Five Positive Outcomes from Amazon’s EC2 Failure
Although the downtime, well, sucked for a lot of Amazon's customers last week, the fact is that this was exactly the kind of situation that had to take place in order to rein in what was an almost out of control movement to the cloud. While I am certain that migrations to the cloud will happen over time, this failure reminds all of us that there are real risks that need to be addressed and that the decision isn't necessarily a foregone conclusion, nor should it be. Here are five positive outcomes that I believe will come from this outage.
- People will pause and think. Let's face it – cloud marketing has been at hype-level for quite some time. This is akin to the days when the web was getting big and vendors everywhere glommed on to the term "Internet" to sell their wares, even "Internet computer desks." Cloud-based businesses are popping up everywhere and the hype factor is huge. Although there have been significant failures at both Google and Microsoft in their "cloud" offerings, these events have been largely forgotten as some of them affected "free" services for which there is not always a high level of expectation. That said, they were still outages. The recent Amazon outage is an opportunity for all companies considering cloud moves to thoroughly review their business justification and ask new questions of their intended vendors.
- Organizations will be forced to understand the infrastructure. One benefit that is often touted when it comes to cloud-based services is the lack of need for internal IT people to handle those mundane, routine tasks that take up so much of IT's time. After all, if you simply throw your services at someone else's doorstep, it's their responsibility, right? Not so fast. If your company simply moves services to outsourced vendors without truly understanding the infrastructures that make those environments tick and reasonable availability precautions are not taken as a result, it's your company that takes the hit. The Amazon outage reminds us that leaving all of the details in the vendor's hands isn't responsible. It's incumbent on the customer to understand all of the availability options and make cost/risk/reward decisions. Customers moving services into the cloud need to deeply understand SLAs and have plans in place for when the inevitable happens.
- Vendors claims will be carefully measured. I don't believe that Amazon ever claimed 100% availability anywhere, which actually, makes me happy because they're being honest unlike so many other providers that promise 100% uptime and then provide a minimal recompense in the event of a violation. In fact the default SLA is 99.95% for each EC2 Region which allows the company 4.38 hours of downtime per region per year. I think it's safe to say that this SLA was violated in a pretty significant way last week. Obviously, most contracts have some remuneration clause when it comes to SLA violation. However, if you hear vendors ever claim that they are "never down" or they provide a 100% uptime guarantee, don't believe it. Sure, some vendors will use a gimmicky 100% guarantee, but there aren't generally the heavy hitters in the cloud space and the SLA violation penalty probably won't come close to covering the real downtime costs if your business is seriously impacted. Don't just rely on vendor claims to make your business decisions. Really read the SLA and make sure that the promise availability is actually attainable and that the penalties are reasonable.
- Companies will plan for failure. This is probably redundant, but bears repeating. Plan for failure. It will happen. Just because someone else is handling your infrastructure needs doesn't mean that they should be allowed to make your business decisions. That's up to you.
- Cloud vendors will all benefit. While some of Amazon's competitors are probably rubbing their hands together in glee, this outage should be a wakeup call to them to carefully review what happened to Amazon and learn from the mistake. I'm absolutely positive that a lot of very smart people at Amazon have learned some very hard lessons this week. Amazon will be stronger for it and all of us, vendors and customers alike, will ultimately benefit.