DB2 Analytics Accelerator vs. Competitor
Attractive hazards: Data mart queries
What’s that sucking sound? It’s the sound of ad-hoc, monster queries hitting your operational data, robbing your core business transactions of the processing cycles they need to generate cash and keep your customers happy. Data marts have long been the answer to making mainframe data available to business departments and analysts, without exposing your core, money-making systems to unruly queries and runaway resource consumers. However, this approach is not without its hazards.
Data marts generally start out as a great way to provide answers to the business. They are attractive because they are quick to build, fulfill an existing business need, and only affect people inside their user community. Because of their quick success, they tend to multiply like rabbits. But they quickly become complex to manage, create data governance challenges, and in aggregate, increase costs.
Efficient alternatives: Analytics and Transactions on one system
A single, hybrid data server composed of the System z platform and the new IBM DB2 Analytics Accelerator, powered by IBM Netezza technology, can provide the same characteristics of a data mart, without the complexity, data challenges and cost. Complex analytic queries can run 100x faster and sometimes even 1000x faster on System z with the accelerator, without changing your applications, moving your data, or managing satellite data marts. Plus you can get up and running in days and support thousands of concurrent users. You can both eliminate unnecessary costs and still take advantage of the reliability, availability and security of the mainframe.
IBM offers a broad portfolio of hardware and software that together creates an enterprise data hub, capable of handling the entire spectrum of workloads of OLTP applications, transactional analytics, operational BI, and complex queries. IBM delivers analytics across the System z stack.
DB2 for z/OS and the IBM DB2 Analytics Accelerator are tightly integrated, via a private, dedicated network and special code that is deeply engrained into both the DB2 and Netezza software stacks. This combination offers, through a single logical server, the best in terms of large scale Symmetric Multiprocessing for highly concurrent, quick running queries in DB2 for z/OS, and an appliance approach to Massively Parallel Processing for complex queries, in the accelerator. All of this is provided to the applications without change. They continue to connect, authenticate and authorize through normal DB2 for z/OS interfaces and mechanisms. In fact, the only access to the accelerator is through System z and DB2 for z/OS, which maintains the security of the environment. This is different from anything else on the market.
Competitive offerings require indexing
Some of our competitors claim to have “one size fits all database servers”, the problem is that those servers must still be configured to the specific workload – OLTP or Data Warehouse as an example. They can’t do both effectively at the same time. Others offer specialized systems for running deep analytics. In all of those cases, you still have to create indexes and keep maintaining them based on what queries you expect to be running. With the IBM DB2 Analytics Accelerator for System z, you do not create any indexes at all. As a result, you can run ad-hoc iterative analytics against your data. And you still run your high volume transactions on the same system too. This hybrid solution can truly handle the entire spectrum of workloads, using the right technology at the right time.
As you will see in this video the competition cannot match the “wicked fast” performance of the DB2 for z/OS and IBM DB2 Analytics Accelerator combination.
DB2 Analytics Accelerator vs. Competitor
Watch IBM System z, running on DB2 for z/OS & the DB2 Analytics Accelerator based on Netezza technology, and a top competitors data warehouse system race against the clock!
This heads up competition will compare the BI Query Performance of the competitors. See if IBM System z can provide the best possible performance for both large transaction processing workloads, as well as extremely complex analytic workloads, all within a single platform environment.
Based on IBM internal tests of IBM solution compared to results of testing of a competitor’s offering executing workloads in a controlled laboratory environment. Competitor solution was available as of Sep, 2010. Actual results will vary based on actual configuration, applications, specific queries, and other variables in a production environment. Users of this document should verify the applicable data for their specific environment. Contact IBM to see what we can do for you.