I have battle scars. I have also had my share of digestive distress and loss of sleep. Why? I relied on vendors to provide my Internet connectivity and hosting.
Over the course of 20 years, serving billions of pages, some of those vendors were less than ideal, resulting in days or weeks of sleepless nights, panicked searches for alternative solutions, and no small amount of crankiness.
That's why, as I see more and more companies dump their servers for life in the cloud, I wonder if they're prepared. I wonder if they've considered the doomsday scenarios I've lived through. I wonder if they realize they're betting their companies' futures on an inherently risky practice.
Before I take you further down this dark and dire path, it's important for me to tell you that I am not advising you to avoid cloud services. Far from it. I use and rely on many myself. No, what I'm advising you to do is be aware, fully aware, of the risks you're assuming when you give up some control over your means of production (your data operations) to an outside vendor.
A corollary to this is pretty obvious: you'll never be able to control everything. In fact, being a control freak can limit your ability to succeed and tie up your resources with activities better delegated to other vendors.
But since you'll never control everything, you need to make sure you have a plan (or at least a sense of awareness) for what can go wrong, for, undoubtedly, something will.
I may seem bitter, but I'm not. Not now, anyway. I have the benefit of decades of experience, some of it hard earned. Of course, when each individual event splattered into existence, yeah, I was probably more than a little bitter. It's also important to know that I ran servers for many years, so I also had plenty of stretches of calm, smooth sailing.
The goal of this article is to help you recognize storm clouds as you sail the Good Ship Server right into Cloud Services Harbor. Yeah, I know. Wow.
In order to help you get a feel for the risks you face, let me point you to a story I wrote all the way back in 2002, when I moved from a T-1 in the office to what eventually became a T-1 in my home. It's not the office-to-home transition I want to highlight. It's not even the problem I encountered serving on port 80, which was the topic of the article.
What I want to point out to you is that vendors make promises that they don't always keep. I know you know this. But when you get a service level agreement with a promise, but are suddenly knocked off the Internet, no document will cover the pain and stress of explaining to your advertisers and customers why you're down.
Another issue is when one company gets bought by another company. You may have been getting fine service from the original vendor, but the priorities of the new owner may not be the same. Worse, the new owner might be a competitor of yours, in which case, you may find yourself in for very rough sailing indeed.
What about when a company transitions business models? I used a small, regional Exchange hosting provider for about 11 years. I was quite happy with their service. But when Office 365 came out, they saw the writing on the wall and transitioned their business model. I was left with more and more outages and worsening service.
Eventual service degradation has been a big problem over the years. If you use a vendor for a while, they are likely to be fine. But if you find yourself with a vendor for a decade or more, you will undoubtedly find yourself working with new people and those people may not be able to provide the quality of service they once were able to provide.
What about jurisdictional issues and the ever-present concern about hacking? If your vendor gets hit with a hack attack, or -- worse -- if a company you're co-tenanted with is caught someone doing something illegal, government agencies might be involved and you might lose access to your equipment. This goes directly to the discussion we had last week about the different kinds of virtual and dedicated servers.
I could go on for hours about the risks, and there will still be some of you out there with smug expressions on your faces. After all, you're not going to have these sorts of problems, are you? You're using Amazon's Web Services, everything can scale, nothing can stop you, and you're smarter than the rest of the world.
Yeah, I can see you. Wipe that smug expression right off your face.
Look, I'm a huge customer of Amazon (Prime subscriptions and all). And while Amazon seems too big to fail, there is risk in that bet, too. Why?
As Benedict Evans of Andreessen Horowitz puts it, "Almost 20 years after it was launched, [Amazon] has yet to report a meaningful profit." Yeah, the company that so many other companies rely on to provide their online infrastructure isn't making money.
Evans does go on to say that the lack-of-profitability isn't a fatal flaw in Amazon's case, and since he's writing for the guy who pretty much invented the Web browser, he's worth listening to. But the bottom-line fact is that Amazon is not making money and betting your business on another business whose business model is not resulting in profit may be risky.
The point of this whole discussion is to make sure you keep the risk factors in mind, no matter which vendors you partner with. Retooling or transferring everything you're doing is possible, but it can be very painful and very expensive.
My recommendation is you develop a backup plan, a Plan B, and always keep it in your back pocket, just in case. Because despite promises, despite reputations, despite years of good experiences, I've long learned that "just in case" happens.