AMIs. OpenNebula provides<br />
an implementati<strong>on</strong> of OCCI based <strong>on</strong> the latest draft of the OGF OCCI specificati<strong>on</strong>, al<strong>on</strong>g<br />
with libraries <strong>for</strong> Ruby <strong>and</strong> Java language. The software that will provide user interfaces<br />
or manage the automatic deployment of cloud resources <strong>and</strong> customer applicati<strong>on</strong><br />
envir<strong>on</strong>ment will make use of the OCCI functi<strong>on</strong>alities in order to interoperate with the<br />
cloud plat<strong>for</strong>m.<br />
We plan to thoroughly test the OCCI implementati<strong>on</strong> of OpenNebula <strong>and</strong> the provided<br />
libraries prior to starting the development of the middleware software.<br />
Possible future cooperati<strong>on</strong><br />
»»<br />
OpenNebula <strong>and</strong> the RESERVOIR <str<strong>on</strong>g>European</str<strong>on</strong>g> project (opennebula.org/)<br />
»»<br />
Claudia Plat<strong>for</strong>m (claudia.morfeo-project.org/)<br />
»»<br />
OCCI (occi-wg.org/)<br />
»»<br />
OVF (dmtf.org/st<strong>and</strong>ards/vman)<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
C<strong>on</strong>tact: Paola P<strong>on</strong>ticelli<br />
C<strong>on</strong>tact details: p.p<strong>on</strong>ticelli@liberologico.com<br />
Web: www.customstore.it<br />
Relevant Links: www.liberologico.com<br />
53
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
54<br />
St<strong>and</strong>ards & Interfaces<br />
OpenNebula - A Reference Open <strong>Cloud</strong><br />
Stack to Enable Interoperable Enterpriseclass<br />
<strong>Cloud</strong> Computing Plat<strong>for</strong>ms<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
OpenNebula is a fully open-source toolkit to build any type of IaaS cloud: private, public<br />
<strong>and</strong> hybrid. The OpenNebula technology is the result of many years of research <strong>and</strong><br />
development in efficient <strong>and</strong> scalable management of virtual machines <strong>on</strong> large-scale<br />
distributed infrastructures. Its innovative features have been developed to address the<br />
requirements of business use cases from leading IT companies <strong>and</strong> across multiple industries<br />
in the c<strong>on</strong>text of groundbreaking projects in cloud computing, such as RESERVOIR.<br />
Additi<strong>on</strong>ally, it is being used as reference open stack <strong>for</strong> cloud computing in several large<br />
research <strong>and</strong> infrastructure projects, such as StratusLab, B<strong>on</strong>FIRE, or 4CaaSt.<br />
The OpenNebula technology has matured thanks to an active <strong>and</strong> engaged community of<br />
users <strong>and</strong> developers. The development is driven by its community in order to support the<br />
most dem<strong>and</strong>ed features, <strong>and</strong> by the internati<strong>on</strong>al research projects funding OpenNebula<br />
in order to address the dem<strong>and</strong>ing requirements of several business <strong>and</strong> scientific use<br />
cases <strong>for</strong> cloud computing. OpenNebula has proved to be a producti<strong>on</strong>-ready soluti<strong>on</strong><br />
that includes enterprise features such as security, robustness, scalability <strong>and</strong> per<strong>for</strong>mance<br />
that many IT shops need <strong>for</strong> internal cloud adopti<strong>on</strong>, either in scientific or in business<br />
envir<strong>on</strong>ments.<br />
OpenNebula is downloaded several thous<strong>and</strong>s times a m<strong>on</strong>th from its site, <strong>and</strong> the code<br />
