« Gartner on search and enterprise social networks | Main | Is Microsoft joining the Lucene/Solr dance? »

April 25, 2012

Vivisimo: Another one bites the dust

Earlier today, IBM announced that it was acquiring Vivisimo for an undisclosed sum. Now the tough question: what’s it all about? For the answer, let's take a quick trip to the early years of the decade.

Vivisimo was founded in 2000 out of Carnegie Mellon University. The first time we saw them, in 2004, they were marketing 'Clusty', a web clustering product that could examine huge numbers of web pages and then associate - or cluster - documents on specific terms. They also had some really strong federation capabilities built in. And the product was highly scalable. In fact, Vivisimo had great success in a number of huge government sites including the US Social Security site, FirstGov, the Defense Intelligence Agency, and commercial sites such as Ely Lilly. One thing all of these sites have in common? Lots of data. We have a term for that now: 'big data'.

IBM has made huge investments in open source search over the last 10 years, specifically yin Lucene/Solr. Hadoop is the Apache answer for big data, and trust me; Hadoop is a hot topic this year.

What does Vivisimo bring IBM? Well... for one thing,  clustering algorithms (and probably patents); a reputation for being able to handle huge data sets; and federation.

What should Vivisimo customers do now? Well, based on IBM's strong customer ethic, I think the answer is "don't panic" = do nothing for now'. Assuming Velocity is working for you, this acquisition should cause you no concern.

If you are evaluating Vivisimo, that's a bit more difficult. Some acquisitions like Verity's acquisition by Autonomy resulted in a wholesale replacement of the platform. Some customers made the switch early on and were happy; others fought to make IDOL work like K2, even with the 'compatibility mode; and never succeeded. You'll also remember that Microsoft, after  acquiring FAST Search, dropped the entire non-Windows platforms a year later which impacted upwards of 70% of the FAST  installed base.

If you are willing to acquire a platform for a couple of years and see what happens, go for it. You may look back and discover you made the right choice. On the other hand, former President Reagan had a saying: Trust, but verify". You might take a look around to see what platform is right for you now and into the future.



TrackBack URL for this entry:

Listed below are links to weblogs that reference Vivisimo: Another one bites the dust:


I've been out of town and so didn't see this post until after my own went up at blog.museglobal.com. I agree the pool of independent federated search vendors is diminishing, but feel the main thing IBM gets out of the Vivisimo acquisition is access to the focused, deep extraction capabilities of Velocity (or any other federated search engine), that can be used to increase the quality and freshness of the data that is fed to their Big Data engine. Working on the GIGO principle (in reverse) this should bring an improvement in their analytics. And don't forget how well Watson will perk up with more focused, richer data.

As for existing and prospective customers, I think it has all been said above - trust, but verify, and look around!

p.s. I will go and make the link to this in my blog.
Thanks Peter; excellent observations. Look to join you soon for lunch perhaps!

One of the hallmarks of Vivisimo has been fairly rapid startup, fast indexing throughput and scalability. As long as IBM does not make it MORE difficult to implement (to increase their service revenue), the engine, out-of-the-box still has lots of potential. Verifying that the business relationship is going to be easy and not unreasonably priced would be my principal concern for SMBs.
Good point Lynda! Remember what happened to Verity and FAST ESP users though... hopefully IBM will be more gentle with Vivisomo's partners and customers!

Any acquisition is risky for customers; support may gradually drop away, you may be pushed onto different and innappropriate alternative products, pricing models may change. This has certainly happened in the past. Without the source code, you have no control of the process.
Absolutely right Charlie. If only there was some way for the average enterprise search customer to gain access, eh?
Thanks for stopping in! /s/Miles

The comments to this entry are closed.