Are We Still Scared of IT Automation?

This thread came in a weird way as a result of seeing yet another re-run of the classic movie “Back to the Future” the other day. Bear with me on this one, and hopefully it will start to make some sense. The story begins when I had just started out in my IT career way back in 1990—yes, it’s been a while. I’ll come back to that in a sec. In the present, I often look at my kids and think of how they take all the ubiquitous tech—the internet, smartphones, etc.—for granted, and they’ve never know a world without constant Wi-Fi and all that goes with it. Actually, it’s not just my kids’ generation, even for me who clearly remembers a world with no Google, no Amazon, no YouTube, no internet period. It’s been so all-pervasive for so long now, it would be a real shock and really hard to go back to the pre-internet way of life. So, back in 1990, I had just moved to Germany from the U.K. working as a technician on big IBM water-cooled mainframes. Computers big enough to fill a decent-sized room. Lots of travel to different cities was involved and having only recently passed my driving test, getting used to driving on the “other” side of the road and getting into the new language, etc., navigating around to customers was a bit stressful at times, especially as being late in Germany is frowned upon. I was never a fan of the fumbling with a (paper) map on go. Let’s face it, it’s a bit dangerous. Especially annoying was when your street was on the fold of the map. Writing out the directions helped, and I got an idea to borrow a shortcut from a mundane but important task: working with the mainframes. When we got a new system in, it was crucial to record the serial numbers of all the TCM (thermally cooled modules). Very large and expensive chips. This was boring and required two people: one calling them out, one writing. I got the idea to record them into my Walkman and note later. So, I thought, for the next trip, why not record the directions into the Walkman and play/pause them back? It worked like a treat (most of the time). OK, it was no scientific breakthrough, but a handy little hack, nonetheless. I was chatting to a colleague about this who informed me that a few companies were working on SatNav, and he proceeded to give a somewhat inaccurate description of how it would work. It all sounded supercool and futuristic. Now, if I was able to talk to my 1990 self and say in 2021, not only will I be able to dispense of maps completely because not only will SatNav be built into almost every car, it will also be a free app on your phone and watch. This would really have seemed nuts. Now, if I went on to say there will be cars and an ecosystem allowing you to not only automatically find the best route based on distance, crowdsourced traffic data, accident blackspots, etc., but these self-driving cars would automatically brake if too near the car in front, would automatically detect lanes and veer you back on course if you get drowsy, and (the clincher) could even drive you safely to your destination while you sleep or whatever… but we would stop at that last bit. Why? Well, of course, it’s complicated: legislation, the what ifs, essentially, is down to trust. Everyone knows how awesome it would be if we didn’t have to worry about the what ifs when it comes to this example of automation or in an IT context. The release from mundane tasks and huge productivity gains, the fact that as per this Network World article citing an uptime institute survey: 70% – 75% of data center errors are caused by human error and 60% of respondents believed their most recent significant downtime incident could have been prevented with better management/processes or configuration.

IT Automation Examples

So, of course we can talk about great examples of this, whether at the network layer automatically deploying proven Gold-Standard configurations, at the storage layer doing automated failovers of services to maintain availability, or at the server layer auto-expiring VMs to name but a few examples—all great to drive efficiencies and ensure predictable, repeatable tasks free of dreaded human error. Alas, the IT world and that which it serves can be somewhat unpredictable. It reminds me of when I heard of the three-letter acronym "SLO" (service level objective) as distinct from SLA (service level agreement) for the first time. Although they look very similar, there’s a great difference in their promise. SLO is more about reality, more akin to in automation in context. In other words, where you predict accomplishing your objective based on the existence of a specific set of conditions. And, if these conditions change, then the overall balance and allocation of resources—be it memory, change of path in a network, or whatever—adjusts in attempt to fulfil the objective, not an agreement. So, just like the car analogy where we have lots of tools to automate the navigation, safety aspects, and so forth, whenever the day comes when we entrust the car to automatically find or change the route and all other adjustments in between without human intervention to get us to our destination in time, this will be a huge step. Similarly, in IT, is automation really just about a series of shortcuts without overall context and lacking the ability to adjust and change in-context, or is it the ability to zoom out with observability of the big picture across the stack with the ability to dynamically adjust the systems to fulfil the objective that was set out? Of course, it’s not trivial. It’s a journey involving many facets: standards agreements, interoperability, artificial intelligence, machine learning, and much more. I’d love to know where you are on the automation journey today. How much do you trust to automation? What types of IT automation solutions would you like to see SolarWinds have in this space? On a final thought, after watching a clip a couple of weeks ago on YouTube of Elon Musk pre-announcing the Tesla bot and rejoicing in the tedium it will relieve us of, imagine if we could talk to our 2050 selves. What would they tell us?