can also be downloaded from the software repository <strong>and</strong> from several commercial <strong>and</strong><br />
open-source distributi<strong>on</strong>s. OpenNebula is used by thous<strong>and</strong>s of organizati<strong>on</strong>s worldwide to<br />
research the challenges that arise in cloud management, <strong>and</strong> also as producti<strong>on</strong>-ready tool<br />
in both academia <strong>and</strong> industry to manage clouds. Users include some of the world’s leading<br />
telecom operators, hosting providers <strong>and</strong> compute centers of leading research instituti<strong>on</strong>s.<br />
The importance of interoperability<br />
OpenNebula emphasizes interoperability <strong>and</strong> portability, providing cloud users <strong>and</strong><br />
administrators with choice across the most popular cloud interfaces, hypervisors <strong>and</strong><br />
public clouds <strong>for</strong> hybrid cloud computing deployments, <strong>and</strong> with a flexible software that<br />
can be installed in any hardware <strong>and</strong> software combinati<strong>on</strong>. The functi<strong>on</strong>ality provided by<br />
OpenNebula <strong>and</strong> the comp<strong>on</strong>ents in its quickly growing ecosystem enable:<br />
» » Interoperability in the private cloud by supporting most comm<strong>on</strong> hypervisors, such as<br />
KVM, VMware or Xen, <strong>and</strong> many other virtualizati<strong>on</strong> stacks through its libvirt plug-in
»»<br />
Interoperability in the public cloud by exposing most comm<strong>on</strong> cloud interfaces, such as<br />
VMware v<strong>Cloud</strong> <strong>and</strong> Amaz<strong>on</strong> EC2; open community specificati<strong>on</strong>s, such us the OGF Open<br />
<strong>Cloud</strong> Computing Interface; <strong>and</strong> open interfaces, such as libcloud <strong>and</strong> deltacloud<br />
»»<br />
Interoperability in the hybrid cloud by supporting the combinati<strong>on</strong> of local private<br />
infrastructure with Amaz<strong>on</strong> EC2 <strong>and</strong> ElasticHosts, <strong>and</strong> any major cloud provider, such as<br />
Rackspace, Go<strong>Grid</strong> or Terremark through a RedHat’s deltacloud adaptor<br />
Adopti<strong>on</strong> of emerging or existing st<strong>and</strong>ards<br />
Our plan is to c<strong>on</strong>tinue our support <strong>for</strong> EC2 <strong>and</strong> OGF OCCI <strong>Cloud</strong> APIs. Both implementati<strong>on</strong>s<br />
are now being used in very large-scale deployments. Our users have reported scalability<br />
results with tens of thous<strong>and</strong>s of virtual machines. EC2 interoperability has been validated<br />
with Amaz<strong>on</strong> AWS. In fact OpenNebula can be used with any of the tools available in the<br />
Amaz<strong>on</strong> ecosystem, such as ElasticFox.<br />
From the perspective of the OpenNebula project, interoperability in the c<strong>on</strong>text of<br />
infrastructure requires openness, adaptability, portability <strong>and</strong> st<strong>and</strong>ardizati<strong>on</strong>. Because<br />
two data centers are not the same, building a cloud computing infrastructure requires the<br />
integrati<strong>on</strong> <strong>and</strong> orchestrati<strong>on</strong> of the underlying existing IT systems, services <strong>and</strong> processes.<br />
OpenNebula enables interoperability <strong>and</strong> portability, recognizing that our users have<br />
data-centers composed of different hardware <strong>and</strong> software comp<strong>on</strong>ents <strong>for</strong> security,<br />
virtualizati<strong>on</strong>, storage, <strong>and</strong> networking. Its open architecture, interfaces <strong>and</strong> comp<strong>on</strong>ents<br />
