AIOps on IBM Z - Group home

Latest updates for IBM Z Performance and Capacity Analytics (February 2021)

By Chris Walker posted Thu February 11, 2021 02:32 PM

It's been about a year since IBM Z Performance and Capacity Analytics V3.1 (IZPCA) was announced. In the intervening 12 months, there have been several updates to the product to further enhance the capabilities around performance analysis, capacity planning and cost management.

In this update, we'll cover a summary of the most recent updates:

Support for z/OS Encryption Readiness Technology (zERT)
z/OS V2R3 Communications Server introduced a new function called z/OS Encryption Readiness Technology (zERT). With zERT, the TCP/IP stack acts as a focal point in collecting and reporting the cryptographic security attributes of application network traffic that is protected using the TLS/SSL, SSH, and IPSec cryptographic network security protocols. Data collected about the network traffic and the security protocols used is valuable to understand if there are potential weak or compromised connections within the environment and potential impacts on application performance due to network connection overheads. IZPCA (APAR PH12196) has been enhanced to process SMF 119 subtype 12 records that provide a summary of network connections using zERT. Using these tables and pre-defined reports, it becomes easier to identify network connection that may be using weak, or even no, protocols for securely transferring data. By identifying the applications and clients involved, network administrations can make quick decisions about client access to applications on the network and build a plan to ensure a more secure environment.

Additionally, the use of this data can also benefit performance analysts in conjunction with the other data sources available in IZPCA. With the overhead of establishing and terminating secure connections, application performance can suffer if this occurs excessively. Looking for correlation between application performance information and the network details coming from zERT can help identify potential areas of focus for improvement by making more efficient use of network connections.

Updated "what if" analysis to identify zIIP offload opportunity
In IZPCA V3.1 we introduced the concept of "what if" analysis to support capacity planning activities through a set of reports that enabled modelling of current workloads at a CEC or LPAR level and then simulate what that workload might look like under a different processor. This functionality has been enhanced with a layout redesign and the ability to apply growth factors, that enables an automatic percentage factor growth or decrease to future workload values based on your own knowledge of upcoming changes. You can also view the report at various levels of granularity (Monthly, Weekly, Daily, Hourly or timestamp level).

Two new reports has been added to provide zIIP offload analysis. The LPAR zIIP Offload Analysis report gives a view on current MIPS workloads on a general CP today and how that could be affected if zIIP-eligible work were to be offloaded. Selecting a specific date for a given LPAR allows a drill-down into the Workload zIIP Offload Analysis to detail which individual workloads would benefit from additional zIIP capacity. Such calculations are important when making investment decisions to maximise benefit and, considering consumption based pricing models like Tailored Fit Pricing, where saving can be made through effective use of zIIP engines.

Accurate forecasting for key events that shift dates
Using the wealth of historical data collected over months and years is a vital input to making accurate plans for future resource needs. Very few industries expect consistent workload levels across 12 months of the year and seasonal variations need to be accounted for. In some cases, these peak periods occur at consistent times each year, for example the US Thanksgiving holiday is always on the fourth Thursday of each November and Black Friday, when retail expects to significant transaction level increase, is the following day. Other peak periods may move around the calendar from year to year. An example here is Chinese New Year. In 2021, this falls on 12 February however in 2022 it will be on 1 February and in 2023 even earlier, on 22 January. If this event drives exceptional behaviors within your environment, a forecasting process that does not take this into account will be misleading and require additional manual planning.

The new Fractal Forecasting capability in IZPCA (APAR PH32126) provides a solution to this challenge. By defining an internal calendar to record these key event days as they vary from year to year, historical data collected from similar dates will be aligned correctly in generated forecasts for future events. This gives planners more confidence in the forecasts they are working with, reduces the need for manual adjustments and simplify the overall capacity planning process.

Additional pre-defined reports for Cognos Analytics, Splunk and Elastic stack
IZPCA supports several reporting platforms to deliver accurate information in whatever is the most appropriate format for the end user. The data curated by IZPCA can deliver insights for many different stakeholders such as management (wanting high level daily, weekly or monthly reports), subject matter experts (wanting detailed information for root cause analysis), or site reliability engineers (wanting to integrate Z data into enterprise wide views).

Recent updates over the second half of 2020 delivered more pre-defined reports for Cognos Analytics, Splunk and Elastic stack enabling users to gain immediate value from the data collected without having to become skilled in report writing. You can, of course, extend these reports and customize them further to fit with the unique needs of your environment.

Key Takeaways
IBM Z Performance and Capacity Analytics is an integral part of the IBM Z AIOps portfolio for leveraging analytics to support the drive to proactive and intelligent decision making. The most recent updates add further support to simplify the processes around performance and capacity management reducing the time to make key decisions around your IBM Z infrastructure. Please reach out to your IBM representative if you have any questions or would like a demonstration or assessment on how IBM Z Performance and Capacity Analytics can support your goals along the Journey to AIOps.