just one more geek in a sea of austin techies

October 30, 2012

Gawker DR Scrutiny #DisasterRecoveryGeek

Gawker, Lifehacker and 
others surprise with no 
Business Continuity plan. 
At the time of this writing the latest "storm of the century" (hybrid storm/hurricane Sandy) is still working its wrath across a major portion of the US. The expected devastation has occurred across a number of states including loss of life, property, coastline, etc. Power outages and flooding were a given but what is unexpected is the loss of some well-known websites that, apparently, were hosted only in Manhattan. As in the island of Manhattan. Who would choose to have their popular, highly-trafficked websites hosted on servers located on a tiny island using an infrastructure already pushed to the limits by a densely-packed population? Here are a few such websites...

Gawking at Gawker
I am finding it difficult to believe a site so immanently popular as Gawker did not have cloud-based copies of itself as a go-to in the event that something happened to their island-based web servers. It's commendable that Manhattan-based Gawker gives its business to an ISP/web host with hyper-local presence (also in Manhattan). But the Gawker Media CTO should be taken to task for not having the Gawker site replicated to hosting locations far outside of the now partially-flooded Manhattan. What exactly is the Gawker Business Continuity plan? The site has been coming back online in bits and pieces so we can assume that there was at least a Disaster Recovery plan. However, in today's reality of dime-a-dozen cloud service offerings there is no reason a well-funded website should ever have to resort to Disaster Recovery.

Gawking Even More (Gizmodo and Lifehacker)
Businesses should not rely 100% 
on island-based data centers.
Let's not overlook the pedigree of Gawker Media's website family. The company operates technology-focused websites Gizmodo and Lifehacker and science-based website io9. Gizmodo and Lifehacker are the very types of sites frequented by IT professionals who are tasked with providing business continuity and disaster recovery options. Here's a Lifehacker post on technical Business Continuity.  Here's a Gizmodo post on Disaster Recovery. Here's another that is currently offline due to flooded web servers.

Rounding out Gawker Media's recognized websites are Kotaku (video games), Deadspin (sports), Jalopnic (cars) and Jezebel (women's interests). Collectively the websites boast visits by billions of unique visitors. I can't say what the threshold is for needing to expand into mutli-location, failover-ready hosting but it's safe to say that Gawker Media is past that point.

Good Plans Sometimes Fail (That's Why We Test)
I don't know anyone directly connected with Gawker Media (GM). I certainly am not privy to the company's operation and contingency plans. It's completely possible that GM did have a business continuity plan and that it simply did not work, resulting in the temporary loss of GM-owned websites. So far, though, messages from the company have only referenced flooding in Manhattan as the source of the problems.

If a business continuity plan was in place but failed then it suggests the plan has not been adequately tested and/or maintained. If there was a DR plan but no BC plan then it reveals a surprising lack of foresight. In either case it seems that Sandy has provided what should be a wake-up call to businesses of all sizes whose disaster recovery plans stop short of answering business continuity concerns.

No comments:

Post a Comment