provide the flexibility <strong>and</strong> extensibility that many enterprise IT shops need <strong>for</strong> internal<br />
cloud adopti<strong>on</strong>.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
C<strong>on</strong>tact: Ignacio M. Llorente<br />
Organisati<strong>on</strong>: Complutense University of Madrid<br />
C<strong>on</strong>tact details: imllorente@opennebula.org<br />
Web: www.OpenNebula.org<br />
55
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
OCCI - Open <strong>Cloud</strong> Computing Interface<br />
specificati<strong>on</strong> set<br />
General overview <strong>and</strong> field of applicati<strong>on</strong><br />
The Open <strong>Cloud</strong> Computing Interface from the Open <strong>Grid</strong> Forum is a RESTful protocol <strong>and</strong><br />
API <strong>for</strong> cloud-related management tasks. OCCI can be used by any device or programming<br />
language that is able to underst<strong>and</strong> HTTP, <strong>and</strong> provides <strong>for</strong> easy m<strong>on</strong>itoring <strong>and</strong> testing<br />
through the HTTP Rendering.<br />
Originally initiated to create a remote management API <strong>for</strong> IaaS-based services, it has since<br />
evolved into a flexible API while still offering a high degree of extensibility. The current<br />
release is suitable to serve many other models in additi<strong>on</strong> to IaaS, including e.g. PaaS <strong>and</strong><br />
SaaS.<br />
The importance of interoperability<br />
The current OCCI specificati<strong>on</strong> set c<strong>on</strong>sists of three documents. Future releases are planned<br />
to include additi<strong>on</strong>al rendering <strong>and</strong> extensi<strong>on</strong> specificati<strong>on</strong>s.<br />
»»<br />
OCCI Core: Provides the <strong>for</strong>mal definiti<strong>on</strong> of the OCCI Core Model.<br />
»»<br />
OCCI HTTP Rendering: Defines how to interact with the OCCI Core Model using the OCCI<br />
API, including how the Model can be communicated <strong>and</strong> serialized using the HTTP protocol.<br />
»»<br />
OCCI Infrastructure: C<strong>on</strong>tains the definiti<strong>on</strong> of the OCCI Infrastructure extensi<strong>on</strong> <strong>for</strong><br />
the IaaS domain; also defines associated resource types, their attributes <strong>and</strong> the acti<strong>on</strong>s<br />
that can be taken <strong>on</strong> each resource type.<br />
56<br />
Adopti<strong>on</strong> by the open source community<br />
OCCI has achieved wide adopti<strong>on</strong> in the open source community <strong>and</strong> has attracted<br />
c<strong>on</strong>siderable interest from the commercial community <strong>and</strong> from other st<strong>and</strong>ards<br />
organizati<strong>on</strong>s due to its built-in inter-compatibility with other RESTful methods.<br />
Implementati<strong>on</strong>s exist that can be downloaded <strong>and</strong> used from a number of projects,<br />
including the following:<br />
1. Implementati<strong>on</strong> of OCCI <strong>on</strong> top of libvirt by the Distributed Computing Virtual Laboratory<br />
at the Robotics Research Institute, Technische Universität Dortmund.<br />
2. A BSD-licensed OCCI implementati<strong>on</strong> <strong>on</strong> top of Apache Tashi by SLA@SOI.<br />
3. An open source implementati<strong>on</strong> of OGF OCCI <strong>for</strong> Eucalyptus under development by he<br />
UK-JISC funded project “Flexible Services <strong>for</strong> the Support of Research.”<br />
4. Adopti<strong>on</strong> of OCCI into the roadmap <strong>and</strong> project plan <strong>for</strong> OpenStack, scheduled <strong>for</strong> the<br />
upcoming ‘bexar’ release.
