The old work-horses of IT, mainframes, are still there and we are still highly dependent on them. Although they run based on a code written in an age-old language called COBOL, even today, the heavy number crunching tasks they undertake and volumes they deliver are terrific! As part of this process, they spew lot of data which tells a lot about their own system’s performance, condition, etc. Now, the million dollar question is why not we use this data to improve the performance of our old-age machines and make them run faster? What helps us to do this?
The answer lies in analytics. Gather and channel data thrown by mainframes in to a proper analytics application. Hurray! You could see you old-number cruncher talking to you through graphs, charts, etc. The analytics application dashboard displays all the data thrown by your mainframe in an easily understandable format after applying all the pre-defined rules, models, etc.
Remember: the technology has moved from ‘management by observation’ to ‘management by automation’ via ‘management by exception’.
As part of the automation, you can even set thresholds and alerts based on these thresholds. Since thresholds change based on workloads, peak hours, etc., you need to set different thresholds when it comes to mainframe environment to help it alert you correctly based on a given scenario. You have to set ‘new normal’ for all kinds of scenarios that your mainframe experiences during its working period. Considering this new normal as the baseline, the system alerts you when there are any deviations.
In summary, linking analytics to mainframes’ performance helps you to reap rich benefits.