Project-Oriented SOA

Recorded at:

Can you please post the PPT, or a link to the PPT. I think there must be an option to view both PPT and the video (@ side corner) in the FULL Screen View.

Thanks Siva Prasanna Kumar .P

First of all I don't understand the need for Service Funding explicitly. Any service exposed by a system and its project is a part of its lifecycle with requests from other systems that help it evolve. I disagree with central funding. Central concepts will soon run into Single Point Of Failures. This inline with Service Oriented Application Development and Enterprise Mashups. IT and Business's are obesessed with having a dashboard which displays the world to them. Why does one central team have to know everything and control everything. I think this is the biggest flaw that has come down from Corporate Governance to IT Governance and has made its way as SOA Governance.

I also don't agree with the concept of SOA Governance. 80% of SOA projects fail and people are talking about how it takes 10-15 years to get a SOA Transition right, no it doesn't usually you get it on the right track within 5-6 months. It is not poor governance that causes SOA projects to fail but rather governance itself.

While a project goals are mostly tactical SOA goals are strategic and I have been using the Rincci SOA methodology with virtually 100% success rates with our Fortune clients. Which marries tactical goals of system evolution with the strategic goals of SOA evolution. WIthout any Central Governance. It works the best with distributed agile teams which are used to change and evolution of systems and use a security net for refactoring, with unit tests for services at the producer end and intergration tests at the consumer end. This is in direct contrast to what you are proposing i.e. Service Lifecycle should be centrally managed and central team should be responsible for service identification, lifecycle management and pipelining acttivities. This is the most anti-case for evolutionary design. This seems like everything we used to do 15 years back with waterfall methods and upfront design and architecture. As an Agilist I have reservations and with our SOA experience I think this is why most SOA projects don't work.

Look at it this way. Its an ecosystem you don't provide services you don't get services. Its an ecosystem that can be self evolving without central control. The most scalable systems are without one Truth, one central controller or one central oracle that knows everything. The analogies from software systems apply to organizational structure as well. As Software systems mimic tasks done by people albeit. in an automated fashion.

I would like to recommend reading my book "Essays on SOA & EAI - A Pocket Guide"

Respectfully yours.
Aditya Yadav
adityayadav DOT com
Author: Essays on SOA & EAI - A Pocket Guide
adityayadav.com/EssaysonSOAandEAIAPocketGuide.aspx

for Agile approach ...
I agree on the safety net (or CI) aspect, and promote and scale well the work of your distributed architects.
To amend video, I would like to recommend browsing through my (openly available) slides from Orlando ScrumGathering "SOA and Scrum"

Cheers
Thoralf J. Klatt
www.linkedin.com/in/tjklatt
www.slideshare.net/tjklatt/klattagilecreationof...
www.scrumalliance.org/resource_download/613