Denis’ Blog

Using open source as a distribution model is a very efficient strategy when well executed. Alfresco in the DMS/ECM world has been taking incredible market shares in just 2 years, as MySQL did in the RDBMs world. Many domains have now their open source challenger: Montavista in the embedded World, Digium with Asterisk in VOIP/e-pbx, Hyperic and GroundWork in the Monitoring, Funambol in the mobile/telco industry, Pentaho in BI or OpenBravo in ERP.

In the Enterprise search space, two companies have demonstrated a clear strategy leveraging an open source model:

• IBM, with Omnifind! Yahoo Edition are also leveraging open source with an entry-level solution. But like IBM is doing with Websphere Express (Geronimo based) and Websphere, that is not a real offer? That’s a teaser to have a foot in the door and to push Omnifind.
• More interesting, SearchBlox is proposing a great value proposal around Lucene and the Carrot Clustered Search Engine. But Lucene lakes enterprise grade features like mapped security and a whole bunch of connectors.

And that’s where come Dieselpoint. Dieselpoint has announced today the launch of OpenPipelines . The future will tell if they are serious in their execution. They know well the enterprise search market and will propose a nice mix of features. But this move from Dieselpoint seems to be a defensive one. In a challenged market (Google Mindshare, FAST/Microsoft/Search 2008, Autonomy leadership, dozens of serious competitors), and given that Dieselpoint has already a technology, this move is not a full open source strategy like Alfresco’s one. It is maybe more a move like the one of CA opensourcing Ingres: I am challenged, I am loosing market share, let’s rejuvenate the beast with an open source sticker :-). I am probably too sarcastic. Intalio in the BPM world has demonstrated that you can leverage open source to go from a 100% proprietary strategy to a “blended” one like our friends of BEA used to say. Wait & See.

So we stay with an open question: who will propose an enterprise grade solution with a full open source distribution model. Enterprise Search is very similar to RDBMs:

• the dual-licencing strategy works, with the right licencing an enterprise search open source company can boost OEM.
• it is a product thousands of consultants and integrators are looking for every day for they application and projects. The bottom-up, low cost sales strategy well described be Matt from Alfresco will perfectly work.
• Established leaders (Autonomy, Google) will be more than challenged with the move

The thing is that in Enterprise Search, unlike DMS or RDBMs, there is key Intellectual Property in the motor. Algorithms are key in the value proposal. At least for now. So moving to open source can be touchy. But given the market trend to commoditization and the strong position of IS leaders (IBM, Microsoft and Oracle), I am convinced that we will see in the next 12 months some pretty aggressive move in enterprise search space towards open source distribution model.

Now the question is. Who will do it? Coveo? Exalead? Intellisearch? Sinequa? Vivisimo?

Take your bet.

I will hold two presentations in Paris next Wednesday. One will be on Enterprise 2.0 and the other on Enterprise Search. Feel free to join, the access is free if registered. I will present our return on experience on Enterprise 2.0 especially with social tagging and bottom-up self organizing communities of practices. For entreprise search, I will present the state of the art of the new holy grail: finding the information you are looking for.

Endeca announced today that Intel and SAP have invested in Endeca. So with these new information, I modify my bet:

• SAP will buy Endeca by the end of 2008. Endeca will not go IPO.
• Autonomy will be acquired by IBM.
• Oracle will not acquire anyone.
• Exalead and Vivisimo will go IPO.

Your bets?