Data center reliability blues?
Hurricane Sandy knocked a lot of major data centers and parts of the Internet offline.
From this article on the San Jose Mercury News:
BuzzFeed began publishing section-by-section content to the blogging platform Tumblr. It also worked with Akamai, a content distribution network, so that some article pages would still load even if the core BuzzFeed site could not. Gizmodo and Jezebel each set up Tumblr blogs, “Sandy 2012 Emergency site” and “Post-Hurricane Emergency Blog,” respectively.
Gawker, too, relied on the easy-to-use Tumblr as a back-up. While Tumblr enabled practically seamless connectedness amid the chaos, the resulting aesthetic was a bit of a throwback to the spartan design of the Internet’s earlier days.
“How dumb to locate data center in a flood zone,” Gawker publisher Nick Denton posted to Twitter on Tuesday morning. “And how dumb to host Gawker servers there.”
So we have lots of data centers panicking and bailing water here, and falling back to Tumblr or other sites just to survive.
How do you survive when 25% of cell towers and broadband go down in a 10 states?
Not all disaster traffic is going into a black hole, actually. Weather.com received almost a Billion page views…
According to advertising analytics shop AdClarity, in only three days Weather.com — The Weather Channel’s website — has seen a whopping 960 million pageviews and 41 million unique visits from people scrambling to see Hurricane Sandy’s path. In fact, the site has had a 20 percent traffic boost this week and 30 percent yesterday, VentureBeat reports.
So the big question is, how do you survive such a disaster in your data center?
Take a cue from Kii Corporation, maker of the world’s largest Mobile Backend-as-a-Service, serving 25M end users.
From the recent TechCrunch article: Kii Cloud Opens Doors For Mobile Developer Platform With 25 Million End Users:
Kii is a second-generation cloud company. Its an engineering company, developed by veteran technologist who have brought syncing and database skill sets to create a service meant for the mobile market. But it’s the international flavor of this company that will have to play a factor in the company’s success. It’s the distribution of a company like NTT docomo that will make the difference. These are companies that have billing built into the services they offer, making Kii an attractive offering for developers looking to reach a global audience.
Kii provides the backend for the NTT DoCoMo service called i-Concier(tm). This service has had tremendous reliability in serving it’s millions of paid subscribers, even through big National disasters.
How does Kii provide such reliability? The same way as the Internet–through multiply redundant and parallel service provision points and a network of reliable switches. Obviously the Kii Cloud enables NTT DoCoMo users to have a backup of the content on their devices, but it’s also important to know that there’s a backup of the backup. In a redundant data center. In a completely different building. In a different location. Encoded into our synchronization protocol is the assumption of failure. 25% of cell towers in 10 states were taken out by Hurricane Sandy. Our synchronization protocols handle degraded connection cases, cut connection cases and have completely resumable data up and download capabilities.
Even the big players are not immune. I spoke to an app maker yesterday who said that fully 60% of their support calls were coming in as a result of the poor synchronization of data provided by Apple iCloud. This was true until they pulled iCloud out in favor of MBaaS. This provided this app maker with more control over their destiny and ultimately a better, more reliable service than even Apple can provide.
During these times of big chaos it helps us to reflect on what we expect from our clouds. The purpose of the cloud is to be without location and without time. The cloud is the eternal, reliable and sustainable soul of the device, the business and the beating heart of the customer relationship.
When selecting an MBaaS provider, please consider Vendor Viability. You want a supplier who has the experience and culture to withstand severe challenges going forward if you are going to rely on them to deliver such a key part of your business.