Cloud Computing and SOA Convergence in Your Enterprise: A Step ...
Cloud Computing and SOA Convergence in Your Enterprise: A Step ...
Cloud Computing and SOA Convergence in Your Enterprise: A Step ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
210 Chapter 11 Mak<strong>in</strong>g the Move to <strong>Cloud</strong> <strong>Comput<strong>in</strong>g</strong><br />
While prov<strong>in</strong>g the performance models, you should leverage performance<br />
test<strong>in</strong>g, determ<strong>in</strong><strong>in</strong>g how well <strong>and</strong> how fast the holistic architecture,<br />
both on-premise <strong>and</strong> cloud-based, will support the bus<strong>in</strong>ess. Moreover,<br />
measure how the system performs dur<strong>in</strong>g an ever-<strong>in</strong>creas<strong>in</strong>g storage, database,<br />
process, <strong>and</strong> service-process<strong>in</strong>g load. If they do not perform well, f<strong>in</strong>d<br />
out where the bottlenecks are: <strong>in</strong> the network the database services If necessary,<br />
work with the cloud provider to correct them.<br />
Select Target Platforms<br />
Once we go through all of the analysis, <strong>in</strong>clud<strong>in</strong>g a service-, process-, <strong>and</strong> datalevel<br />
underst<strong>and</strong><strong>in</strong>g of our problem doma<strong>in</strong>, <strong>and</strong> have considered both security<br />
<strong>and</strong> governance, compiled a list of c<strong>and</strong>idate systems, <strong>and</strong> completed the<br />
validation test<strong>in</strong>g, it is time to pick the cloud comput<strong>in</strong>g platforms.<br />
This step is pretty easy consider<strong>in</strong>g that any issue around the platform’s<br />
ability to meet the requirements of the architecture, <strong>and</strong> thus the bus<strong>in</strong>ess,<br />
should be well known <strong>and</strong> understood by now. Also, keep <strong>in</strong> m<strong>in</strong>d that it is<br />
more likely that the f<strong>in</strong>al selection of the suite of target cloud comput<strong>in</strong>g<br />
platforms is very different from what you first envisioned, but if you did your<br />
homework <strong>and</strong> followed each step <strong>in</strong> this book, they should be the proper<br />
platforms for your architecture.<br />
Also worth mention<strong>in</strong>g is the ease of switch<strong>in</strong>g, or should we say, the relative<br />
ease of switch<strong>in</strong>g, from one cloud comput<strong>in</strong>g platform to another if for<br />
some reason you make the wrong call, or more likely, if some bus<strong>in</strong>ess event<br />
occurs with the cloud comput<strong>in</strong>g platform, such as the cloud comput<strong>in</strong>g<br />
provider go<strong>in</strong>g out of bus<strong>in</strong>ess or a merger or acquisition changes or removes<br />
that platform. Of course, this depends on the cloud comput<strong>in</strong>g provider you<br />
selected, its use of st<strong>and</strong>ards, <strong>and</strong> your ability to f<strong>in</strong>d another provider that<br />
offers similar characteristics <strong>and</strong> features.<br />
The bus<strong>in</strong>ess issues are more important if you are look<strong>in</strong>g to create an<br />
<strong>SOA</strong> us<strong>in</strong>g cloud comput<strong>in</strong>g, s<strong>in</strong>ce that scenario depends entirely on the<br />
cloud provider to stay <strong>in</strong> bus<strong>in</strong>ess <strong>and</strong> keep up <strong>and</strong> runn<strong>in</strong>g. You need to carefully<br />
consider<br />
The viability of the provider <strong>and</strong> the likelihood that it will provide ongo<strong>in</strong>g<br />
support for your cloud comput<strong>in</strong>g platforms.<br />
The provider’s ability to recover from hardware, software, <strong>and</strong> network<br />
failures, dynamically <strong>and</strong> with m<strong>in</strong>imum downtime.