This post originally appeared on the Packet Pushers’ now-defunct Ignition site on October 28, 2019.
I was recently asked a question about the best business outcome of automation.
My immediate thought was improved speed of operations by mechanizing operational tasks, like automated software upgrades, creating VLANs, updating ACLs or routing, and so forth. This is the most common automation business value I see discussed because it improves the technology function.
That said, I don’t think automation is the most valuable business goal.
Once you have mechanized (automated) common operations, the most critical aspect of operations become visibility. Did the change make a difference? Did it work at all?
People have an inherent distrust of mechanization, and there is powerful need to validate that mechanized systems work correctly. As mechanization gets more complex, the need for visibility and validation increases.
Historically, the poor quality of the data obtained from SNMP or CLI scraping meant that network monitoring was, well, difficult. It’s a circle of despair because vendors mostly half-assed their SNMP implementations because no one was using it, and thus no one used SNMP because it was so half-assed.
Let’s not underestimate the impact of automation in other infrastructure areas. The operations console of virtualization, storage, and operating systems have added visibility to their tools. People expect to have some insights into what is happening.
Working as part of an IT infrastructure team is a social activity. It’s easy to forget that sometimes.
On the bright side, mechanization means more time available to do new activities. Instead of doing poor-quality work on configuration, time is now available to upgrade you value into working on that monitoring and visibility project that you always wanted.
