MAINFRAME > Administrator > Db2

Making DB2 Zip With IBM’s zHyperWrite


New and expanding workloads—including applications running in the cloud and support for analytics, social media and mobile device access to business applications—increasingly demand more from enterprise mainframe systems. They drive the need for improvements in security and the processing demands that requires. These workloads bombard IT systems with requests for information and provide new data to existing business applications.

The corresponding, ever-increasing I/O activity is stressing the current IT systems’ capability to respond. Mobile users generate additional transactions that flow from systems of engagement to systems of record. Both often reside on the mainframe. At the same time, accelerating database transactions is good for business. Consistently fast transactions yield higher user productivity and greater user satisfaction.

IBM’s zHyperWrite* storage technology was introduced in the IBM DS8870 storage subsystem for accelerating DB2* transactions on the mainframe when using IBM’s HyperSwap* technology. zHyperWrite technology became available in December 2014 with the DS8870 7.4 firmware release and corresponding software exploitation PTFs for z/OS* V2.1 (APARs OA45662, OA45125 and OA44973), and for DB2 10 and DB2 11 (APAR PI25747).

zHyperWrite technology speeds the DB2 log writes by eliminating a majority of the time required for the synchronous replication of data. This improves both DB2 log throughput and transactional response time. zHyperWrite technology allows applications running under middleware having the right I/O profile and disaster recovery mechanisms to run primary and secondary write operations in parallel. This parallel processing reduces the overhead of synchronous replication technologies, such as IBM Metro Mirror, to I/O service times. zHyperWrite also supports multitarget Metro Mirror, now available in the same DS8870 firmware release. DB2 for z/OS is the first middleware to exploit the zHyperWrite technology.

The Challenge

As mentioned, accelerating database transactions is good for business. Close examination of a banking workload’s transactional latency with DB2 shows the typical components of a transaction’s elapsed time (see Figure 1).

Faster CPU technology delivered via new server releases helps reduce DB2 server time. Software enhancements regularly improve scale, reduce contention and improve performance, and new networking technology enhancements can help reduce network delays. However, for many applications, a very significant component of transaction latency comes from synchronous I/O delays when writing to the DB2 log. Unlike other possible synchronous read I/O operations, clients cannot tune around the DB2 log write. Acceleration of DB2 log writes offers a significant opportunity for accelerating transaction execution.

Acceleration of the synchronous log write will have a number of positive effects on DB2 in terms of scale, transactional latency, reduced client costs and improved resilience. zHyperWrite accelerates the DB2 log writes by eliminating most of the time required for the synchronous replication of data. Positive effects include:

  • Transactional latency is reduced because up to 43 percent of the write I/O service time to the DB2 log is eliminated. In the client example in Figure 2, this 43 percent improvement in DB2 log write latency resulted in a 28 percent reduction in DB2 transactional latency.
  • Scale is improved by increasing DB2 log throughput up to 179 percent as shown in Figure 3. Reduced I/O service time for log writes yields corresponding improvement in DB2 log throughput. This allows increased transaction volume growth without adding additional DB2 data sharing instances, z/OS LPARs or processors.
  • Reduced client costs result because you can defer splitting DB2 databases into additional data sharing instances, growing more within the same hardware and software footprints.
  • Improved resilience occurs because your infrastructure is better positioned to handle unexpected workload spikes that may occur.

Calline Sanchez is the VP of IBM Systems Storage program management.

Harry M. Yudenfriend is an IBM Fellow with Systems and Technology Group, System z and Power who joined IBM in 1980. He was named an IBM Master Inventor in 2001 and has achieved his 33rd invention plateau.

Peter Sutton is an IBM Distinguished Engineer for IBM Z core technology.


comments powered by Disqus

Advertisement

Advertisement

2019 Solutions Edition

A Comprehensive Online Buyer's Guide to Solutions, Services and Education.

An Intuitive Approach to DB2 for z/OS SQL Query Tuning

Real-world samples provide valuable lessons.

MAINFRAME > ADMINISTRATOR > DB2

A Db2 Utilities Migration Project

IBM Systems Magazine Subscribe Box Read Now Link Subscribe Now Link iPad App Google Play Store
Mainframe News Sign Up Today! Past News Letters