IBM® DB2® Analytics Accelerator V4.1 expands the value of high-performance analytics. DB2 Analytics Accelerator 4.1 opens to static SQL applications and rowset processing, minimizes data movement, reduces latency, and improves availability.

The IBM^® DB2^® Analytics Accelerator for IBM z/OS^® (simply called DB2 Accelerator or just Accelerator for this IBM Redbooks^® Solution Guide) is a high-performance appliance that integrates the IBM zEnterprise infrastructure and the IBM Netezza technology to accelerate data-intensive and complex queries in a DB2 for z/OS highly secure and available environment (see Figure 1). DB2 and the Accelerator appliance form a self-managing hybrid environment, concurrently and efficiently running online transaction processing and online transactional analytical processing together with business intelligence and online analytic processing workloads.


Figure 1. DB2 Analytics Accelerator

IBM DB2 Analytics Accelerator V4.1 expands the value of high-performance analytics. DB2 Analytics Accelerator V4.1 opens to static SQL applications and rowset processing, minimizes data movement, reduces latency, and improves availability.

The IBM Redbooks publication Reliability and Performance with IBM DB2 Analytics Accelerator Version V4.1, SG24-8213 helps technical decision makers understand the benefits of the Version 4.1 Accelerator when used with DB2 11 for z/OS. The installation of the new functions and the advantages to existing analytical processes as measured in a controlled test environment are described. The IBM DB2 Analytics Accelerator Loader for z/OS V1.1, a tool that facilitates the data population of the DB2 Accelerator, is also introduced.


Did you know?

By investing in the DB2 Accelerator, query elapsed times can be significantly reduced, CPU cost for query processing can be moved to the Accelerator, and storage cost for data that is no longer modified can be reduced by moving that data onto less costly disks attached to the Accelerator, all while protecting the IBM System z® investment.

The Accelerator has also been enhanced in recent releases to handle not only dynamic SQL but also static SQL. This improvement can extend query processing to the transactional and batch workloads for those queries that qualify for acceleration. The possibilities that this improvement brings to System z is almost endless.


Business value

“Wicked fast!” was the phrase used to describe the DB2 Accelerator by a customer during the implementation of a proof of concept (PoC).

The DB2 Accelerator, using Netezza technology, and combining with DB2 for z/OS creates a solution capable of increasing query processing by 2000X. It allows for the completion of analytics queries in seconds that used to take hours (see Figure 2). Most important, it allows queries that would never have completed because of their complexity to finally produce an answer, in only seconds or minutes.


Figure 2. DB2 to Accelerator query comparisons

In the Accelerator’s more recent release, it can process static SQL in addition to dynamic SQL. This allows it to be used for long running production SQL that in the past would have been relegated to batch and second and third shift. With the ability to complete an SQL in seconds or minutes rather than hours, It frees up resources and restriction of when SQL can be executed.

Although not its primary purpose, it can also reduce mainframe MIPS (millions of instructions per second) usage. If a query that ran for 2 hours for example on the mainframe is now executed on the Accelerator, not only does it complete in a significantly shorter elapsed time, the MIPS that would have been consumed by running that query on the mainframe are now available.


Solution overview

Business Analytics are no longer considered a “nice to have” with low performance expectations. Data warehousing and Business Analytics are becoming mission critical applications with the same quality of service (QoS) requirements as traditional Online Transaction Processing (OLTP) workloads. The applications must be reliable, available, and secure: attributes synonymous with System z. Sometimes, there is also a need to have the analytics workload coexist with traditional OLTP in a mixed workload environment. The ability to run mixed workloads is a major differentiator when determining a warehouse-hosting platform.

Today’s analytic solutions must incorporate the same ideals of all mainframe applications: