Referral Management

Both SOA and cloud computing are on the rise. However, SOA, having been around much longer is climbing the slope of enlightenment. Cloud computing is still on its way up to the peak of inflated expectations. They are both given a 2-5 year to mainstream adoption. I think that cloud computing can learn from SOA, its more advanced technology cousin.  They are starting to become connected and one lesson is the importance of testing and validation as I wrote about recently, continuous testing and validation of all these combined data sources from the cloud mitigates the risk of invalid data getting mashed up in your workflows, and your business rules going off track. It will be interesting to see how the cloud adopts this SOA quality lesson.

What we've found is that there is a bigger constraint to establishing complex, critical business workflows, whether you are calling it SOA or considering a cloud model. The bottleneck for SOA is one of availability. What do I do if I cannot develop and test my intended functionality against the services I depend on? I need a virtualized model of the services, mainframes and data I depend on to build my own successful workflow.

Here again, SOA's need for Virtualization of application behaviors is a canary in the coal mine for Cloud concerns, where your dependencies are even more abstract and ephemeral. In relying on Cloud resources, there is a lot of variability of data - making it hard to hit development targets without a "virtual cloud" model in place. We'll address this topic in more detail soon.