System Management Facilities (SMF) offer a way of keeping track of what is happening on your mainframe. These reports help fill in the picture of mainframe activity. And, while this might seem simple at first, SMF records (in their numerous iterations) confound mainframe managers due to their overwhelming, often complex, breadth of information and unapproachable format.
Line after line of uncontextualized data on billing, system performance, data archiving, and more — it’s enough to make anyone’s head spin. There are 255 SMF records available that help deliver information about all functions and products running on your z/OS environment. So, let’s explore what, exactly, SMF records are, the information they deliver, and how these documents can be used to uncover new mainframe performance opportunities.
SMF Records and Performance Insight
Mainframe performance relies on the dependability of several, interconnected components. No one function operates in a silo — every mainframe action depends on a hierarchy of operations.
For example, a customer’s initial query depends on the speed of CICS; CICS depends on the speed of DB2; DB2 depends on the speed of DASD; and so on and so forth. This means that the time it takes to deliver information to front-end consumers is the sum of all preceding response times. This is all to say that an inefficiency, at any stage, can bottleneck response times since all components depend of the speed of one another.
This is why SMF records are useful — with granular insight into machine activity users can identify, and remedy, any kinks in the chain of command. One of the uses for SMF reports is to help users better understand what is contributing to mainframe performance issues. Although there are 255 SMF records, most companies focus on the top 20 SMF records that are most closely aligned with mainframe KPIs.
| SMF type | Function | Description |
| 14-15 | Non-VSAM | Non-VSAM file closed details |
| 30 | Job or step statistics | Details about jobs and programs running on the mainframe |
| 42.6 | SMS Dataset | Dataset IO statistics per interval |
| 64 | VSAM | VSAM statistics |
| 61, 65, 66 | Catalog dataset | Catalog dataset operations |
| 70 | CPU Monitor | RMF CPU activity |
| 72.3 | WLM activity | RMF WLM metrics by service classes |
| 73 | Channel path activity | RMF channel path activity per interval |
| 74.1 | Device activity | RMF device IO statistics per interval |
| 74.5 | Cache activity | RMF cache subsystem device activity |
| 74.9 | PCI/e activity | RMF |
| 78.3 | LCU activity | RMF LCU / Hyper PAV activity per interval |
| 80 | Security activity | RACF/TSS security events for audit |
| 101 | DB2 accounting | DB2 thread activity (IFCID 003) |
| 110 | CICS performance | Transaction execution details |
| 113 | Hardware performance | Hardware Instrumentation Services |
| 225 | AutoSoftCapping | zCost ASC - Software capping activity |
his table illustrates the sheer extent of information that can be conveyed through SMF reports. From Catalog Dataset Operation to Transaction Execution Details, tracking system performance through SMF records can be intimidating. So, if this all seems overwhelming it’s because it is.
Luckily there is a better way to leverage SMF records.
How to Get More Out of SMF Records
SMF records deliver a staggering amount of mainframe data. What do you do with it?
There is no intuitive way to analyze SMF records by default in z/OS. Some users might choose to analyze these documents in a spreadsheet, and identify anomalies by hand, but this as tedious as it is inefficient.
Zetaly Service Intelligence offers a better way — this series of application collects SMF records in real time and displays information in user-friendly interface. This helps support system governance and analytics.
This software suite collects all top 20 SMF records, in real time, and stores them in a centralized database outside of the mainframe. This way, IT can correlate all information from different SMF record types to help easily uncover new information about the z/OS environment.
For example, the MSU portal within Zetaly Service Intelligence displays two key SMF records related to mainframe performance:
- SMF 113 — Hardware Instrumentation Services (HIS). HIS breaks down hardware cache usage and offers insight into how instructions are interacting with the various levels of cache.
- SMF 72.3 — Workload Management (WLM) Service Class metrics.
By visualizing these metrics, Zetaly Service Intelligence makes mainframe optimization more intuitive. This helps to streamline performance analysis and support decision making for z/OS environments.
Other SMF records are displayed in Zetaly Service Intelligence so users can easily gain a global overview of system activity. Zetaly Service Intelligence consists of a series of applications:
- CPU Explorer – Job consumption analysis
- MSU Explorer – Hardware & configuration performance analysis
- CICS Explorer – CICS performance analysis
- DB2 Explorer – DB2 performance analysis
- IMS Explorer – IMS performance analysis
- I/O Explorer – I/O consumption analysis
- DASD Explorer – Storage performance analysis
- Dataset Explorer – File management analysis
- Smart Production – Production monitoring
Zetaly Service Intelligence extracts SMF records before they reach system storage. This means that users gain analytical capability without having to tap mainframe resources. An intuitive analytical framework that doesn’t tax your mainframe resources.
Contact Zetaly for more information on how Zetaly Service Intelligence, and other software solutions, can help support mainframe optimization.