5. A reference implementati<strong>on</strong> of the OCCI specificati<strong>on</strong> by the OpenNebula project,<br />
scheduled to be updated to the latest versi<strong>on</strong> in the near future.<br />
6. An implementati<strong>on</strong> of the OCCI protocol/API as part of the Service Sharing Facility (SSF)<br />
<strong>for</strong> the German Research Project DGSI, developed by Plat<strong>for</strong>m Computing.<br />
All of the above implementati<strong>on</strong>s except <strong>for</strong> OpenStack are already fully functi<strong>on</strong>al, <strong>and</strong><br />
many have been in the <strong>for</strong>m of working code <strong>for</strong> existing projects <strong>for</strong> some time. The latter<br />
implementati<strong>on</strong> includes demos <strong>for</strong> Job Submissi<strong>on</strong> (SaaS/PaaS), a KeyValue store (PaaS)<br />
<strong>and</strong> an included skelet<strong>on</strong> implementati<strong>on</strong> of the OCCI infrastructure model, which can be<br />
bound to any available hyper-visor to create an IaaS based cloud.<br />
The OCCI specificati<strong>on</strong>s are designed to allow boundary-level interfaces to be built using<br />
RESTful patterns over HTTP, <strong>and</strong> can thus be applied to almost any existing software<br />
infrastructure comp<strong>on</strong>ent or layer to provide a st<strong>and</strong>ards-based way to adapt it to the<br />
cloud. This feature accounts <strong>for</strong> their high degree of interest <strong>and</strong> adopti<strong>on</strong>.<br />
Possible future cooperati<strong>on</strong><br />
Formal release of the OCCI specificati<strong>on</strong>s is expected by the end of 1st quarter 2011. The<br />
OCCI group welcomes notificati<strong>on</strong> regarding additi<strong>on</strong>al implementati<strong>on</strong>s <strong>and</strong> further work,<br />
<strong>and</strong> collaborates through its ogf.org working group pages supplemented by a dedicated<br />
web site at occi-wg.org that hosts links to downloadable examples of the implementati<strong>on</strong>s<br />
described above.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
C<strong>on</strong>tact: Alan Sill<br />
Organisati<strong>on</strong>: Open <strong>Grid</strong> Forum<br />
C<strong>on</strong>tact details: alan.sill@ttu.edu<br />
Web: occi-wg.org<br />
Relevant Links: www.ogf.org<br />
57
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
Legal, Ec<strong>on</strong>omic, Ethical <strong>and</strong> Security Issues<br />
<strong>Cloud</strong> computing <strong>and</strong> its ethical challenges<br />
Abstract<br />
The paper analyses some important ethical challenges posed by cloud computing,<br />
c<strong>on</strong>cerning ownership, safety, fairness, resp<strong>on</strong>sibility, accountability <strong>and</strong> privacy.<br />
Ownership, possessi<strong>on</strong>, <strong>and</strong> use<br />
<strong>Cloud</strong> computing is part of the c<strong>on</strong>temporary tendency towards the deflati<strong>on</strong> of the<br />
noti<strong>on</strong> of ownership <strong>and</strong> the uniqueness of what is owned. The underlying idea is that use<br />
does not imply ownership, <strong>for</strong> it might require <strong>on</strong>ly temporary possessi<strong>on</strong>. Owning <strong>and</strong><br />
there<strong>for</strong>e maintaining large <strong>and</strong> complex hardware resources is a limiting, expensive <strong>and</strong><br />
often unsustainable overhead <strong>for</strong> users. The issue here is that, while the ownership of the<br />
hardware supporting computing activities is not needed or wanted anymore, the ownership<br />
of the outcome of such activities remains vital.<br />
Safety, reliability <strong>and</strong> data insurance<br />
58<br />
Storing large amounts of potentially sensitive data <strong>on</strong> hardware facilities owned by private<br />
companies poses the problem of how <strong>and</strong> why the storage provider should be trusted in<br />
managing them properly. The soluti<strong>on</strong> here seems to lie in the improvement of the legal<br />
