Ronan Bradley's FinanceTech Directions

I don’t know who first said that you learn most from your own mistakes but they missed the follow-on: Learning from other’s mistakes is a lot less painful! In that context, the FBI is continuing to give us all learning opportunities as they attempt to modernize and integrate their complex systems.

I wrote about the FBI Virtual Case Manager debacle a while back: After they ditched their first attempt at solving this problem (at a cost of only $104m to the US taxpayer), they went SOA and created project Sentinel . However, since then they seem to have gone through a rapid and painful learning curve on SOA. Jerome "Jack" Israel, the FBI's chief technology officer puts it like this:

“We've gone from maybe hyped-up about it to the cold realization that 'Hey, this SOA is a lot harder than industry is making it out to be,”

In particular, the discontent seems directed at vendors who think that all it takes is some minor product modifications and a new marketing gloss on their old proprietary (often EAI ) products to magically make them SOA platforms. Dennis Nadler, a former chief engineer at the Defense Information Systems Agency, puts it rather bluntly:

"On the edge, sure, it may look like an enterprise service bus . . . but behind the scenes, it's all proprietary,"

Dragging old products into new markets is hardly a new phenomenon. However, SOA is particularly at risk because the focus on architecture has allowed some vendors to suggest that their own pre-SOA products can cut the mustard. They mostly don’t: Take note IBM, Microsoft et al – renaming products or simply declaring them ESBs or SOA platforms isn’t enough.

To finish with some good news, all of this doesn’t mean that the FBI is facing Virtual Case Manager debacle v2 – rather they have focused on the best returns on investment – which in their situation is integration with regional data exchanges (which sound similar to the UK CJIT approach).