In October, Amazon Web Services suffered an outage on the East Coast of America resulting in a number of websites going down. Amazon offers companies the opportunity to host enterprise applications and big data projects in the cloud and supports sites such as Reddit, Airbnb and Flipboard – all of which were affected by the outages.
Initially a group of hackers known as ‘Anonymous’ claimed that they had caused the outage saying on Twitter, “I’m attacking Amazon for about 4 hours but just posted now I had to have sure it was all down.” However, Amazon denied the claims.
In a blog post Amazon said, “At 10:00AM PDT Monday, a small number of Amazon Elastic Block Store (EBS) volumes in one of our five Availability Zones in the US-East Region began seeing degraded performance, and in some cases, became “stuck” (i.e. unable to process further I/O requests). The root cause of the problem was a latent bug in an operational data collection agent that runs on the EBS storage servers.”
Google has suffered similar issues, with users being unable to reach many, if any at all, of Google’s services for six minutes. They declined to comment beyond an apology but Wired said: “Google declined to discuss the matter further. But this massive outage — however brief — shows how tenuous our “digital lives” can be and how much we’re dependent on Google in particular. Google has gone to extreme lengths to minimize outages. But it too is fallible, and clearly, multiple services can go down in the event of an engineering mistake, technical malfunction, or natural disaster.”
So if two companies the size of Amazon Web Services and Google can suffer from outages, then it should make companies think about how important minimising outages is. Here are a few tips to protect your company:
1. George Crump of Information Week says that by using an Onramp solution, companies can adopt a cloud mirroring policy, meaning that data would be written to two cloud providers simultaneously. Although this may seem over cautious, it would mean that if one cloud provider were to suffer an outage, then your data would be safe with another provider, with the chances of two providers being struck at the same time being incredibly low.
2. You should never assume that someone is protecting you from an outage. Backing data up may not seem like the most technical idea but it is effective. Whether you take care of it yourself or you make an arrangement with your provider, having backup data ensures that when an outage does occur that your business is not left in the lurch. Learn more about data center development here.
3. Make sure that your service level agreement highlights the consequences of a loss of service, microsoft azure direct connect is one of the popular business cloud solutions. It is important to keep some of the application or data available if and when there is a disruption to service. Crystal Nichols of Unitiv Blog says, “Define an acceptable level of availability for the given cloud application. For example, if you’re relying on a cloud provider to give you disaster recovery services, you might require 99.999% availability.”
4. Consider what the risks of moving to the cloud are. If you are not prepared to risk losing availability due to an outage then the cloud may not be for you. It might be best to buy vps hosting instead. Although the cloud can no longer be considered a fad it doesn’t mean that every organisation should look to migrate. Crystal Nichols says, “It’s not that cloud solutions can’t meet your needs; you just need to go in with eyes wide open and be ready to accept the associated risks.”

About Christian Lawley

Account Executive at Prohibition PR - far too interested in football.