How Db2 Analytics Accelerator Improves Analytics on z/OS

IBM promotes Db2 for z/OS as a solution for business intelligence and data analytics. Its latest version, Db2 12, improves scalability, data ingestion capabilities, and performance, according to IBM’s Robert Catterall in a recent SHARE presentation.

Central to Db2’s ability to offer powerful and cost-effective analytics capabilities is a tool that’s actually been around for several years. Db2 Analytics Accelerator uniquely extends Db2’s ability to handle complex or data-intensive queries, Catterall explained.

The accelerator helps redefine conventional wisdom around running analytics on the z/OS system by introducing space-saving and performance-improving capabilities that make it much more cost effective to run analytics on the mainframe.

Db2 Analytics Accelerator is part of what makes z/OS such a great platform for big data analytics, despite perceptions that it’s exclusively an operational workhorse. Here’s what IT professionals need to know about this important tool.

What is the Db2 Analytics Accelerator?

Physically, the accelerator is an appliance that sits near the server running Db2 for z/OS, connected by a dedicated high-speed connection. Logically, it’s invisible, with no separate user interface required to run queries through the accelerator.

The big difference is performance: standard complex Db2 queries run far faster when executed by the accelerator.

The Benefits of an “Extended” Db2 System

Beyond sheer speed of execution, the accelerator saves time and costs in several other ways.

For example, it allows database administrators (DBAs) to run high-performance analytics in z/OS without having an analytics-oriented design for the databases being analyzed. An admin looking to analyze an operational Db2 table with ad hoc queries won’t first need to put that operational data into a star schema, Catterall said. Instead, target tables are copied to the accelerator unchanged, and from there, programs can perform quick analysis on that data – in some cases, a query that might take an hour will complete in seconds.

Other features of the accelerator broaden what it actually means to keep data in Db2 for z/OS, Catterall said. Data can be physically present solely on the accelerator while appearing only logically in the front-end Db2 database, which pays off in the form of storage and performance benefits.

Similarly, the accelerator can bear the brunt of costly SQL-based data transformation processes by allowing administrators to specify certain tables as “accelerator-only” tables. Physically, those tables will exist only in the accelerator, but are logically still available to the front-end Db2 database.

Using z/OS for Analytics

Beyond the Db2 Analytics Accelerator, z/OS is well-suited for analytics because it supports the analysis of both highly structured transactional data, as well as unstructured data (e.g., clickstreams, geospatial information and z/OS operational data like SMF data), Catterall said. As a result, businesses can find correlations across disparate data sources, from structured transactional data to unstructured information like clickstreams and geospatial data. Tools like the z/OS Platform for Apache Spark make analysis on z/OS even more powerful and cost effective with capabilities like rapid ingestion for unstructured data. 

This all helps enterprises fulfill the promise of big data analytics: to find useful information that can make a difference to the business’s bottom line. It’s not just about analyzing historical data, either. Businesses are using z/OS to gain a real-time view of customer purchase behavior and preferences, or to predict future sales or operational outcomes. Ultimately, z/OS can provide the technology foundation to breakthrough business insights.

To learn more about the analytics capabilities available on z/OS, watch the full video from Robert Catterall’s presentation in the SHARE Content Center.


Recent Stories
Security Experts Are Getting Excited About FHE (Fully Homomorphic Encryption)

SHARE Vice President Brian Kithcart Shares Why ‘Collaboration’ Is Key to Mainframe Work

Message From SHARE: A Look at Virtual Summit 2021