Clearspace Recent Blog Comments Syndication Feed

RE: Four Fundamental Differences Between TwinFin and Exadata

dclegg@netezza.com — Wed, 11 Aug 2010 07:37:45 GMT

I like the restrained irony. And the vivid analogy, so I'll pursue it one stage further if i may.

I may or may not be a blacksmith (i don't have huge hairy forearms) but are you saying that Exadata doesn't work yet, but you confidently predict that it will?

RE: Four Fundamental Differences Between TwinFin and Exadata

Phil Bailey — Fri, 06 Aug 2010 12:35:47 GMT

Having experienced an implementation of an NPS system back in 2003, I can safely say that this eBook does a fantastic job of explaining all the things that I have struggled to convey myself without sounding overzealous. It's a simple case of old technology vs new technology for me - with an obvious winner in so many respects...

RE: Four Fundamental Differences Between TwinFin and Exadata

dbirmingham — Thu, 05 Aug 2010 13:40:51 GMT

Read the ebook - short review: "spot on".

One of the most dramatic realizations that the data warehouse owners find, sometimes like an epiphany, is that over time the warehouse has gradually assimilated complex structures, processes, workarounds and spot-solutions that begin to debilitate the warehouse. Then they want out. Teasing-apart this problem is usually not painful, just tedious. People avoid fixing it because it will take significant effort, the risk of breakage is too high, with no guarantee (on their current hardware) that such an effort will make things any better. Netezza has so much power in the machine, that turnaround time alone will accelerate their project to a positive closure, and often they will have lots of fun doing it. Who couldn't use a little more of that?

So the obvious question is: Why would someone embrace this complexity (e.g Exadata) in the core hardware architecture, from the outset, if the complexity is a known liability already? So your observation - simplicity is a huge asset - is not lost on those who have experienced this effect. Functionally speaking, most any technology allows us to define data structures and issue queries. But when billions of rows are staring at us, and the hardware architecture is a commodity construct that doesn't solve the physical performance needs, it reduces "purist" notions of functional elegance into a mind-numbing problem that can make intelligent people doubt their purpose in the universe. It need not be so.

The ebook offers a good overview with the right detail, and answers the primary questions in a way that invites the reader through a mental thought experiment: Netezza is answering questions that one may not have thought to ask, but here's why they are important, and why Netezza has answered them. For many new data warehouse initiatives, the principals have read the industry white papers and believe they have grasped their solution as an application, rather than what it really is - a living environment (with another set of rules). Once they dart down the path and stand up the warehouse, they think they are free to move on to the next project. Alas, if they have done it poorly, they will have more work before them to fix it. If they have done it well, they become a lightning rod for more work. At this inexorable juncture, they need a painless way to either reset or move forward, but it requires so much physical power that the solution stalls out and languishes. Netezza blows through this wall like a locomotive.

Periodically I get feedback from owners of competitor products who claim that their (favorite product name here) does everything Netezza does. I can usually challenge this assertion, but first ask the obvious: Does it do those things well?

RE: Four Fundamental Differences Between TwinFin and Exadata

T Dolan — Wed, 04 Aug 2010 22:27:51 GMT

Very interesting and objective commentary from the VP of Product Marketing for Netezza. I’m certain there were plenty of blacksmiths standing around Gottlieb Daimler proclaiming victory that he would never get the combustion engine to work…

RE: A Little Late Musing on the Greenplum Acquisition

Chuck McDevitt — Wed, 14 Jul 2010 07:59:20 GMT

Greenplum sells boths ways... Software only do it yourself, or several choices of prebuilt appliance where you get hardware and software together.

For now, EMC says they will continue this policy... Time will tell.

RE: EMC Swallows a Green Plum?

Michael Crane — Thu, 08 Jul 2010 14:29:25 GMT

Phil - Excellent, well written post!

One strong reason for the acquisition is GP's growth in China. Early on GP made a strategic gamble to pursue business in China even before it went after US and EU business. The gamble is paying off big! It's my view this this acquisition sets the stage for other, complementary EMC acquisitions in segments such as BI, ETL, ELT, replication ( EMC still has no logical replication capabilities), MDM, predictive modeling, data mining, and text analytics. One distinct possibility is acquisition of a specialized vendor of Hadoop tools ( nice to hear so much about Netezza and Hadoop recently) to supplement Greenplum’s strong MapReduce capability. I'm a pretty big fan of Hadoop, btw.

RE: Netezza Migrator™: A Point of Clarification

jfeinsmith — Thu, 01 Jul 2010 03:33:40 GMT

In addition, Netezza Migrator's version of EDB*Loader (Oracle SQL*Loader compliant API) has direct integration with nzload, allowing for full-throttle load rates into Netezza.

RE: Netezza Migrator™: A Point of Clarification

pfrancisco — Wed, 30 Jun 2010 12:27:11 GMT

@Steven: Since Netezza Migrator’s primary purpose is to aid companies in moving their Oracle proprietary PL/SQL, DLL, and DML to Netezza and not as an ETL tool, we understand the issue raised by the commenters on DBMS2 (Todd Fin and others) who asked about load rates. Migrator was designed to be suitable for many incremental data movement scenarios with the convenience of being driven via Oracle SQL and making use of Netezza's high-speed load/unload API.

When many terabytes of data movement are required from Oracle into Netezza, we recommend the use of third party tools capable of reading the Oracle data and/or log files directly since the Oracle SQL interface is typically far too slow for such tasks and not suitable for log-based CDC.

RE: Netezza Migrator™: A Point of Clarification

steven.rogers — Mon, 28 Jun 2010 01:08:05 GMT

Migrator could be a nice tool when it works as announcement stated. You did not say much about the Data migration though. I have read the referenced post in dbms2 blog about Migrator where it had raised valid questions.
Does Migrator eliminate that barrier finally or people will continue having a hard time dealing with it? Regards, Steven

RE: Boston Globe Names Netezza Top 100 Massachusetts Company!

Mark_w — Sun, 23 May 2010 14:06:31 GMT

Congratulations to Netezza for getting into the Globe 100. Great to see the next generation of Boston tech companies making their way to the top!