Today, Amazon Web Services begins rebooting something like 10 percent of all Amazon EC2 instances to do a Xen hypervisor patch. That's a lot of instances. But if you're affected, you would have gotten an email, and you've probably planned for it! You're fine. You're totally fine. Probably. It may even be done by the time you read this.
So why am I bringing it up? Because, dang, Amazon Web Services: For the cloud service provider that basically invented the idea of being a cloud service provider -- for a division that may be worth as much as $50 billion by next year depending on which analysts you believe -- AWS does a terrible job at telling anybody what it's doing.
Amazon customers started getting the aforementioned emails on instance reboots earlier in the week, at least one of which made its way to our colleagues at Gigaom, prompting a mild but persistent state of panic amongst AWS users -- instances were getting rebooted but nobody was really sure why, except that it was to fix a "host vulnerability."
AWS Evangelist Jeff Barr published a blog entry on Thursday, confirming that the maintenance began last night and goes through Tuesday. It came with additional confirmation that the critical bug is with the Xen hypervisor, plugging a hole that was revealed last week.
Smarter people than I have explained this patch in more detail, but it's a doozy, and requires a reboot. Which is cool, and good on AWS for taking steps to protect users. But until that post went live, people could only guess why their often-critical production instances were getting a forced reboot -- even the article linked before uses a lot of guesswork to figure out which Xen vulnerability in particular was getting the patch.
Amazon Web Services is the Big Guy in the cloud. It was doing cloud infrastructure-as-a-service before it was cool, and continues to outlast the competition. The ever-shrinking margins on cloud infrastructure, by way of the eternal cloud price wars between Microsoft, Google, and Amazon, do call into question how long this party can last, but for now, AWS is far and away the leader. Just about every startup and a pretty healthy percentage of large enterprises have at least some kind of presence on Amazon Web Services.
So when Jeff Barr opens that Amazon Web Services blog entry with "Today I've received a few questions about a maintenance update," it raises some red flags. No, Amazon isn't beholden to anybody to reveal more about AWS. And in total fairness, Amazon Web Services has really picked up its game with talking to the developers who make up its core audience with more local events and its Re:Invent conference. But some days, it seems like all we hear about Amazon Web Services are price cuts, outages, and bizarre server reboots. It doesn't always inspire a huge degree of trust. So maybe -- just maybe -- opening up a little bit wouldn't be the worst thing.
By not being transparent about this one, AWS caused a lot of FUD where there needn't have been any, if it had made the public announcement first. As Drake says, communications saves relations. And while this turned out to be a non-issue, AWS is just too big and too crucial to maintain this veil of secrecy.