What's happening in the world of data warehouse appliances?
Releases and re-launches
Comment Quite a lot has happened in the data warehousing appliance (DWA) space recently, so here's an update.
Perhaps the most significant event has been the latest release by DATAllegro. In product terms the most interesting part of this announcement is the introduction of compression.
IBM, of course, introduced this last year and DATAllegro and IBM won't be alone in supporting this for long. However, the most interesting thing about compression is not just that it reduces the amount of disk you need, but it actually speeds up processing as well. This is because the overhead of having to de-compress the data is more than made up for by the gains in disk reads (or, in the case of IBM: the amount of data held in buffer pools).
But this release from DATAllegro is less about product features and more about technical partnerships. While it will continue to put together its own appliances, in future DATAllegro will be building these out of Dell processors (still based on Intel chips) and EMC disks rather than building the processors itself and using a rival disk supplier.
The most interesting part of this is the partnership with EMC and I understand that the EMC sales force will be entitled to full commission on sales made by DATAllegro. Thus, DATAllegro expects to be invited into more deals at an early stage, thanks to the involvement of EMC.
Talking about new deals, Greenplum has been approaching this from a different direction: both raising new venture capital funding and aggressively opening new offices. In particular, it has opened offices in the Far East, where otherwise Netezza has more or less had the field to itself.
As an aside, and depending on how you measure these things, it could be argued that Greenplum is the second leading DWA supplier after Netezza (and not counting IBM here), based on number of sales. However, most of these are fairly small, though Smart Communications has a 24Tb system, which is certainly respectable. Like DATAllegro, Greenplum is actively leveraging a major partnership, this time with Sun.
Another company that has recently announced additional funding is Vertica, which has yet to come to market, though it will be pre-briefing analysts in the near future. Paraccel, on the other hand, is a little further ahead, in that such briefings are already going ahead. What unites both of these products is that they both employ column-based databases.
This raises a number of questions: will they both go for the same market (Paraccel is targeting SQL Server users)? How have they overcome the scalability issues that the likes of Sand and Alterian were not able to overcome? Actually, I can answer this one, because it's obvious: by implementing parallelism.
Perhaps most interesting is the question as to how this impacts on the market: the likes of Greenplum and DATAllegro often claim that customers don't like Netezza because it is "proprietary" - in my view, this is not just spurious but is as much wishful thinking on their part as anything else, however that's not the point here - the point is that both Vertica and Paraccel will be introducing more "proprietary" solutions into the market: will that make them more or less acceptable? My guess is the former: if such solutions are commonplace then they cease to be an issue.
Finally, what of the other vendors in the market? Nothing much has happened with IBM since its roadshow towards the end of last year and nor have I heard much from Kognitio or anything from Calpont. Netezza will have a new release sometime in the not too distant future (but you could say that about anyone) and as for HP...well, the Neoview Platform is up there on its website but the company is hardly creating a storm with its marketing - I expect a re-launch at some point.
Copyright © 2007, IT-Analysis.com