It is now evident that the threat of COVID-19 can also be measured through I/O activity in z/OS datacenters. The daily graphs provided by EADM are showing customers that I/O numbers are impacted and have dropped sharply since emergency measures requested large Enterprises to have their personnel work from home.

 

 

 

On March 17 2020,  a confinement plan was implemented in France.  That day, a Bank’s I/O activity dropped 50% between 10 and 11 AM during Transaction Processing, showing bank employees had problems accessing z/OS resources. Did employees working remotely have adequate network resources and did they have the right authorizations to work from home?

 

A similar analysis with weighted averages over 8 hours of TP shows a much lesser impact.

It was therefore critical to analyze I/O performance with specific hour frames in order to observe that there were issues with remote access to the Enterprise’s z/OS resources.

 

 

The impact of the Coronavirus lockdown is much less significant during BATCH because program chains remain the same: Save / Sort  / Save.

 

Conclusion:

Careful, weighted averages can hide important information in z/OS I/O performance analysis .