Jim Martin, U.S. Representative, Fundi Software
IBM® Mobile Workload Pricing (MWP) for z/OS® optimizes sub-capacity pricing by reducing the impact of transaction workload spikes on your Monthly License Charge (MLC). MWP is an opportunity for IBM IMS™ customers to leverage the reliability and scalability of IMS Mobile, while controlling the cost of the increased workload.
To use MWP, you must agree with IBM on a method to capture and identify (or “tag”) your mobile workloads. The agreement usually involves generating a monthly file of mobile workload CPU usage. The data consists of general purpose processor CPU seconds for each mobile transaction program summarized hourly by each LPAR for all machines that process mobile transactions. MWP includes coverage for IMS, CICS®, DB2® for z/OS, WebSphere® MQ, and WebSphere Application Server for z/OS.
The mobile workload CPU usage file is analyzed by IBM Mobile Workload Reporting Tool (MWRT). MWRT is a free tool (similar to IBM Sub-Capacity Reporting Tool, SCRT) that calculates an adjusted sub-capacity value for an eligible program (such as IMS) based on the proportion of capacity deemed to be “mobile.”
While the desktop-based MWRT tool is simple to use, the process of collecting the required data, formatting it into the comma-separated values (CSV) layout supported by MWRT, and retaining the appropriate data for reporting purposes may have significant costs.
The cost of reporting and managing MWP can be understood in terms of these factors:
- Collection costs: Collecting and retaining the required data, such as IBM System Management Facilities (SMF) records and subsystem-specific logs.
- Processing costs: CPU and hardware costs of extracting reporting data from the original z/OS-based log records.
- Support costs: Staffing and managing the processes associated with MWP. This cost is a subtle one that can easily grow over time.
While the primary goal of MWP is to help reduce CPU costs, you might need to leverage the process to provide additional value to help offset the cost of creating the CPU usage file. Leveraging the process for additional value can allow you to change the process from one of pure additional cost to one that provides valuable understanding and analysis of your mobile workload processing.
This approach can provide the following benefits:
- Potentially decrease software licensing costs.
- Find the distribution of mobile workloads by time.
- Find exceptions to service levels for mobile workload.
- Investigate and fix transaction performance problems.
Figure 1 shows IBM Transaction Analysis Workbench for z/OS, a tool that provides the processing needed for MWP from the complex and heterogeneous components of your mobile transactions. In addition, it provides z/OS transaction problem analysis by exploiting all the available logs, accounting, and traces collected as part of normal transaction processing.
Transaction Analysis Workbench capabilities include:
- Exception analysis identifies potential problems in subsystems such as IMS, CICS, DB2 for z/OS, WebSphere MQ, and WebSphere Application Server for z/OS.
- An interactive dialog with friendly formatting, as well as navigation and search aids, allows you to quickly identify transactions that need investigation.
- Transaction tracking correlates the events across all participating subsystems to provide a replay.
- Timing aids that show either a progressive response time or event latencies so any delay in transaction processing becomes apparent.
TRAN = ‘MOB*’ AND
IPADDR = ‘123.456.789.123’ AND PORT = 63
PTF UI23621 allows you to use these filtering criteria to tag mobile workloads in log data and build MWP CSV output for all supported subsystems.
One major benefit of Transaction Analysis Workbench is that it’s not just a tool for mobile workload pricing. It’s a comprehensive reporting tool that provides insight by way of existing collection infrastructure (established around the intrinsic logs and traces of various subsystems as well as z/OS) to:
- Automate selection of all the input sources.
- Extract meaningful transactional data from various logs to build a reduced file that’s less processor-intensive to analyze.
- Identify exceptional transactions that failed to meet service level objectives or that terminated abnormally.
- Track logs from across subsystems for given exceptions, allowing you to replay the flow of that transaction.
- Tag workloads matching complex criteria across subsystems.
- Output data to support other tooling requirements, such as CSV for MWRT, or JSON and CSV output to support big data and log analytics engines operating off the host.
Transaction Analysis Workbench can help you obtain a positive return on the investment of costs associated with mobile workload data collection. Beyond mobile workload data collection, Transaction Analysis Workbench supports your longer-term goals with the necessary analytic capability to map all of your workloads and allow you a deep-dive analysis of the z/OS transactions that drive them.
See the following web page for more information: Transaction Analysis Workbench for z/OS.