Amazon Outage Raises Questions About the Cloud
Powerful storms rolled through a number of states, including Virginia, on Friday evening, knocking out power to one of Amazon's cloud computing data centers. The resulting outages present another stain on Amazon's record and, by association, the record of public cloud computing as a whole, but questions remain as to whether Amazon and the cloud are really at fault.
A severe storm with straight-line winds, called a derecho, rolled through Virginia on Friday night, leaving a significant path of destruction in its wake. Amazon's cloud computing data center in northern Virginia lost power during the storm, going down and taking popular services like Netflix, Instagram and Pinterest with it, along with a host of other services.
The outage began around 11:20 pm EST, and while Amazon reported that APIs were operating normally a few hours later, it still took several more hours for Amazon to repair the instances and get all services back up and running, according to Venture Beat.
This Amazon outage is the latest black eye for the cloud computing industry leader, which suffered an outage in Europe a few months ago due to a lightning strike and suffered a much larger outage last year, taking down some major sites for several days. This incident also comes on the heels of an announcement by Google that the search giant is entering the cloud computing space with the Google Compute Engine.
Cloud Adoption Concerns
Since the company is by far the industry leader, any damage to Amazon's reputation is also damage to the concept of the public cloud, but is it really fair? Cloud computing proponents highlight the fact that having a cloud computing partner with multiple data centers allows you to avoid the risk of a single physical event taking your service offline, but this obviously didn't happen in this incident.
New reports show that the blame may not fall squarely on the shoulders of Amazon. As this ZDNet article discusses, it may just be that the affected services did not want to pay to have their instances appear in multiple data centers, guarding against this exact type of event. While some may say that Amazon should offer this service without a charge, the reality is that running a second set of services in real-time would essentially double operating costs, which isn't something even a massive provider like Amazon should do for free.
While details from those involved on who is to blame are not likely to come out, especially if these services decided to save money by not opting for physical redundancy, the fact of the matter is that cloud computing as a whole is taking a hit in the arena of public opinion.
Some of the blame has to fall at the feet of the outspoken cloud proponents, who tout the cloud's benefits while skimming over the risks and not correctly associating the benefits with an appropriate cost structure. Some of the blame can also fall on Amazon's shoulders. While it's impossible to completely protect a data center from an event like a derecho, the length of last year's outage was definitely the company's fault.
New studies show that midsize businesses are the slowest business group to adopt cloud computing, and events like these may only strengthen the argument that the cloud isn't ready. However, proponents of that argument often forget that these same physical events can happen to any data center, and many would not be up and running again within a few hours.
While it's always painful to have to turn customers away during an outage that was completely out of your control, the reality is that 100-percent uptime is just a fantasy, regardless of whether you use the cloud or not. The public cloud has a number of strong benefits and some risks regarding data security, but a handful of minor outages shouldn't be a major part of the discussion when choosing a cloud provider. As long as IT managers at mid-market companies don't make promises that the cloud can't deliver, only good things can happen when moving your business into the cloud.
This post was written as part of the IBM for Midsize Business program, which provides midsize businesses with the tools, expertise and solutions they need to become engines of a smarter planet. Like us on Facebook. Follow us on Twitter.