Gray Clouds: Recent Slips Worth Noting

Intuit Services Outage Draws Angry Comments From SMBs

News item: Wikipedia Suffers Fourth of July Outage As explained, Wikipedia’s data center in Florida suffered a power outage, impacting the world’s largest open-source encyclopedia. Wikimedia officials said that, anyway, they’re now going to add a separate data center to minimize disruptions.

News item: Amazon.com Experiences Online Service Disruptions. Amazon.com, arguably the world’s largest provider of data center infrastructure in addition to its online retail business, experiences a site outage that makes accessing the site difficult if not impossible for several hours. Amazon provides only a thin answer about what actually went wrong.

Within the span of a couple of weeks, three of technology’s out-and-out superstars experienced significant service disruptions in the cloud, with millions of dollars in lost business -- with very little outrage. It’s almost as if dropped cloud or disrupted online service is now brushed off with the same, aggravated-but-’oh-well’ acceptance of a dropped cellular call.

After all, these haven’t been the only big outages this year, there have been a number of others.

It’s likely that these three outages were just poorly timed coincidence. But it’s also worth keeping our eyes open for further disruptions in the cloud. Large data centers that host large-scale business operations are pretty complicated beasts. They require massive investments to build, massive investments to maintain and massive investments to upgrade. Conditions outside the data center -- such as power supplied by third parties - can play havoc with operations. (We’re seeing more and more publicly traded cloud companies warn investors, for example, that they could be subject to the negative risks if their data centers can’t get enough electricity from local power companies.)

It’s also going to be worth seeing how transparent companies are in the future. Frankly, when they suffer an outage, they need to err on the side of providing too much information, rather than not enough information. Customers and partners need to know specifics: Was it a software problem or a hardware problem? Was it a power issue with failed redundancy? Was it an act of nature in a particular geography?

And these are just for outages. The issues of latency and performance are an entire discussion altogether.

Right now, there does not appear to be much outage outrage - at least nothing that is more than a one-day story. But let’s hope these stories don’t start stringing together any more than we’ve already seen.