I’ve had a day to ponder the deal and collect opinions around the office and I don’t feel better about the situation. EMC signing a new agreement with Autonomy was not a total shock. It was a measured business response to the changing ecosystem of enterprise software partnerships. I’ve been in many conversations since the FAST/Documentum integration was released where people complained about the early performance problems and complexity it introduced. The flexibility of the approach made sense to everybody, but for the thousands of poor souls who had (and still have) to upgrade to the platform and introduce N number of additional processes to monitor, migrate customizations, change ops procedures, et.al. – it was daunting if not downright overwhelming.
I can remember complaining to my sales rep when D5 came out – “Why doesn’t EMC just buy a search engine and get it over with already?” They’ve got enough change between the sofa cushions to pull off a deal. I feel the same way about the app server and it was a shame that the BEA deal fell through.
The argument has been made that the current plug-able architecture will allow for a simpler transition to other search engines. Interestingly enough – I know it has already been done. (hint: check out the search implementation of the Documentum OEM edition – starts with an ‘L’). My problem is not the technical viability – I am sure it can be made to work eventually. My problem is the operational impact and the burden on the community it will create. If I am a customer struggling to get FAST to work the way I want – my incentives to upgrade to D6 at all are reduced when I have to do it all over again with a different one on the next release. It may be good for me as a consultant, but as an advocate for the user community – it’s quite a dilemma.
This is a case of business relationships rather than user needs driving product architectures. It’s the real world but too much of it could eventually create a backlash. This scenario begs the question, when do core changes in infrastructure requirements begin to affect the viability of a product at a customer? Moving from Tomcat to BEA as the embedded app server has MANY welcome advantages, but whole communities of administrators have to retool to support it. There is talk of changing this again. Introduction of the messaging server was a great enhancement – this is yet another thing people have to deal with. If they force a change in the search infrastructure again I am afraid some Documentum admin somewhere will snap and we’ll be dodging bullets from a bell tower. Put simply, when upgrades exceed the level of effort for initial deployments, customers will eventually rethink their choice for the CMS and open the door to arguably less capable products that appear more approachable.
For EMC – I agree the acquisition of FAST and a theorized change to Autonomy is a wash technically speaking. I call this theorized because I’ve seen nothing in print about how EMC proposes to handle a transition, or if one is even mandated. Past behavior however suggests that the next major release will drive a replacement of FAST with the new OEM agreement and you should begin thinking about the implications to your organization sooner than later. Your mileage may vary..