The announcement that EMC is to acquire Greenplum is not altogether a surprise—it was always a possibility that EMC might decide to move into the data warehousing space for itself, given the number of its existing partners in that space—if it chose to do so the main question was who it would choose to acquire. Now we know the answer.
In terms of the specifics of the deal, EMC will be setting up a new Data Computing Product Division of which Greenplum will be a part. The company will be introducing a reference architecture for Greenplum and it has also said that it will be making a physical appliance, though no details are available yet. The company has said that existing partnerships, on both sides, will remain in place. I doubt, however, whether the likes of ParAccel will be happy to maintain this relationship unless EMC can prove that it has clear lines of demarcation between its storage and data computing divisions.
Clearly, this acquisition has the potential to significantly ramp up Greenplum’s presence in the market, in particular in Europe, where it is currently relatively weak, and it therefore looks like a good deal for both points of view. However, this does not mean that there are not issues.
The first question is about technology. I’ve been on lots of calls of this type over the years and the people being acquired are always wildly enthusiastic about the new opportunities being given to their technological babies. Well, they would be wouldn’t they? However, the truth is that often these people walk away after their lock-in period is over: either because they want to spend all that money they’ve now got, or because they don’t like large corporate cultures or because they’ve have had another bright idea that they want to go and develop. What can easily happen then is that product development loses its innovative drive and becomes more of a me-too product. That isn’t to say it won’t do well but it can be a significant factor going forward.
The second issue is that Greenplum has always extolled the fact that it is open and not proprietary. With the announcement of reference architectures and appliances that looks like it will lose its force as an argument. In particular, if you are going to build an appliance then surely you want to leverage all the features of the hardware that you can to optimise performance. So even if it will not technically be a proprietary solution it will look like one and be perceived as such. Personally, I don’t place much emphasis on this argument but some people do and those people will be disappointed.
Thirdly, there is the question of marketing. Can EMC market data warehousing? Clearly, we don’t know the answer to that question: the proof will be in the pudding. What I can say is that I think Greenplum needs a shot in the arm when it comes to marketing. There was a time when it was the clear second place new generation vendor after Netezza but that has changed: it has been out-competed by Vertica at the high end, InfoBright at the low end, and out-MapReduced (to coin a phrase) by Aster Data. In particular, the company has failed to make much headway in Europe, though it has done well in Asia-Pacific (where it doesn’t face so much competition). So—a bigger sales force and more marketing dollars so see it pick up market share and make it a significantly more important competitor.
Finally, I should consider the implications of this acquisition. What does it say for the other newer vendors in the space? Well, it means there is one less exit strategy. Unless other hardware suppliers, like HDS, decide to follow EMC’s example we are running out of potential acquirers unless it is from existing players simply buying companies (and then letting the product slowly die) for the sake of building market share. From that point of view, and from the perspective that Greenplum should be a much more formidable competitor going forward, this has to be bad news for all of Greenplum’s competitors, even if it is good news for that company’s customers.