Netezza rolls out its compression story
The proximate cause for today’s flurry of Netezza-related posts is that the company has finally rolled out its compression story. In a nutshell, Netezza has developed its own version of columnar delta compression, slated to ship May, 2008. It compresses 2-5X, with the factor sometimes going up into double digits. Netezza estimates this produces a 2-3X improvement in overall performance, with the core marketing claim being that performance will “double” from compression alone.
To understand how these claims go together, it helps to recall that for Netezza the performance gating factor is usually I/O from disk. FPGA? Oh, we have unused gates sitting around. Inter-node communications? No problem to date, and if there ever is one we have sockets ready to go that would double bandwidth anyway. CPU? Oh, we use small, low-power ones, that occasionally send random messages telling us they’re bored to tears … All kidding aside, probably CPU is sometimes a gating factor too; otherwise, the relationship between compression and overall performance would be even more direct than it is. But for the most part, faster data can be gotten off of disk, the faster a Netezza system answers queries. And since data comes off at close to a fixed speed, the less data volume there is to fetch, the better.
Frankly, I fail to be convinced that this announcement is much more than me-too catchup. DATAllegro has boasted similar compression for a while. Vertica not only has similar or better compression, but actually processes its data in compressed form. Infobright has mind-boggling compression figures, which is why they can be in business without having shipped an MPP version yet.
But after they get blue in the face from arguing against that judgment, Netezza’s marketing folks quickly catch their breath and turn the argument around. Their sales figures clearly suggest that they’re winning a lot of deal against the smaller startups, no matter what the smaller guys like to claim.* Well, if they have a big performance gain coming down the pike soon, won’t that make their competitive position yet stronger? I guess we have to see what the competitors come up with over the next half year themselves.
*Note: That kind of disagreement isn’t necessarily a function of deception or delusion. It can also be a sign of an industry that sports well-managed sales forces, each of which is good at qualifying deals and walking away early from the ones it’s apt to lose.
Comments
Leave a Reply