Are we putting too much trust in the cloud?

This week Amazon suffered another outage to its Elastic Cloud Compute service – a technical blip which eventually took down Amazon Web Services (AWS) at its North Virginia data center. The outage had widespread implications, affecting massive sites such as Reddit, Pinterest, Github and Foursquare, and taking down smaller businesses’ sites with it. TeamworkPM, an Irish provider of project management SaaS, saw its service collapse for over two hours on Thursday 22nd. AirBNB’s website also went down, and Minecraft was completely unavailable to gamers.

It’s not the first time, of course. Between April and August 2012, AWS suffered three separate outages. All of them were catastrophic. The causes ranged from server issues to power failures triggered by freak storms near its data center in Ashburn. The domino effect brought Netflix, Instagram, Quora and other enormous content hubs to their knees; it also brought down some of the most popular social sites on the web.

The cloud is robust, at least according to the statistics. Computerworld reports that AppNeta recorded the best providers’ uptime at 99.9994 per cent, with even the worst topping 99.9 per cent each year. But when the cloud fails, the foundations come crashing down, and suddenly that seemingly tiny 0.01 per cent becomes a major issue. Bringing failed services back up can be time-consuming, too.

Enormous cloud providers such as AWS are now propping up some of the busiest sites on the web. DeepField Networks, a cloud intelligence company, says that around one-third of us access services in the Amazon cloud every single day. As much as 21 per cent of the traffic pumped through Amazon’s cloud services was traced back to Truste.com, a company that provides online privacy solutions to some of the biggest brands in the world, including Apple.

Are we placing too much trust in the cloud? Claims of self-healing technology are commonplace, and cloud hosting is sometimes marketed as a ‘100 per cent uptime’ product. Some analysts believe businesses are becoming complacent, placing important data and services in the hands of one provider, cutting corners on hosting costs and failing to consider the implications of an outage. But it’s easy to see why businesses assume AWS is ultra-reliable, but delve into the Terms of Service and you’ll quickly realised it doesn’t have an uptime SLA.

However, the fact remains that cloud computing is the most affordable way for big websites to shovel data across the internet, and a good resiliency plan can guard against problems when it’s part of a company’s overall availability plan. Using AWS, it’s possible to place all data in one location for a relatively low cost, but businesses do have the option of purchasing multiregion mirroring from Amazon to protect against outages. It’s a more expensive option, but after three problems in seven months, perhaps we’ll see more businesses spreading their bets in the cloud.

If you are looking for a unique marketing campaign, submit a brief now!

Claire

Images from: Free Digital Photo

Leave your reply

Click here to cancel reply.