AWS outage exposes cloud failings

A cloud with a circle filled with an X on it.

Posted on Thursday, March 9, 2017

When Amazon Web Services went offline for several hours last week, businesses across the globe were affected. The inability to access key systems had a serious impact on operations – and revenue.

Amazon has not specified exactly what caused the outage, leading to speculation that a cyberattack or faulty router were to blame. If nothing else, the AWS outage shows that cloud systems are not completely invulnerable as many resellers claim.

The reality is that businesses need to re-evaluate their infrastructure, checking to see whether too much faith is being placed in hosted services.

Ignoring basic redundancy principles

Placing all of your systems in the trust of a single provider has always been a risk – even before the advent of hosted cloud services. Basic redundancy principles suggest that building a single point of failure into infrastructure will inevitably cause problems.

The distributed nature of cloud data centers should mitigate these risks in part, but if the issue is vendor-wide, you are still left with a single point of failure. Which is why an AWS-wide outage caused so many problems.

Time to revisit the private cloud

The flexibility of on-demand resources is incredibly attractive, allowing your business to adjust workloads within minutes. But there’s no reason this same versatility cannot be deployed on-site using a software defined storage infrastructure to build a private cloud.

This on-site cloud can be seamlessly linked to AWS (or other cloud service) providing the flexibility you need and helping to solve the single point of failure problem created by relying on a single provider. You can still allocate resources however you choose, with on- or off-site provisions able to take over in the event of an outage.

Best of all, a private cloud built using software defined storage can make use of commoditised hardware – including your existing post-warranty assets – to provide capacity at a very low cost. With SDS in place, you are free to build a single or multi-site infrastructure using replication and load balancing for increased resilience.

And when you do need to expand using Cloud services, the join with your SDS infrastructure is seamless. SDS allows you to take full advantage of built-in QoS to truly maximize performance of your storage infrastructure.

In doing so, your business will be better prepared to deal with an outage like that experienced by AWS customers last week.

Next steps

To learn more about repurposing your storage assets in an on-site private cloud scenario, please get in touch.