Business Disruptions due to Amazon Cloud Outages: Lessons Learned

amazon cloud outageLast Friday is yet-another-cloud-outage coming from Amazon – the 2nd outage in 2 months time, actually. The outage has outraged the tech and business communities, and raised a classic question: Is the cloud reliable? Well, since the outage brings down major websites, the buzz around the web is that the cloud is – as many skeptics say – not ready for the prime time.

As found in Amazon’s service health dashboard, the cause of the outage is due to the power issues in Amazon’s North Virginia data center, most likely due to the severe storms in the region. The service impacted is Amazon EC2.

There are large sites – along with many smaller counterparts – hosted on EC2: Amazon brought down Netflix, Instagram, Pinterest, and Heroku – to name a few. The scale of the outage has made the Amazon outage talks last for a week – and perhaps more.

The outage has costs Amazon EC2’s users dearly – and the domino effect is really devastating. Service interruptions anger customers, and in some businesses, customers’ patience is very, very short. Just ask Brandon Wade, the owner of Whatsyourprice.com, a dating site. As reported by Wired, his web business boasts 400,000 users, and during the outage, he received 1,000 complaint emails per hour. Tarnished reputation and trust, the Amazon outages single-handedly bring a business down – with unknown ETA to full-recovery.

The 2 outages in 2 months have caused Amazon its reputation. And gaining trust is more difficult than ever, in my opinion. Amazon apologizes profusely, but it’s too late, I suppose. Things happen, and what Amazon (and its clients) should do is picking up the pieces and move on.

Lessons learned

Perhaps marketers – including those hired by cloud companies – should transform their pitches. The cloud once was touted as always available. On paper, the cloud technical ability is to take over the loads whenever a server is down. The fact is, when power is turn off abruptly, what’s look good on paper doesn’t look so good anymore in the real world.

Perhaps cloud solution providers should educate prospects about the potential of the cloud – beyond the virtually impossible 100% uptime and cost savings… simply because, in some cases, the cloud is not that reliable and can cost you more than non-cloud solutions.

Just like what Vincent Arden said in his comment on one of our article, the real power of the cloud is actually it its ability to enhance businesses’ ability in responding to changes, as well as in delivering innovation that can bring them competitive advantages. And for those purposes, private cloud solutions – or hybrid ones – are better than public clouds.

Finally, perhaps you must learn what Zynga does: Using Amazon EC2 as a backup, when Zynga’s own cloud – called zCloud – experience technical issues or overloads. Using the public cloud solution as a backup of your private cloud. Now that what I call smart move.

Your thoughts?

You might also like

  • Considering the volume, depth and phenomenal grown of Amazon’s services – their track record is awe inspiring. The public cloud is a great place to learn & launch but once you start to grow and have a business model dependent on a 3rd Party whose resources are shared with many others – you need a back up plan and like Zynga has learned that back-up plan is to move Private & Hybrid cloud deployment model. Private Clouds within your enterprise’s data center or hosted for you at a secure Private Cloud Hosting service provider’s facility are dedicated to your customers and come with SLAs, and when combined with Private PaaS allow you to still use public services like Amazon when appropriate (launching new games or testing new applications) – without being dependent on a single shared resource provider.

    Learn more http://bit.ly/PathToHybridCloudGlory

  • I agree with Diane. The ability to leverage Public cloud for development and testing has far reaching benefits such as financial, quick deployment of R&D, as well as controlled environments outside your private cloud.

    Thank you for sharing the “dark side” of the cloud in this article as you have clearly pointed out that the cloud still has the dependency on power and sometimes no matter how stellar a track record or how well designed and funded a cloud solution might be. There is still the potential for an outage.

    Bottom line – no matter what a cloud provider promises you in their SLA, it is the head of IT who is ultimately going to be responsible to have the proper business continuity plan in place that can be executed when the cloud experiences turbulent times in order to protect the interests of your clients or employer.

    Looking up

    @vincearden

    • Ivan

      Vince,

      Thanks for your comment – yes, I agree that regardless of the SLA, business continuity plan should be enforced to avoid “mishaps,” as the cloud’s growing pains will continue somehow (just like the recent cloud outages impacted major services like Twitter, Google Talk and Microsoft’s Azure.)