Blogs

Raising real-time transaction and analytic processing to the next power

Post Comment
Big Data Evangelist, IBM

Transactional data is the lifeblood of enterprise analytics. The vast majority of what's extracted, transformed and loaded into your operational data warehouses comes from transactional databases. And many of your deployed online transaction processing (OLTP) databases are also used for online analytic processing (OLAP) applications such as real-time operational business intelligence.

OLTP and OLAP are joined at the hip on enterprise database platforms such as IBM DB2 with BLU Acceleration. Some refer to this converged application support as online transaction and analytic processing (OLTAP). If you can achieve an order-of-magnitude OLAP performance boost, thanks to in-memory computing technology embedded in an OLTAP database, you can truly dominate in your industry. Also, in-memory OLTAP is a knowledge-worker productivity booster of major proportions, supporting real-time, in-the-moment queries and decision support.

BLU Acceleration.png

In-memory OLTAP, leveraging the best of columnar and row-based approaches, is the core capability of the new release of DB2 10.5 with BLU Acceleration. In-memory columnar databases have clear, well-known advantages over row-based approaches in support of low-latency applications of all sorts. Where analytics are concerned, in-memory columnar's chief performance advantage comes when it is used to accelerate structured, repeatable scans and queries against very large aggregated tables.

In the latest enhancement release to BLU Acceleration, known as DB2 "Cancun Release," IBM has accelerated real-time OLTAP through several key new features that are now available on Linux, UNIX and Windows platforms.

New performance features

Chief among the new DB2 "Cancun Release" features are BLU “Shadow Tables.” In a nutshell, DB2 Cancun maintains in-memory columnar "shadows" of row-based operational-data tables. Creation and refreshing of the shadows is through automatic, incremental synchronization of the corresponding row-oriented DB2 tables. This new feature enables DBAs to boost query speeds automatically without having to manually create, manage and tune complicated indexes. With the DB2 "Cancun Release" enhancements, BLU Acceleration automatically routes analytic queries to the appropriate shadow tables, which leverage IBM's BLU in-memory technology. Not only does this enable faster analytics against DB2, it also ensures continued high throughput on row-based OLTP applications and much less manual administration.

Another important new performance feature in this release is caching-facility memory optimization. This enables in-memory speed with mixed-workload agility. It does so by automatically allocating BLU Acceleration memory based on OLTP and OLAP workload requirements. As with shadow tables, these optimizations boost OLAP performance, maintain OLTP service levels and reduce the overhead associated with administering any or all of these applications on a DB2 "Cancun Release" instance. Similarly, other new features, such as improved support for high availability and disaster recovery, will benefit all applications running on an instance.

How fast are BLU Shadow Tables?

With BLU Shadow Tables, the performance of analytical queries can improve by 10 times or more, with equal or greater transactional performance. This is based on internal IBM testing of sample transactional and analytic workloads by replacing 4 secondary analytical indexes in the transactional environment with BLU Shadow Tables. Performance improvement figures are cumulative of all queries in the workload. Individual results will vary depending on individual workloads, configurations and conditions.

Learn more about this latest release of BLU Acceleration and more on the IBM BLU Hub