Towards a Platform for Widespread Embedded Intelligence - ERCIM

SPECIAL THEME: Embedded Intelligence
Networking Semantic Services
for Pervasive Computing
Networking Semantic Services for Pervasive Computing
by Valerie Issarny, Nikolaos Georgantas and Sonia Ben Mokhtar
With computing and communication
By Valerie
capabilities
Issarny, Nikolaos
now embedded
Georgantas
in more
and
and
Sonia
service
Ben Mokhtar
behaviour protocols. A signifi-
more physical objects in the surrounding environment and with most users cant enabler for dealing with these issues
With computing and communication capabilities now embedded in more and more physical
carrying wireless computing devices, the pervasive computing vision pioneered is the Semantic Web and in particular the
by Mark Weiser is being realised
objects
more
of
than
the surrounding
ever. Nevertheless,
environment
such an abundance
and most users
Semantic
carrying
Web
wireless
Services
computing
paradigms,
devices, the
of ambient capabilities in
pervasive
the open, dynamic
computing
pervasive
vision
environment
pioneered by
naturally
Mark Weiser
which
is
allow
more
description
than ever
of
becoming
service
possible.
comes with high heterogeneity, Nevertheless, which hinders such their abundance ad hoc integration of ambient into rich capabilities semantics in and the behaviour open, and dynamic rigorouspervasive
pervasive applications. Within environment the INRIA ARLES naturally team, comes we are investigating with high heterogeneity, efficient reasoning which about hinders the former. their ad Tackling hoc integration
solutions to pervasive heterogeneity into rich pervasive building upon applications. semantic Within service-oriented the INRIA both ARLES semantics team, and we behaviour are investigating hetero- efficient
architectures. solutions to pervasive heterogeneity building upon semantic geneity in service-oriented a combined way architectures.
is still an
open, challenging issue, further exacer-
Various pervasive applications Various may bepervasive
call for applications computing paradigms may be that provisioned enable bated in today’s by the fact networked that semantic environments. reasoning Figure 1
provided in today's networked illustrates environexploitation
loose, dynamic of the devices coupling in among the surroundings hetero- is to a create compute-intensive smart spaces for task, users. tooIn
such an
ments. Figure 1 illustrates exploitation environment, of the geneous nomadic resources. user may Service-oriented
in particular benefit demanding from a large for the display wireless, that resource- is nearby to watch
the networked devices to multimedia create smartcontent
architectures provided (SOA) by a have local emerged media server, as an while constrained the content devices will of automatically the pervasive be switched
spaces for users. In such an environment, to the user’s smart appropriate phone display computing as soon paradigm as other users environment. get close to preserve the user’s privacy. This
the nomadic user may, in particular, application ben- may resolving further be a extended number of to heterogeneity
share experience with others. For instance, assume that the user
efit from a large display nearby is actually to watchin
a issues. holiday However, resort, watching pervasive an advertisement environ- As for part available of the excursions. work of the Then, INRIA the user may
multimedia content provided will by to a local get feedback ments have about further the excursion, introduced which new chal- may be ARLES obtained team, through on the blogs development maintained of by other
media server. The content will nearby automati- visitors lenges on their for handhelds SOA both at and/or middleware the Internet. and distributed The user may systems even enabling will to the look perva- at alternative
cally be switched to the user's excursions, smartwhether
at application advertised layer. at nearby Regarding resorts, the at tourist sive information computing vision, offices we or are even devising on the Internet.
phone display as soon as other Actuator users get resources former, like devices displays, need to and dynamically software resources enablers like for pervasive blogs, multimedia services. We streams and
close, to preserve the user's privacy. advertisements This detect on the services Web, all available constitute in the networked open net- capabilities have, in particular, that may investigated conveniently interop- be composed
application may further be to extended realize a topervasive
worked application environment enabling and to interact a holiday withplanner
erability for nomadic enablers and users. their integration
share experiences with others. For
instance, assume that the user is actually
in a holiday resort, watching an advertisement
for available excursions. The
user may wish to get feedback about the
excursion, which may be obtained
through blogs supplied by other nearby
visitors on their handhelds and/or the
User detection
Internet. The user may even wish to look
at alternative excursions, whether advertised
at nearby resorts, at tourist information
offices or even on the Internet. Figure 1: Integrating resources in user-centric smart spaces.
Actuator resources such as displays and
Figure 1. Integrating resources in user-centric smart spaces
software resources such as blogs, multimedia
streams and advertisements Pervasive on the computing them. This environments requires enabling as interoper- the one exemplified within the base above SOA embed architectural networked style. devices,
Web, all constitute networked possibly capabili- wireless, ability from of both various discovery application and communi- domains, e.g., The mobile resulting and SOA personal style for computing, pervasive consumer
ties that may conveniently be electronics, combinedand
cation homeprotocols, automation as services domains. are This imple- diversity computing of devices offers along two with key functionali- the ones of networks
to dynamically create a pervasive and of appli- software mented infrastructures on top of diverse call for middleware computing paradigms ties, each that one enable vertically loose, integrating dynamic coupling
cation providing a holiday among planner heterogeneous for platforms resources. (eg, UPnP Service-oriented used in the home architectures related advanced (SOA) have features emerged in the middle- as a well-serving
nomadic users. such paradigm and resolving Jini in a number the mobile of heterogeneity domain). issues. ware and However, in the application pervasive environments layer, as have
further introduced Regarding new challenges the latter, applications for SOA both need at tomiddleware
depicted in and Figure at application 2. First, semantic-rich layer. Regarding the
Pervasive computing environments, former, devices like compose, need to dynamically in an ad hoc fashion, detect services available interoperable in the service open networked discovery enables environment and
the one shown above, embed to networked interact with available them. This at the requires specific enabling time and place, interoperability the locating of both of discovery networked and services communication
devices, possibly wireless, from protocols, various as services which have are implemented been developed on top indepen- of diverse according middleware to their platforms semantic (e.g., descriptions. UPnP used in the
application domains, eg, mobile home and and perJini
in dently the mobile without domain). a priori mutual Regarding knowl- the latter, In the applications middleware need layer, to discovery compose in isan
ad hoc
sonal computing, consumer fashion electronics, services edge. available The major at the challenges specific are time enabling and place, achieved, which despite have been the possible developed coexis- independently
and home automation domains. without This a priori common mutual understanding knowledge. Major and adaptation challenges tence concern of various enabling discovery common protocols, understanding by and
diversity of devices, along with adaptation those of of the of the semantics of of the the services being associated, interworking and between interoperability them, which between fur- service
networks and of software infrastructures, behavior protocols. associated A significant and interoperability enabler for between dealing with ther includes these issues middleware is the Semantic data type Web and in
particular the Semantic Web Services paradigms, which allow description of service semantics and
behavior and rigorous reasoning about the former. Tackling both semantics and behavior
36 ERCIM News No. 67, October 2006