New Amazon outage raises Cloud resilience questions again

Last week power generators failed causing an outage to Amazon web services including Netflix and Instagram. The outage went on trip a bug in Amazon’s own load balancing systems that would normally allow work load to be distributed across different zones and maintain uptime. In this case the capacity distribution failed. This doesn’t appear to be something that Amazon could have predicted, but as we know, prevention is better than cure.

Here are four best practices you can implement when it comes to mitigating risk for cloud and minimising disruption.

1. Application Delivery Networking. It’s imperative that users have access to on-demand cloud applications. Application Delivery Networking can dynamically manage the provision of your apps, even if the instances that make up these apps are being delivered via multiple virtual servers.
    
2. Geographic Load Balancing. While the Amazon capacity distribution application may have failed, it’s important to recognise that many Geographic Load Balancing Systems work incredibly well to ensure uptime and distribute work across multiple geographic locations. This is a must to avoid having all your eggs in one basket.
    
3. Virtualised application security. Security in the cloud is just as imperative as it is all areas of IT. With virtualised application security you can control who is using your cloud based apps and also deploy blocks as and when required. This can also help protect from less visible attacks such as a Level 7 DDOS incursion.
    
4. Virtual web application acceleration. If your site loads slowly then you’ll not only get abandoned by users, but also Google too, who penalise slow loading pages in search rankings. Acceleration for virtual web applications minimises the time it takes to shuttle data between data centres and deliver your assets to the user. Network level optimisation allows us to affect both the physical and the virtual to insure that your web apps are delivered promptly.