c<strong>on</strong>straints that can make providers trustworthy <strong>and</strong> in transferring the full ownership <strong>and</strong><br />
c<strong>on</strong>trol of the data access <strong>and</strong> usage from the provider to the user.<br />
Fairness <strong>and</strong> digital divide<br />
<strong>Cloud</strong> computing c<strong>on</strong>tributes to a democratisati<strong>on</strong> of computing resources through their<br />
potential wider distributi<strong>on</strong> at a lower cost. Yet the digital divide is also a problem of accessibility<br />
<strong>and</strong> usability, <strong>and</strong> in these two respects, <strong>Cloud</strong> computing may easily exacerbate it.<br />
C<strong>on</strong>trol <strong>and</strong> resp<strong>on</strong>sibility<br />
<strong>Cloud</strong> computing shifts the c<strong>on</strong>trol of a computati<strong>on</strong>al infrastructure from the provider<br />
to the user. Users remain legally resp<strong>on</strong>sible <strong>for</strong> their wr<strong>on</strong>gdoing but they are not preemptively<br />
incapacitated to misuse the provided infrastructure. They are assumed to be
entirely resp<strong>on</strong>sible of their computing activities because they are fully empowered. This<br />
leads to a more complicated issue, the relati<strong>on</strong>ship between accountability <strong>and</strong> privacy.<br />
Accountability <strong>and</strong> privacy<br />
Accountability is used to en<strong>for</strong>ce resp<strong>on</strong>sibility, so it may be seen as a positive factor in<br />
the management of <strong>Cloud</strong> computing. However, accountability has a direct impact <strong>on</strong> the<br />
levels of privacy <strong>and</strong> an<strong>on</strong>ymity of the users. In order to be accountable, users’ acti<strong>on</strong>s<br />
need to be traceable <strong>and</strong>, as such, their physical identity must be knowable to the provider,<br />
while their acti<strong>on</strong>s must leave meaningful traces that can be used to identify, prove <strong>and</strong><br />
quantify the damage or offence caused by reckless behaviours. Arguably, a principle should<br />
be endorsed <strong>for</strong> which, am<strong>on</strong>g all the available implementati<strong>on</strong> of accountability, the <strong>on</strong>e<br />
that minimizes the erosi<strong>on</strong> of the right to privacy <strong>and</strong> to an<strong>on</strong>ymity is chosen. For this<br />
reas<strong>on</strong>, soluti<strong>on</strong>s based <strong>on</strong> federated authenticati<strong>on</strong> <strong>and</strong> authorisati<strong>on</strong> <strong>and</strong> policed logs<br />
access should be preferred to those based <strong>on</strong> proactive <strong>and</strong> invasive practices, like deep<br />
packet inspecti<strong>on</strong> or proactive log mining.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
C<strong>on</strong>tact Pers<strong>on</strong>: Luciano Floridi<br />
Organisati<strong>on</strong>: University of Hert<strong>for</strong>dshire<br />
C<strong>on</strong>tact details: luciano.floridi@philosophy.ox.ac.uk<br />
Relevant Link: www.philosophyofin<strong>for</strong>mati<strong>on</strong>.net/Welcome.html<br />
C<strong>on</strong>tact: Matteo Turilli<br />
Organisati<strong>on</strong>: University of Ox<strong>for</strong>d<br />
C<strong>on</strong>tact details: matteo.turilli@oerc.ox.ac.uk<br />
Relevant Link: www.oerc.ox.ac.uk<br />
59
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
VENUS-C Study <strong>on</strong> ec<strong>on</strong>omic <strong>and</strong> legal<br />
implicati<strong>on</strong>s of sustainable scientific clouds<br />
The importance of interoperability<br />
The interoperability of the Venus-C infrastructure technologies is a key element, which<br />
may be very useful to develop new findings in a collaborative <strong>and</strong> more efficient manner.<br />
One research questi<strong>on</strong> of the VENUS-C project will be whether interoperability, am<strong>on</strong>g<br />
different authorisati<strong>on</strong> systems, can be enabled by mapping the comm<strong>on</strong> language defined<br />
by the plat<strong>for</strong>m to <strong>on</strong>es defined by the local infrastructures. A loss of interoperability may<br />
