In the interest of full disclosure my employer is Netezza and at one time it was Oracle. I worked with Oracle software for over 20 years and a little about 15 years ago became very involved in Data Warehousing mostly with database marketing solutions. In 2002 I found Netezza and fell in love with the technology because it was simple and allowed me to focus on business value and data assets rather than a constant battle with performance and tuning.
Before I go into my feeling of Greg Rahn’s response to the Netezza eBook let me make something perfectly clear. If I were hired tomorrow to architect a very high volume OLTP system, Oracle would definitely be considered. These systems are highly complex and require the execution of millions of short transactions per day by millions of users and all the features that make Oracle so complex are needed to accommodate these transaction volumes and required response times.
These complexities are greatly reduced for Business Intelligence and Analytics if something is designed to specifically address them like Netezza. The complexities are an inherent part of Oracle because it was not designed for Analytics and therefore needs to be retrofitted to attack this specific area of processing that is becoming more and more important in the digital age compounded by the tremendous growth of data.
Now lets get the BLOG by Greg Rahn in response to Netezza eBook and Daniel Abadi - Defending Oracle Exadata. It is very clear to me in his response that Greg looks at these overly complex activities for implementation as a benefit and I guess most people that make a living dealing with those activities have blinders on when it comes to business value and benefits to the bottom line. For me the simplicity of Netezza was a godsend to the business and our analytical challenges. It freed up valuable resources to focus on driving business value and not putting myself or others in a position where because of the complex technology we were needed to keep shit running. This has absolutely no monetary or business value.
You need to be an engineer with a tremendous understanding of both infrastructure and Oracle software in order to make it work. With that said, I need to ask how many companies have access to a Greg Rahn who seems to be very talented that can handle these complexities and if you don’t what will it cost you to get them on board?
In my previous life I used Oracle and it was the complexities and need for very experienced people that led me to using Netezza for Analytic solutions. The DBA staff could now focus on getting real value out of the data to help improve business and the bottom line. My head started to spin as I went through the number of manuals just to understand the best approach to partitioning, indexing, striping and a slew of operational and physical design activities that need to be understood for producing an optimal configuration. Never mind the array of initializing parameters that also needed to be understood.
Lets just take one comment that Greg made:
Bottom line: There are numerous ways that Exadata can restrict the data that is set to the database servers and it’s likely that any query with any predicate restrictions can do so. Certainly it is possible even with the analytic question that Netezza mentions.
To me this says it all. Numerous ways implies that there is no simple right way to do this but is dependant on many other technical decisions related to physical design. Netezza is a true appliance and there just isn’t a hell of a lot to think about. I guess it all comes down to simplicity, cost and performance.
What I read in Greg’s response to Netezza makes me feel a whole lot better about my decision to move away from Oracle for DW and analytics and never look back. We used Oracle where it made sense for the business and not for the benefit of engineers.
“Fools ignore complexity. Pragmatists suffer it. Some can avoid it. Geniuses remove it.” – Alan Perlis
FirstLiberated

Follow FirstLiberated on Twitter
