Amazon Outage is a Heads Up for CIOs: You can outsource the service, but not the responsibility
The recent events at Amazon are notable for multiple reasons, but for those of us who have spent time actually working in IT there are no big surprises there. Technology fails. Big, integrated technology systems fail bigger, and Amazon is definitely one big, honkin’ system. No matter who hosts it—even Amazon—if it is technology-based, it’s gonna fail sooner or later.
The right enterprise management investments can notify you when a failure happens (or is about to happen), but they can’t prevent human error, hardware malfunctions, security problems, or outright sabotage.
Does the “Amazon problem” mean that companies should move workloads out of the cloud?Bring the service in-house? Or move to competitors? Of course not. It simply means that IT executives have one more piece of information for making decisions about how and when to outsource.
I have my own private guess about what happened, technology-wise, at Amazon, and I’ll be interested to see what Amazon finally has to say. Meanwhile, Amazon’s problems pose an interesting question for CIOs, particularly those considering moving some portion of their IT services to the “Cloud”. Discussions in Computerworld (http://www.computerworld.com/s/article/9216160/Who_gets_blame_for_Amazon_outage_?source=CTWNLE_nlt_dailyam_2011-04-26 ) and elsewhere are asking, “Who gets the blame?” And it’s likely the same question is being asked in many executive suites.
To put the question more generally, when problems occur with an outsourced service, who has the ultimate responsibility? From my perspective, the answer is pretty simple. I chimed in back in 2008 in an article I wrote for CIO Update entitled “Outsourcing: Honeymoon Over, Reality Sets In”– a title also drawn from personal experience.
Here’s an excerpt:
“There is no “one size fits all” outsourcing strategy. Picking and choosing from the multiple offerings on the (cloud) menu requires business knowledge, understanding of the industry vertical in which the business operates, the kinds of IT services needed to stay competitive in that vertical, and the availability of skills and expertise.
This is where today’s CIOs have an opportunity to earn their keep, because this “big picture” perspective is one that requires significant experience, knowledge and judgment. Within high performing IT organizations, strategically selecting outsourcing options, and skillfully combining them with in-house expertise and services, is high art. Such organizations understand what outsourcing is and one thing it is not: it is not an opportunity to cede responsibility for the IT services being delivered. While execution can be outsourced, responsibility cannot. And from a management perspective, multiple companies have testified to the fact that outsourcing creates as many management challenges as it solves.”
My view on this has not changed. Amazon is a business and it is obviously in their best interest, as well as that of their customers, to maintain the highest service levels possible. And while CIOs can opt to deliver all or part of a company’s IT service catalog from Amazon or any other cloud vendor, he/she can’t outsource the responsibility for delivering that service. The buck stops at the executive suite —which is why the bucks flow upward.
I had an interesting discussion with Michael Crandell, CEO of RightScale last week about how his customers are using IaaS. RightScale is a full featured platform for managing applications at scale, with scale consisting of either resources or users. RightScale runs across IaaS providers, including Amazon and Rackspace, and I asked whether any of his customers were requesting replication across cloud providers. My question arose because I know that “real world” IT often contracts with multiple network providers for redundancy of critical circuits. Michael told me that he is getting such requests, which was no surprise.
So here is the bottom line: In companies seeking cross-cloud redundancy, someone at the helm is using his/her knowledge and judgment to protect critical services from outages such as Amazon experienced. My guess is that more companies will be doing just that in the near future.
More expensive? Definitely. But if I were a CIO, I’d make the investment, knowing that I’ll look like a hero the NEXT time a cloud provider—or my own IT organization—goes down for the count.