Author: Don Jones
Business information technology has come a long way. From the mainframe days of yore, to the LAN-centric networks of the 1980s and 90s, to the domain-based infrastructure and the virtualized “sprawl” of the late 2000s and beyond, we’ve had a pretty good run. The unifying theme of this progression is scale. Our environments have simply gotten bigger and bigger, with more assets to care for. Many organizations have done an excellent job of keeping up by implementing private cloud technologies to help reduce operational overhead, diving into DevOps to add agility, and more.
But some businesses are still managing their IT estates like it’s 1999. Their practices have become increasingly out of date and unable to keep up with the expanding scale of IT. In many cases, the response to this is to slow down or stop the growth of IT. This is a horrible prospect for any business that relies on technology for its daily needs.
In most of these cases, executives and mid-level directors don’t realize how precariously balanced their IT teams are. Things don’t break on a daily basis, they reason, and so things must be fine. But the truth under the hood is a lot scarier; many of these environments are teetering on the edge of collapse. It’s like running a business that ignores the existence of social media; you can get away with it for a while, but eventually it’s all going to come crashing down around you, and it’ll likely come as a huge surprise when it does.
So how can you tell if your IT environment is being managed using completely outdated techniques? How can you tell if you’re running one of these legacy IT teams that’s poised for failure? There are a few obvious signs that we’ll cover here.
1. Static IP addressing
If your team still statically addresses server and other datacenter assets, you’re definitely living in the 90s. This tactic originally developed because the earliest dynamic addressing solutions were finicky. We didn’t want our servers to somehow end up without an IP address, and so we provisioned them statically, often using spreadsheets as ersatz databases to track those addresses. That’s right, we ran our network using Excel and somehow believed that was a good idea. Today, static addressing is essentially the result of superstition. “If the DHCP server fails,” the reasoning starts, at which point you should be jumping in and shouting, “why do we have a DHCP infrastructure that isn’t highly available?!?!?”
2. Brand loyalty
Have you ever seriously thought about seating your Linux and Windows admins in different buildings, just to end the arguments? Have you ever implemented that idea? Or, if you’re a single-OS environment, ever ask why? Is it because your team has convinced you that their chosen OS is the Right-Tool-For-All-Jobs? It’s kind of like saying a two-bedroom apartment is the perfect fit for a family of any size. The only person who can get away with saying that is a builder of two-bedroom apartments (and you’re probably not going to take him very seriously). The "I only know how to support one thing" mindset should terrify IT leaders to the depths of their souls. You’re basically being told that you can’t have flexibility; you can have this one thing, and you’d better hope it’s enough. Literally, no other part of your organization could get away with that.
3. Next, next, finish
When you ask your administrators to make a change in the environment, do they whip out a graphical user interface? Worse, do they initiate a remote connection to the server itself, and run the graphical tools there? If so, be afraid. That administration modality was never a good idea, and smart people walked away from it years ago. It’s less secure, it drains server performance, and it’s the sign of an administrative team that, frankly, doesn’t know what they’re doing. Sure, some legacy applications in your environment may still need that, or a similar approach, but your team should be complaining for you to replace those things, not perpetuating a terrible process. A modern IT team will focus on units of automation, like scripting, and attempt to solve every problem by writing a script to solve that problem--now and when it reoccurs in the future.
4. Servers with beards
If your servers are old enough to have gray whiskers, running operating systems that were in fashion when MTV still stood for music television, then you’ve spotted a key driver of about-to-fail IT. You can’t have fresh, modern, flexible IT when you’re running ancient software (anything more than six years old). Think about it: The security threats, business challenges, and administrative techniques of today didn’t even exist six years ago. How can an older server operating system expect to keep up? Worse, environments dependent on old software attract ossified IT teams who don’t like change. The team stops learning, stops growing, and stops being interested. Stroll in one day and ask them to innovate, and they’ll probably groan when you leave the room, and go back to propping up your aging infrastructure.
Fixing the problem
The fixes for this, and the keys to getting your IT environment stabilized, are threefold:
2. Modernize at least a bit. You don’t have to upgrade every server to the latest OS, but you do need to upgrade a few. Give your team something modern and interesting to work with. IT people (good ones) got into the field for the novelty. They like the new shiny. Let them run a pilot project with something new and shiny, so that they have a reason to get engaged and enthused, and to flex their brain muscles.
3. Educate. You may think that there’s no budget for education, but that’s insane. You can’t possibly expect a team to remain useful and interested without continually growing. And if your team isn’t interested in continually growing, then you’ve probably already lost the good ones, and now you’re at the bottom of the bin. Try to get your team engaged. Send them to a local conference, get them some inexpensive video training, or let them buy a book or two. Insist that they identify outdated, archaic management practices and propose ways to update them. Insist on more automated and less hands-on IT environment, and then let your team chart a path to learn how to make it happen.
Join Don, along with Jason Helmick, Jeffery Hicks and Greg Shields as they discuss emerging trends in the security industry at Pluralsight LIVE, our first ever in-person conference.