On July 23, 2010 I attended the announcement of the IBM Smart Analytics Optimizer (ISAO).  This was IBM’s first iteration of heterogeneous scale-out to enable mixed OLTP and OLAP workloads to run on the same platform.  Subsequently IBM acquired Netezza and the Netezza appliance became the foundation for the IBM Db2 Analytics Accelerator (IDAA).  In IDAA workshops I would refer to the ISAO as IDAA v0.5.

Fast forward to today when IBM announced IDAA v7.1.  “What is old is new again.”  Think of this delivery of IDAA as “ISAO v4.0” (there was some good-natured debate about the numbering).  Why do I frame it those terms?

1. IDAA has received a “brain transplant”.  Instead of Postgres, the database engine for IDAA will be Db2 Warehouse.  This will give IDAA greater compatibility with Db2 for z/OS, including broader SQL and datatype support, TIMESTAMP consistency, and more.  By exploiting Db2 with BLU Acceleration features such as in-memory columnar (rather than row-organized) data, Single Instruction, Multiple Data (SIMD) chips, actionable compression, and “data skipping” (avoiding extents that don’t contain data of interest) IDAA v7.1 provides a robust platform for better performance with a roadmap for continuous improvement and enhancement.  IBM has also consolidated its data repository development efforts by focusing on fewer, higher-value initiatives.
2. IDAA has received a “heart transplant”.  The IBM Integrated Analytics System (IIAS) appliance on which IDAA v7.1 is based, no longer depends on Field Programmable Gate Arrays (FPGA) that were the “secret sauce” of the Netezza (aka PureData System for Analytics – PDA).  FPGAs might be reintroduced at a later time.

So, a few questions – and a fewer answers.

• How long will the legacy Netezza/PDA “Mako” N3001 be supported?  Until 2023.  So the installed base of existing IDAAs will be protected for five years.
• Will the Database Partitioning Feature (DPF) – IBM’s MPP shared-nothing data warehouse scaling technology – be exploited by IDAA v7.1?  Dunno.
• Wither the IDAA cloud deployment introduced in v6?  Nary a mention today.

Topics for a future blog post:

• “True HTAP” – Hybrid Transaction / Analytic Processing
• IDAA deployment on IBM z – using Docker images sitting on Integrated Facility for Linux (IFL) specialty processors.

From November 9 – 14 I will be visiting one of the most charming cities on the planet when I attend the International DB2 Users Group (IDUG) conference in Prague, CZ.  I will present “IBM Data Retrieval Technologies : RDBMS, BLU, IBM Netezza and Hadoop”.  It will be session E7 on Tuesday, November 11 from 3:15 to 4:15 p.m.  Please consider attending IDUG Europe and joining me at this session.

Abstract:  IBM offers four distinct data retrieval technologies: the traditional RDBMS, which primarily relies on indexes to speed access; the new BLU Acceleration columnar compression database; IBM PureData System for Analytics (IBM Netezza), which deploys racks of Field Programmable Gate Array (FPGA) processors to parse the data; and IBM InfoSphere BigInsights, which is the IBM distribution of Hadoop. Choices are good to have, but how do you choose which technology to apply to a particular business use case? In this session, you learn how these techniques differ, including their relative strengths and weaknesses, to help you make an informed choice.