Tuesday, March 14, 2006

Somethings just don't make sense

There are many things in life that just don't make sense. For example, how can you have South Park without Chef? Whatever happened to Cold Fusion? Or the entire "plot" of Highlander II! But probably the one that gets to me at the moment (because it's related to my job) is how on earth can there be such a thing as a legacy ESB?!

Since it sprang to life, the ESB term has been used to mean different things to different people. For example, to some it's JMS with a few extra bells and whistles, to others it's Web Services, and to yet others it's the saviour of EAI. However, over the past couple of years, one thing that almost everyone can agree on is that it should be an infrastructure for SOA applications. Now we all know that one of the things SOA is good for is leveraging existing infrastructural investments: letting you use your old stuff in new and interesting ways.

So it always seemed to me that a product that helps you implement SOA applications (quick note, SOA is not something you can get out of a shrink-wrapped box), shouldn't be part of the legacy problem! And yet when I was with Arjuna, I came across a number of large and small companies that were already being forced to work with legacy ESBs! These were products (names removed to protect the guilty parties) that, once integrated into the user's infrastructure, couldn't cope with changes in user requirements a matter of mere months or years later; plus they were so intertwined with that infrastructure that they simply couldn't be removed (or it wasn't worth the effort). So these companies then had to rely on yet more ESB implementations to talk to their legacy ESBs! And the vicious circle went on. Not exactly a good return on your investment!!

Now I'm not saying that we're going to reverse that trend, but I'd like to think we'll try not to fall into the same holes that others have. However, it's still an interesting contradiction, sort of like military intelligence.

No comments: