Rackspace.com had downtime?

It’s a pity that Rackspace, a company with a stellar reputation for uptime and “Fanatical Support” experienced hours of downtime for clients in the Fort Worth data center. After putting in millions of dollars into infrastructure, how could something like this happen?
First, disclosures: I used to be happy Rackspace customer. I studied the way they did business and implemented it in my own dedicated server solutions business back in 2004. I’m friendly with the founders/former management of Rackspace and Serverbeach, and I’ve learned so much from their advice.
Just a few weeks ago, Serverbeach experienced a power outage. They took the necessary steps to get people back online within a few hours, but it was sure to cost clients thousands of hours of man power and money to fix. The damage is still there – downtime puts full time businesses out of order.
When a hosting company goes down, so many levels of the business chain are hurting. For example, I used to have servers at Sagonet, and when they had problems, all of my racks there were down. When my racks were down, thousands of websites were shut off from the world. Thousands of websites shut down mean that not only are my clients mad at me, but their clients are mad at them. And those clients may have to use these websites to serve THEIR clients, and so on. The loop can go on forever, and it all comes down to a datacenter hickup.
At that point, there’s nothing you can do but wait for the machines to come back online. And once they’re online, a shit ton of complain emails are bound to go to the datacenter, the leasers, the resellers, the clients, and even the clients of the clients. Companies like Rackspace who manage both the datacenter and the actually sold services are forced to compensate for lost time, or face terrible press and a huge loss of clientele.
It’s particularly frustrating for Rackspace that all of this happened from a stupid truck accident that brough a transformer down. The open letter says that they had to shut some infrastructure down just to get to the accident victim. I’m sure the high level Rackspace clients wouldn’t have minded leaving that poor soul to suffer if it meant keeping their websites running.
When I was in the business, problems weren’t as terrible, because I targeted small businesses. Many of them wouldn’t even notice downtime! Rackspace, on the other hand, serves clients such as Motorola and JC Morgan Chase. And when their services go down, the entire world is bound to know.
I’ll give Rackspace credit for doing the right thing. They wrote an open letter to their clients, took responsibility, and promised compensation.
Unfortunate for them, their reputation may be somewhat tarnished from this stupid accident.
Jessica Mah is a 17 year old entrepreneur, blogger, and sophomore in college. She’s currently the founder of a startup, managing editor at Startupism.com, and Jessicamah.com. In her free time, she enjoys the prospect of being an underage angel investor.
November 13th, 2007 at 2:31 pm
I’m not one to know much about hosting beyond the dreamhost for personal use+a couple dedis for others, but I’d imagine if I was running a business where downtime mattered that much, I’d have planned for situations like this. Downtime is inevitable, redundancy is expensive, but if downtime is expensive and inevitable albeit rare, multi-site redundancy can’t be a bad investment.
Just my two cents.
November 15th, 2007 at 6:33 pm
My first product (www.themindcanvas.com) uses rackspace.
We rent only one server from them, and we didn’t even notice the downtime. We got a personal call from our rep over there anyway.
They really do have “fanatical service”. (Unfortunately, they are also “insanely expensive”, which is why I don’t host big projects like SlideShare with them).
December 10th, 2009 at 3:22 pm
In the hosting business, no matter how many redundancies you build into an infrastructure, there is always going to exist a possibility of failure and downtime. We actually use Rackspace for higher end services like Microsoft Exchange having selected them for their responsiveness, which in the few downtimes we have experienced, has been excellent. Find a few good providers, and put your eggs in several baskets. And keep searching for other good providers.