It seems that the network industry thrives on complexity. We like complicated command lines, scripting languages, and network engineers are always fiddling with the network to try and make things better. However, does this continual fiddling really make the network run better? A study done by ZK Research shows that human error accounts for 37% of network downtime. That’s right if this were Family Feud and Richard Dawson asked “Name a cause of network downtime” and then yelled “SURVEY SAID!”, we would hear the “bing!” and see “Human Error” as the number one response — ahead of telco outages, hardware failures, power outages and everything else.
So why is this? Well, first of all, I think network managers really haven’t embraced the visual front-end over the years, and frankly, I don’t blame them. As a former network manager, I tried many of them, and to be honest, many of them caused more problems than they fixed. I remember one that Cisco itself made called “Config Maker” that the companies own TAC referred to as “Config Breaker”.
Well, I’m here to say, the tools are getting better. I recently ran across a product called LiveAction, from a company called Action Packed! Networks, which I think has one of the best visual front ends I’ve seen. Here’s a link to a video they created showing how much faster the automation capabilities in LiveAction are when isolating and fixing a Spanning Tree problem compared to using the good old CLI. The company’s website is actually filled with examples where the automation capabilities beat CLI.
Frankly, I’m not sure I need a video to convince me that a visual, automated way of doing things is better than a manual, command-line-driven way. Is configuring your PC easier through the “system” icon, or is it easier to go to DOS and edit the .sys file? Automation is better — as long as you can trust it, and that’s where the rub has been since the tools of yesterday didn’t deserve the trust.
I do believe, though, that we’re entering an era of IT where we need to start to trust these automation tools more as the network becomes more important. I’ve said this over and over, but it’s worth repeating: Cloud and mobile computing are network-centric compute models and the success or failure of these initiatives largely depends on the network. That being the case, we need to make sure “human error” moves way down the list of causes of network failure. It may have been OK in the Richard Dawson, best effort era, but it certainly doesn’t cut in the mission critical, Steve Harvey era of networking.
So for all, you automation skeptics out there, take another look at tools like Action Packed! and find one that works for your organization — you won’t be able to scale network operations much longer doing things the old way.