cause relevant ec<strong>on</strong>omic but also immaterial costs <strong>for</strong> the scientists, because here resides<br />
the added value of the <strong>Cloud</strong> plat<strong>for</strong>m. However, interoperability generates legal <strong>and</strong><br />
ec<strong>on</strong>omic issues.<br />
Potential Ec<strong>on</strong>omic issues of Scientific <strong>Cloud</strong>s:<br />
60<br />
»»<br />
Marginal costs: of operating <strong>on</strong> a <strong>Cloud</strong> provider’s infrastructure at certain volumes of<br />
data traffic may become more expensive than providing the necessary IT infrastructure<br />
in-house.<br />
»»<br />
Service interrupti<strong>on</strong> or disrupti<strong>on</strong>: this could cause significant damages <strong>and</strong> a loss of<br />
scientists’ reputati<strong>on</strong>. If a large amount of pers<strong>on</strong>al <strong>and</strong> sensitive data is lost, how can<br />
some<strong>on</strong>e quantify this serious damage that is not <strong>on</strong>ly legal but also ethical?<br />
»»<br />
Licensing costs: if the scientific communities want to modify legacy applicati<strong>on</strong> to<br />
functi<strong>on</strong> in the <strong>Cloud</strong>, after the stipulati<strong>on</strong> of the c<strong>on</strong>tract, this may cause high costs.<br />
»»<br />
Supply chain failure <strong>and</strong> problem with transfer of data <strong>and</strong> software am<strong>on</strong>g different<br />
cloud service providers: when a <strong>Cloud</strong> provider outsources some of its chain services to<br />
third parties, the level security of data can be reduced.<br />
»»<br />
Availability of programming skills to modify legacy applicati<strong>on</strong> to functi<strong>on</strong> in the<br />
<strong>Cloud</strong>: may cause high costs <strong>for</strong> the scientific communities.<br />
»»<br />
Increase of CPU-based licensing costs when we moved to a cloud plat<strong>for</strong>m: in this case<br />
licensing costs of transferring could be very high <strong>and</strong> could reduce the value of a <strong>Cloud</strong><br />
plat<strong>for</strong>m.<br />
Potential legal issues <strong>and</strong> st<strong>and</strong>ards of scientific <strong>Cloud</strong>s<br />
In order to make data cooperati<strong>on</strong> secure <strong>for</strong> researchers, we need to identify the most<br />
relevant issues <strong>and</strong> st<strong>and</strong>ards that allow them to trigger their discovery processes:<br />
» » Data protecti<strong>on</strong>: in the scientific <strong>Cloud</strong> envir<strong>on</strong>ment sensitive <strong>and</strong> c<strong>on</strong>fidential data<br />
can be shared ethically if researchers obtain in<strong>for</strong>med c<strong>on</strong>sent to do so. C<strong>on</strong>sent is<br />
also needed <strong>for</strong> the participati<strong>on</strong> to the research, obtaining c<strong>on</strong>sent <strong>for</strong> the publicati<strong>on</strong>
of results in which their data are included, protecting the identity of the participants,<br />
deciding if restricti<strong>on</strong>s <strong>on</strong> data access applies to the in<strong>for</strong>mati<strong>on</strong> as a whole. In the case<br />
of scientific <strong>Cloud</strong> communities, which exchange data over different countries, the “EU<br />
Binding Corporate Rules” provide a scheme, which may really help also the Venus-C<br />
project to reduce this problem <strong>and</strong> to ensure the data safety to researchers.<br />
»»<br />
Privacy <strong>and</strong> c<strong>on</strong>fidentiality: in the Scientific envir<strong>on</strong>ment, there are two kinds of<br />
c<strong>on</strong>fidentiality data risks: an identificati<strong>on</strong> disclosure risk <strong>and</strong> an attribute disclosure<br />
risk. Often the <str<strong>on</strong>g>European</str<strong>on</strong>g> Directives <strong>and</strong> the ECPA are not sufficient to protect users, the<br />
Venus-C project should trigger two different approaches (the restricted data <strong>and</strong> the<br />
restricted access) to ensure the privacy safety of data.<br />
