Amazon Cloud Outage: The Risks of Trusting Your Entire Business on a Cloud Vendor

amazon cloud outageHere is another sad story of cloud computing failure: Amazon cloud has experienced server problem, affected many sites, including the popular Reddit, Foursquare, Quora and many others. The outage teaches us a lesson: Regardless all the benefits of cloud computing, don’t rely too much on the cloud – have backups on your local machines.

First reported yesterday, there were connectivity issues affecting the database. Then Amazon also reported that their EC2 (Elastic Compute Cloud) were also having troubles. All in all, here is the biggest problem of all: Amazon cloud redundancies have failed to stop a mass outage.

We know the pitches of cloud storage services – “our data and information is stored in multiple locations in such a way that there is no chance for your data to be lost” – and something else like it. For websites, this backup redundancy will mean that your site would always be on, 24/7. In practice, it wasn’t that easy – at least in Amazon’s case. There was a glitch in the system that causes Amazon services to collapse.

For sure, the outage losses Reddit and the others money and productivity. I’ve visited Reddit and I found out that the website has this message on homepage for users: “reddit is in emergency mode right now because Amazon is experiencing a degradation. they are working on it but we are still waiting for them to get to our volumes. there is no ETA at this time, but we are trying to work some magic and will very slowly be bringing the site back up. please stand by.”

Reddit is definitely unhappy with Amazon cloud, and so does Quora, Foursuare and the thousands of others.

Lessons learned

Many experts have warned you and me. Unfortunately, we are exposed to all the buzz of what positive things cloud computing can offer you. I do believe that cloud computing can be very beneficial for businesses, but it seems that people are forgetting something that is more important: What if things don’t go so well, as advertised?

Before the cloud, when things go wrong, the impact is “localized” – your business servers failure won’t affect other businesses; and vice versa. With the cloud, the impact of failure is magnified and amplified – one single glitch can impact a great number of websites.

So, is this Amazon’s fault? Well, while their servers do experience problems and they do fail to deliver, but we shouldn’t put all the blame on Amazon. You and I – the cloud service users – should know all the potential risks of bringing our entire business to the cloud.

We should understand the caveats before we sign up with any cloud business service vendors:

  • Switching from one cloud vendor to another is a difficult task, if not impossible, as every cloud service is not built for such flexibility.
  • If we overly trust on one particular cloud vendor, our business fate will be highly related to the cloud service we sign up with.
  • Cloud computing is a new tech advances that evolves rapidly – you and I should expect hiccups to happen occasionally.

If you insist on bringing your entire business to the cloud – as experts suggested – please consider adopting the hybrid cloud model to allow you manage resources on-premise and externally – so that when things got rough from your public cloud vendor, you can always lean back on your private cloud.

So, to remind you: In cloud computing adoption, be sure you understand your risks and always be prepared for outages, regardless of what cloud vendors pitch you.

You might also like