I wondered here how IBM would position DB2 with BLU versus Netezza. Please have a look before you go on… and let me admit here and now that when I wrote this I chickened out. As I sat down this time I became convinced that I should predict the end of Netezza.
Why?
In the post here Bob Picciano, the general manager of IBM’s Information Management Software Division, made it nearly clear. He said that DB2 BLU is for systems “under 50 terabytes” only because BLU does not cluster. I suspect that if IBM converted all of the Netezza clusters with under 50TB of data to BLU it would knock out 70% or more of the Netezza install base. He states that “most data warehouses are in the under-10-terabyte range”… and so we can assume that Netezza, precluded from anything under 50TB, has a relatively small market left. He suggests that Netezza is for “petabyte-size collections”… but as I suggested here (check out the picture!), Hadoop is going to squeeze the top away from Netezza… while in-memory takes away the bottom… and IBM is very much into Hadoop so the take-away will not require a fight. Finally, we can assume, I think, that the BLU folks are working on a clustered version that will eat more from the bottom of Netezza’s market.
We should pay Netezza some respect as it fades. When they entered the market Teradata was undisputed. Netezza did not knock out the champ but, for the first time, they proved that it was possible to stay in the ring… and this opened the market for Exadata, Greenplum, Vertica and the rest.
DB2 has limitations on the max table size, whilst in Netezza you can load 128Tb data in a single table on a N1001 full rack, and 192Tb on a N2001 full rack. This is a single data object in the database and can be manipulated and loaded as a single entity.
With a scan rate of 450Tb/hour now possible in the new range systems, Netezza for pure big data volumes will be ahead of the functionality of DB2 BLU, at the current time, and the capability for it to execute these large data volumes.
Whether this remains as the Pure System stack changes and matures will remain to be seen. There are a number of other offerings from IBM which conflict with their own offerings but each has a niche in business which differentiates it from other offerings.
Hi Gary,
Thanks for the input. I have no doubt that there are several interesting features in Netezza that are not in DB2… But the question is whether these features define enough of a market to make it worthwhile to keep Netezza going. IMO there is not enough room between DB2 at the low end and Hadoop at the high end to make it worthwhile.
Rob
I am not surprised to see that in the long run Netezza may have the same fate just like Red Brick and Informix database.
I agree with Rob’s analysis. 64TB table size limit is not a problem for most customers, and once IBM remove the current single node limit of BLU then 192TB and beyond will become possible.
Regarding Weidong Zhou’s comment about Informix, I assumed it was ironic until I saw that he is so firmly tied to Oracle products… and also the standard Oracle habit of spreading misinformation, clearly.