Justifying Service Oriented Architecture in economic downturn
by Goutam Mukherjee
As we all know, the global economy is going through a crisis. It started showing its initial sign from mid 2007, continued through 2008 and is staring at us even as we move into 2009. Very few experts are willing to take a risk on forecasting on how long this will continue. Many are even tagging this as the biggest one since the Great Depression. So now the big question is how are the enterprises going to tackle this situation and beat this downturn? How can they cut costs, remain agile and stay competitive? And in this context, the relevant question from the SOA practitioners’ viewpoint is how the corporations can leverage SOA in the best possible way to achieve the benchmarks that they are striving for even in this down turn.
Adopting SOA always has some upfront costs. So in the first glance it may look like that SOA has an inherent flag to tend towards a NO-GO. That indeed is the case if you focus on the traditional way of looking at an IT initiative from the Costs and Benefits standpoint. SOA has a high upfront cost and a gradual accumulation of benefits and so it becomes very hard to justify – especially in these tough economic conditions. For this specific reason, we should look at the approach from a quantifiable ROI angle.
When we talk about ROI, two most important components come out – "Investments" and "Returns". The "Investments" for a typical SOA initiative consist of – a. Assessment / b. Infrastructure / c. Implementation / d. Governance. On the other hand, the "Returns" from a typical SOA initiative comprise of – a. Business Agility / b. Service Reuse / c. Infrastructure Reuse / d. Integration Component Reuse / e. Pre-existing Governance Model.
Once we have defined the major components and the corresponding constituents for coming up with the quantifiable numbers, we can focus on minimizing (or smart-sizing) the investments and maximizing the returns. The measurement model for the investments should clearly capture each constituent, the factors influencing the constituents and very importantly the prioritization of the constituents. For example, instead of investing in costly ESBs for a medium to low sized SOA initiative, can we look at any alternative of custom-built framework that can be built as a part of any current web-enabled infrastructure and later easily plug that into an ESB? Or take another example, do we need a full blown governance model or can we take an incremental approach to that?
Similarly, the measurement model of the returns should clearly capture all the constituents and the factors. We should remember here that all of them may not be able to be quantified in the beginning and they can only be benchmarked based on future experiences. One clear example of this will be ‘business agility’. There should be some factors built into the model for constituents like these. Another important factor is that business prioritization of the benefits should be factored into calculating the effective returns.
Another critical factor in this context will be to actually measure the authenticity of the ROI numbers in an incremental manner with each application moving into SOA umbrella. A valid Chargeback Model will be essential for this purpose. Chargeback Models can be based on multiple formats. Few of the popular ones are - usage based, subscription based and peak-level based. Based on the business context, one of these or a combination of these should be decided upon and the projected ROI figures should constantly be monitored against the actual figures.
So in a nutshell, a pragmatic approach to SOA with intense focus on quantifiable "Investments" and "Returns" coupled with a constant check against the real operational figures will definitely help a long way in justifying SOA - and yes, even in these troubled times.
