Question: Netezza vs. Vertica
I received the following question last week from Mike T:curious as to your evaulation between vertica and netezza. i have used netezza for several projects at a couple of clients.
My response, unfortunately, is that I can't quite say yet.
I've been using Netezza for about two years now, and I'm consistently impressed with the people and the performance, if not always the robustness of the product. Now, my software isn't perfect either, so I'm not really throwing stones here (or at least I'm trying not to), but I think it's fair to say that there are so many 'moving parts' in Netezza's system - hardware, software, networking, SQL, APIs, etc. - that it's unrealistic to expect all the pieces to play nicely together all the time. I am thoroughly impressed by what they've done, but let's not claim that it's perfect.
At least not yet. I've always been more impressed with the people than the product - which is saying something, given how much I like the product - so I'm confident that in time the kinks will indeed get worked out.
As for Vertica, my experience has been fairly short-lived (a couple months) and has been disappointingly hands-off (by my choice, not theirs): I've read lots of documentation (you should start here) and provided data sets/query workloads for benchmarks, but I have not yet installed, configured or directly used a Vertica system. The architecture itself is so blindingly obvious that one can't help but believe in it, and the performance results I've seen have backed that up. The cast of characters over at Vertica certainly makes for some good PR as well, so I'm confident that we'll be hearing a lot about them and a lot from them in the future.
So, if you don't mind some unrequested prognostication, here's how I see this playing out: Vertica will be successful because Netezza paved the way for non-traditional database systems. But their success will be somewhat niche - a very large niche, but niche - partially because the system is spectacular for some things but absolutely hideous for others. More importantly, however, Michael Stonebreaker wants all databases to be specialized, and thus having a database that was good for everybody would defeat his larger purpose. So a very large niche it will be.
Netezza, then, will become the general purpose large-scale data appliance of choice. It's overall performance profile, lower administration costs and generally stellar performance characteristics will land it in a somewhat-orthogonal niche to Vertica. A broader niche, but still a niche.
Oh, and lack of the word 'warehouse' in my description of Netezza is not an accident. Part of Netezza's success will come from moving beyond the data warehousing space. Stay tuned.
At least not yet. I've always been more impressed with the people than the product - which is saying something, given how much I like the product - so I'm confident that in time the kinks will indeed get worked out.
As for Vertica, my experience has been fairly short-lived (a couple months) and has been disappointingly hands-off (by my choice, not theirs): I've read lots of documentation (you should start here) and provided data sets/query workloads for benchmarks, but I have not yet installed, configured or directly used a Vertica system. The architecture itself is so blindingly obvious that one can't help but believe in it, and the performance results I've seen have backed that up. The cast of characters over at Vertica certainly makes for some good PR as well, so I'm confident that we'll be hearing a lot about them and a lot from them in the future.
So, if you don't mind some unrequested prognostication, here's how I see this playing out: Vertica will be successful because Netezza paved the way for non-traditional database systems. But their success will be somewhat niche - a very large niche, but niche - partially because the system is spectacular for some things but absolutely hideous for others. More importantly, however, Michael Stonebreaker wants all databases to be specialized, and thus having a database that was good for everybody would defeat his larger purpose. So a very large niche it will be.
Netezza, then, will become the general purpose large-scale data appliance of choice. It's overall performance profile, lower administration costs and generally stellar performance characteristics will land it in a somewhat-orthogonal niche to Vertica. A broader niche, but still a niche.
Oh, and lack of the word 'warehouse' in my description of Netezza is not an accident. Part of Netezza's success will come from moving beyond the data warehousing space. Stay tuned.
Trackbacks
Trackback specific URI for this entry
This post over at DBMS2 got me thinking - is Netezza's materialized view functionality an unintended - or maybe just un-promoted - column store?
NZ's materialized views work by extracting a column or columns from a table and storing them separately, in
Professor David DeWitt posted the following comments in reply to my Netezza/Vertica predictions. I've reproduced them here for visibility, and so I can respond to them more properly.
It's important not to misinterpret Mike Stonebraker's comments about
Well, the word is finally out: Netezza is making user-defined functions available. I've been waiting to blog about this for a while, and I'm rather ticked off that Monash and others beat me to it. Oh well, guess that's what I get for not reading all NZ
Mike's vision of the future in built upon the belief that: 1) the market will trend towards superior performance through specialization, and 2) that different data sets -- or maybe even the same data set -- will be managed by different types of database systems depending on the target application. For example, a stock tick stream might be managed initially by a stream database system in order to provide real time answers to continuous queries, but then it will be loaded into a data warehouse where it will be managed by a system like Vertica or Netezza for historical analysis.
It is a mistake to label a system that is tuned for a specific function such as data warehousing as serving a "niche" market. Just as Oracle and SQL Server are designed for transaction processing, Vertica, Netezza, Greenplum, and Teradata are built specifically for handling ad hoc queries on multi-terabyte data warehouses. And even within this latter space, the different products are specialized in different ways. Rather than relying on the indices used extensively by database systems targeted for the OLTP marketplace, Netezza focuses on very fast sequential scan performance using custom hardware. Vertica takes this specialization to the next performance level by observing that a database system based on a column-oriented architecture is much better for OLAP applications than one that uses a row-oriented architecture -- an architecture designed originally for OLTP applications.
[David DeWitt, professor of computer sciences at the University of Wisconsin and Advisor to Vertica Systems, Inc.]