On February 25, 2020 IBM announced IBM Z Table Accelerator V1.1, an application optimization solution aimed at helping application owners boost the performance of mission-critical IBM Z applications and enabling cost-conscious IT operations leaders to reduce resource consumption. This brand-new offering from the IBM Z Operations Analytics portfolio delivers high value by enhancing performance while reducing the Total Cost of Ownership (TCO) associated with your z/OS environment.
As clients are challenged to manage increasingly complex 24/7/365 operational environments, the ability to provide excellent end-user application experiences, improve overall operational efficiency, and preserve existing infrastructure despite resource mitigation is essential. The main benefits of IBM Z Table Accelerator are twofold:
- You can boost the performance of key batch, online, and mobile workloads which frees your workforce to focus on delivering new value while maintaining your business SLAs (Standard License Agreements) and ensures additional system capacity for other applications.
- You can reduce resource consumption (CPU / MIPS) counting against the 4-hour-rolling-average and/or Tailored Fit Pricing usage levels enabling you to reallocate funding toward the more innovative projects you’ve been wanting to tackle. As more clients migrate to Tailored Fit Pricing, more opportunities arise to optimize applications across the entire month rather than just at the 4-hour-rolling-average peak.
IBM Z Table Accelerator code path – the secret sauce
By copying the frequently accessed reference data into IBM Z Table Accelerator high-performance in-memory tables, data access can be up to 100 times faster than is possible using other methods. Reference data represents the ~20% of data accessed most frequently for business transactions where costs tend to accumulate and offers significant opportunity to increase performance and decrease operating costs.
High-performance in-memory tables are accessed using the shortest possible code path as close as possible to the calling application. Accessing buffered data uses the same code path that is used for accessing data from disk; it reduces I/O (In/Out), but still uses the same code path. The top half of the following figure shows an approximation of the Db2 code path taken for accessing data on disk and from buffers. It typically goes through 4 stages of buffering (SQL server analysis services, relational data system, data management, and baseline management) to pull information from the DASD (Direct Access Storage Device).
The bottom half of the figure shows the code path taken to access data from the high-performance in-memory tables. It’s a direct path to the data with no buffering at all. This is why accessing just a small amount of often-accessed data using IBM Z Table Accelerator can make such a dramatic improvement to an application’s processing speed.
In one example, a batch workload was reduced from 7.5+ hours to 15 minutes! In another example, this technology was used to optimize more than 7.5% of an entire Z-shop codebase and has been responsible for up to a 97% reduction in CPU seconds in some applications.
An example where IBM Z Table Accelerator shines
Conventional wisdom says you must expend resources to save resources, right? So how is IBM Z Table Accelerator able to save enough CPU consumption to justify itself? The answer is probably best described in an example:
Consider a hypothetical large credit card company, “Acme Credit Cards”. Acme keeps a database (in Db2 or VSAM) table of all known fraudulent credit cards. As a part of all new credit card transactions, Acme must run a query against the fraudulent credit card table to ensure that the credit card being used in the transaction is not, in fact, fraudulent. As you can imagine, Acme’s system has to run this query millions, if not billions of times per day. While executing one query is relatively quick and simple, the daily load could equate to a sizeable amount of CPU consumption. Instead of running the query normally, IBM Z Table Accelerator can store the database in-memory and shorten the code path (using its proprietary algorithm) required to reference the fraudulent credit card table. While substantially reducing the CPU consumption on a single query will have virtually no effect on overall MIPS usage, reducing consumption on millions and billions of table calls could have significant implications both for Acme’s applications and cost savings.
Key Takeaways
IBM Z Table Accelerator V1.1 is a new IBM offering that adds significant value to customers seeking to optimize cost and skills in their Z operations environment. This offering is geared toward optimizing frequently accessed reference data in Db2 and/or VSAM tables powering any z/OS applications.
You can learn more in the Announcement Letter. There, you can also find information about the other offerings in the IBM Z Operations Analytics portfolio aimed at helping clients digitally transform with resilient AIOps solutions. Please check out our data sheet or reach out to your IBM representative if you have any questions.