The Washington Post recently released a map showing various places the Internet went dead following Hurricane Sandy. A short article also appeared on io9, which highlighted the map and shared some interesting thoughts from io9 writer/creator/editor-in-chief Annalee Newitz. “In the future,” she writes, “disasters don’t just rob you of power and water, they rob you of information as well.”
Since your info is in the cloud, and since the cloud doesn’t work if your Internet access is down, no Internet means no info. Newitz also notes that because of outages like these, the cloud is a lie, saying, “There is no special place in the air that your data can go and remain unaffected by life back on Earth.” But really, did anybody ever think there was such a place?
The term “cloud” is simply a fancy name for web-based programs, platforms, storage, and so on. It’s beneficial for those selling cloud services to make the cloud seem infallible, but that’s not our approach.
When we talk about disasters, we’re obviously not talking about global devastation that affects the entire earth—if that’s ever the case you won’t be worried about your data, you’ll be worried about survival. Instead, regional disasters are the focus. Your data will be safe from a regional disaster if it’s in a data center outside of your area. This means that to a certain extent, the cloud is a magical place that’s unaffected, not because it’s floating in the sky, but because it’s outside of your area and is therefore away from the local emergency. Suppose, however, that your data center is affected by the same calamity you are. Mirroring your data to two data centers means at least one of those data centers will hypothetically have your data. But that still leaves the connectivity problem.
The more you move into the cloud, the more dependent you are on it. As our partner Guy Baroan mentioned in our most recent ebook, “Making Disaster Recovery Easy”, every time somebody moves anything to the cloud (especially production equipment), Internet access becomes critical. In these instances, a smart business will have a secondary Internet service provider, just in case the first one goes offline. When your business fully relies on the cloud, it fully relies on the Internet, meaning that with only one ISP, you’ve got one point of failure that can shut your business down following an emergency like Sandy.
Total Internet blackouts can happen, but if you’ve got two ISPs, it’s much more likely that you’ll still have one of them in the face of disaster. And according to the Washington Post, the map shows locations with more than a 30 percent decrease in the number of listening hosts, which means the Internet wasn’t totally out for everybody. If you look at the dots on the map, the outage isn’t totally devastating considering some of these areas may only have a 30 percent decrease in connected hosts. For any businesses that may have had two ISPs, it’s even less likely that they’d have suffered an outage, meaning they’d have had access to all of the cloud-based services they may have been using.
In any case, disaster-prone areas are at risk of having an Internet outage. Heck, even areas that don’t commonly experience disasters can lose Internet access—it happens all the time. That’s why it’s useful to have redundant Internet access, especially for those that use cloud services.
The bottom line is that the cloud isn’t magical or bulletproof, and that’s why the best thing, as always, is to make sure you’ve got redundancies in place that will help you reduce dependencies, and eliminate downtime.
If you’re interested in learning more about the hype surrounding the cloud, check out our Recovery-Ability guide “Cloud, Fog, or Smog.”