Lightning Strikes Twice as Amazon Becomes A Lightning Rod.

Amazon customers must feel a little like Roy Sullivan. You may have heard of Roy—he holds the Guinness World Record of being struck by lighting the most—seven separate times during his 35 years as a U.S. National Park Ranger.

Amazon’s EC2 service outages have been well publicized over the past six months. And lightning struck again—literally, this time—in Dublin, Ireland on August 7^th, sparking a blaze that took down EU-WEST-1, Amazon’s only datacenter in Europe. Adding insult to inferno is the fact that, even as Amazon’s only European datacenter suddenly became darker than a Cormac McCarthy novel, many of Amazon’s European customers are left with very few options. That’s because data compliance mandates force them to keep their data in the region.

Later in the week, Amazon suffered yet another serious failure that left customers Foursquare, Reddit and Netflix without service (the second major outage for Amazon’s US-EAST datacenter this year so far).

During this latest outage outrage, some customers complained of “little or poor feedback from Amazon” regarding the cause or anticipated recovery timeframe. Hmm, it might be worth reviewing Two Lawyers Walked into a Presentation, my May 26^th blog that addressed negotiating service contracts with cloud providers.

I blogged about the first major service outage at Amazon’s US-East datacenter on April 28th, and offered suggestions to help protect companies that outsource workloads to the cloud. You might want to check out that blog (and review your cloud outsourcing strategy). My previous recommendations bear repeating:

1.         Don’t put all your eggs in same cloud-computing basket, or your eggs my end up getting fried.

2.         Manage your IT assets with the same rigor as you do in your datacenter regardless of where they are—even in the public cloud.

When it comes to managing and monitoring your private cloud, NetIQ offers very powerful tools. In fact, these tools also help you prepare for unforeseen outages, allowing you to perform contingency planning and define and automate best-practice responses to various failure scenarios—before a crisis occurs. If you haven’t already done so, I’d suggest checking out Novell Operations Center which provides a dashboard view of the interrelationships that ensure delivery of critical business services.

Well, I’ll leave you with one more weather factoid to consider: The most common type of lighting is known as intra-cloud lightning—lightning that occurs within a single cloud—or is that a single cloud provider?

–Richard