»»<br />
Intellectual Property Rights: within this c<strong>on</strong>text, who may have the recogniti<strong>on</strong> of being<br />
the author of the work if the research results stem from a shared process of generati<strong>on</strong>?<br />
Within the shared <strong>Cloud</strong> envir<strong>on</strong>ment does it still make sense to talk about Intellectual<br />
Property? The issue is to find a good balance between IPR protecti<strong>on</strong> <strong>and</strong> open access to<br />
research results.<br />
»»<br />
Identity <strong>and</strong> Access Management: in the case of scientific <strong>Cloud</strong> communities, it is<br />
not clear how to identify which kind of st<strong>and</strong>ards <strong>and</strong> protocols should apply to the<br />
in<strong>for</strong>mati<strong>on</strong> exchanged in the scientific <strong>Cloud</strong> envir<strong>on</strong>ment, those related to the single<br />
users or those related to a large community?<br />
Be<strong>for</strong>e starting a research project in a <strong>Cloud</strong> Computing envir<strong>on</strong>ment, the scientific<br />
communities involved should agree to adopt comm<strong>on</strong> st<strong>and</strong>ards that evaluate who is<br />
resp<strong>on</strong>sible <strong>for</strong> data security. Relevant st<strong>and</strong>ards are the ISO/IEC 27000-27001-27002<br />
series <strong>and</strong> SAS70.<br />
At present, st<strong>and</strong>ards <strong>for</strong> scientists regarding Service Level Agreements, do not take <strong>on</strong><br />
board security issues. Instead they focus <strong>on</strong>: reliability, throughput, durability, elasticity,<br />
linearity, agility, automati<strong>on</strong>, customer service resp<strong>on</strong>se times <strong>and</strong> load balancing. All these<br />
issues have also been widely analysed in our study.<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
C<strong>on</strong>tact: Francesca Spagnoli<br />
Organisati<strong>on</strong>: Engineering<br />
C<strong>on</strong>tact details: Francesca.spagnoli@eng.it<br />
Web: www.venus-c.eu<br />
61
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
The <strong>Cloud</strong>: Underst<strong>and</strong>ing security, privacy<br />
<strong>and</strong> trust challenges<br />
The overall objective of The <strong>Cloud</strong>: Underst<strong>and</strong>ing the Security, Privacy <strong>and</strong> Trust Challenges<br />
study is to advise <strong>on</strong> policy <strong>and</strong> other interventi<strong>on</strong>s which should be c<strong>on</strong>sidered in order<br />
to ensure that <str<strong>on</strong>g>European</str<strong>on</strong>g> users of cloud envir<strong>on</strong>ments are offered appropriate protecti<strong>on</strong>s,<br />
<strong>and</strong> to underpin a world-leading <str<strong>on</strong>g>European</str<strong>on</strong>g> cloud ecosystem. <strong>Cloud</strong> computing is<br />
increasingly subject to interest from policymakers <strong>and</strong> regulatory authorities. The <str<strong>on</strong>g>European</str<strong>on</strong>g><br />
Commissi<strong>on</strong>’s recent Digital Agenda highlighted a need to develop a pan-<str<strong>on</strong>g>European</str<strong>on</strong>g> ‘cloud<br />
strategy’ that will serve to support growth <strong>and</strong> jobs <strong>and</strong> build an innovati<strong>on</strong> advantage <strong>for</strong><br />
Europe. However, the c<strong>on</strong>cern is that currently a number of challenges <strong>and</strong> risks in respect<br />
of security, privacy <strong>and</strong> trust exist that may undermine the attainment of these broader<br />
policy objectives. Our approach has been to undertake an analysis of the technological,<br />
operati<strong>on</strong>al <strong>and</strong> legal intricacies of cloud computing, taking into c<strong>on</strong>siderati<strong>on</strong> the <str<strong>on</strong>g>European</str<strong>on</strong>g><br />
dimensi<strong>on</strong> <strong>and</strong> the interests <strong>and</strong> objectives of all stakeholders (citizens, individual users,<br />
