Nearly one-third of the total organizations that are pursuing SOA are using an enterprise service bus (ESB).
This is where we can see that for one third of the enterprise SOA starts with ESB and it should. What about the rest. As I analyze this, following questions come to my mind.
- Is the ESB market place confusing?
- Is it lack of success stories?
- Will the ground realities within an enterprise that would take some time to address?
- Is it the lack of knowledge on ESB?
Let us go through these points.
- Is the ESB market place confusing?
ESB in current market place has different implementations, at least two; one that comes with a service container and one without. Though, ESB by its definition does not mandate a service container embedded to the product, yet is a compelling option to look for- as pitched by some vendors . A variant of this has been introduced by vendors those who have come up from the application server market and entered into ESB space. I see this more of retaining existing investment and to counter the marketing slogan “ESB with a service container” started by Sonic and followed by TIBCO. Though a compelling argument can be started on which option is better (the one that can host only J2EE based service and integrate other technologies; or the one that can host services of multiple technologies along with integration capability) yet there definitely exists two distinct options for a prospective buyer. Certain products like WBIMB do not have a service container and is promoted as "Advanced ESB". Are these not enough to confuse a sensible buyer?
Let us add more to it. If we take into account that many see BPM (that many people relate to SOA)implementation as an entry point to SOA, we see vendors coming up with a new product on top of another every now and then . Is this not enough to make a buyer think that these products are yet to mature? In my opinion consumers are stuck between brand and value. - Not many success stories
Let us have a look at one of the key value driver; “Reusability”- this can be leveraged if services are used. People from different divisions (within an enterprise) obviously will have different requirement from the same service. When they say requirement, more often than not it translates to business rules. Whether this comes from a pure political point of view or not that is arguable. Irrespective of anything, it is better to design services those are configurable from business rules point of view. Unfortunately, I have not come across many service designs those promote this kind of design.
Customers adopting a new technology look out for success stories, But unfortunately human tendency make people to look at the downsides (which in this case are failures) also to come up with a positive view on the technology. Does the factor mentioned in previous paragraph is the reason why we do not hear many success stories in the ESB world? - It will take time to address ground realities within enterprise
SOA by its very nature is pervasive and demands a fall of “The great Wall” that exists between organizational groups. In many organizations each group maintains privacy of their systems and strategies. A sensible CIO quickly understands this as a constraint because the systems that are candidate for SOA transformation primarily are the systems with high criticality. With this, the CIO has two challenges at hand; One, winning the mindset of maintaining status quo by the group; Two, doing so without disruption of major systems. Some find a solution to this in a stringent way. But, majority of them take a two step approach. Step one- Enabling web services; Step two- introduce architecturally elegant components e.g. ESB.(it is worthwhile to mention that apart from achieving the known benefits, the aim of step one is to be aware about what is where). In my opinion above mentioned point has elements that contribute to why still the significant two third are in SOA but not yet pursued ESB. There exists another breed of CIOs who has grown up with the culture of satisfying the subordinates and has climbed the hierarchy because of their popularity. They are the one who adopts SOA at its basic (may not be with web services) for mere management eye wash. - Do buyers lack of knowledge on ESB?
There are three kind of enterprise. One, who do not know; Two, who know but are confused; Third, who know but want to maintain status quo because they hold the key to the current system; At this age where each alternate day either a system integrator or a product vendor pitches its capability to the enterprise it is difficult to comprehend the first possibility for all but few. My point of view on second and third option is being addressed in the previous sections.