companies, cloud service providers, regulatory bodies <strong>and</strong> relevant public authorities). We<br />
undertook literature <strong>and</strong> document review, interviews, case studies <strong>and</strong> held an expert<br />
workshop to identify, explore <strong>and</strong> validate these issues in more depth. The present paper<br />
represents the final c<strong>on</strong>solidati<strong>on</strong> of all inputs, suggesti<strong>on</strong>s <strong>and</strong> analyses <strong>and</strong> c<strong>on</strong>tains our<br />
recommendati<strong>on</strong>s <strong>for</strong> policy <strong>and</strong> other interventi<strong>on</strong>s.<br />
C<strong>on</strong>tact: Neil Robins<strong>on</strong><br />
Organisati<strong>on</strong>: R<strong>and</strong> Europe<br />
C<strong>on</strong>tact details: neilr@r<strong>and</strong>.org<br />
Relevant Link: Full report cordis.europa.eu/fp7/ict/security/publicati<strong>on</strong>s_en.html<br />
62
Glossary<br />
API<br />
BES<br />
CDMI<br />
CMWG<br />
DCI<br />
DEISA<br />
DMI<br />
EIRO<br />
HPC<br />
HTC<br />
IaaS<br />
CNSR/IBCP<br />
ISV<br />
JMS<br />
JSDL<br />
MoU<br />
NGI<br />
OCCI<br />
OS<br />
OVF<br />
PaaS<br />
SaaS<br />
SAML<br />
SCAP<br />
SLA<br />
SM<br />
SPG<br />
SRM<br />
VEEH<br />
VEEM<br />
VM<br />
VO<br />
VOMS<br />
VOMS<br />
VRC<br />
Applicati<strong>on</strong> Programming Interface<br />
Basic Executi<strong>on</strong> Service<br />
<strong>Cloud</strong> Data Management Interface<br />
<strong>Cloud</strong> Management Working Group<br />
Distributed Computer Infrastructure<br />
Distributed <str<strong>on</strong>g>European</str<strong>on</strong>g> Infrastructure <strong>for</strong> Supercomputing Applicati<strong>on</strong>s<br />
Desktop Management Interface<br />
<str<strong>on</strong>g>European</str<strong>on</strong>g> Internati<strong>on</strong>al Research Organisati<strong>on</strong><br />
High Per<strong>for</strong>mance Computing<br />
High Throughput Computing<br />
Infrastructure-as-a-Service<br />
Nati<strong>on</strong>al Centre of Scientific Research/ Institute of Biology <strong>and</strong><br />
Chemistry of Proteins (Ly<strong>on</strong>, France)<br />
Independent Software Vendor<br />
Java Message Service<br />
Job Submissi<strong>on</strong> Descripti<strong>on</strong> Language<br />
Memor<strong>and</strong>um of Underst<strong>and</strong>ing<br />
Nati<strong>on</strong>al <strong>Grid</strong> Initiatives<br />
Open <strong>Cloud</strong> Computing Interface<br />
Open Source<br />
Open Virtualizati<strong>on</strong> Format<br />
Plat<strong>for</strong>m as a Service<br />
Software as as Service<br />
Security Asserti<strong>on</strong> Markup Language<br />
Security C<strong>on</strong>tent Automati<strong>on</strong> Protocol<br />
Service Level Agreements<br />
Service Manager<br />
Security Policy Group<br />
Storage Resource Management<br />
Virtual Executi<strong>on</strong> Envir<strong>on</strong>ment Host<br />
Virtual Executi<strong>on</strong> Envir<strong>on</strong>ment Management<br />
Virtual Machine<br />
Virtual Organisati<strong>on</strong><br />
Virtual Organizati<strong>on</strong> Membership Service<br />
Virtual Organizati<strong>on</strong> Management Service<br />
Virtual Research Community<br />
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
63
<strong>Cloud</strong>Scape III - Taking <str<strong>on</strong>g>European</str<strong>on</strong>g> <strong>Cloud</strong> Infrastructure Forward<br />
64<br />
Disclaimer<br />
The views expressed in the use cases <strong>and</strong> positi<strong>on</strong> papers in this document are those of<br />
the authors <strong>and</strong> do not necessarily reflect the view of the <str<strong>on</strong>g>SIENA</str<strong>on</strong>g> project or the authors’<br />
organisati<strong>on</strong>s <strong>and</strong>/or affiliates. Copyright <str<strong>on</strong>g>SIENA</str<strong>on</strong>g>.
<str<strong>on</strong>g>SIENA</str<strong>on</strong>g> (RI-261575) is funded by the <str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong> under<br />
Framework Programme 7 (2007-2013) Research infrastructures projects<br />
<str<strong>on</strong>g>European</str<strong>on</strong>g> Commissi<strong>on</strong>