Strive for IT Excellence

Build a Resilient Architecture That Handles Cloud Outages

The cloud offers high resiliency, but it isn't perfect: read the service level agreement closely and you'll see the guaranteed uptime isn't 100%. Even big cloud providers like Amazon Web Services have had outages causing widespread impacts on businesses. Amazon's outage in early 2017 caused an estimated $150 million to the impacted businesses.

The costs of a cloud outage, like any on premises outage, include the cost of lost business, the cost of the recovery effort, reduced productivity, and potentially fines and penalties for noncompliance with regulations or contractual commitments. None of this means that you shouldn't use cloud, but part of your planning for cloud should be planning for a cloud outage.

Understand Your Responsibility

A research study by Veritas found that most IT workers believe the cloud provider is responsible for ensuring uptime and preventing application outages. While it's true that the cloud provider is responsible for cloud uptime, they aren't responsible for your business uptime. You still need to be prepared for an outage with plans that minimize downtime and get your business operational as quickly as possible even for applications in the cloud. Find out what your cloud provider's disaster recovery plan is and build on it.

 This means you need to be able to recognize a problem in the cloud and respond to it. You need a plan for monitoring your cloud in order to know when a situation is developing and a documented plan for responding.

 Extend The Cloud With Your Own Resilient Architecture

You can't just rely on the cloud provider for resiliency to keep your applications up. First, not all cloud issues originate in the cloud. The cloud can be running 100% normally from the cloud provider's perspective, but you've still got a business outage if there are network issues. You can reduce that risk by using virtual private networks or network services from multiple internet service providers (ISPs), though even separate ISPs can have shared resources that provide a single point of failure.

Build redundancy into your cloud deployment by using multiple regions or multiple cloud providers. Multiple regions from the same provider are a relatively straightforward solution. When you have a multi-cloud solution that spans different providers, you introduce issues of compatibility and data replication and synchronization. Hybrid cloud, where you have a private cloud on premises in addition to the public cloud, also has these challenges.

You should also build resiliency into your applications. Design them with flexible configurations and the ability to fail gracefully when services are unavailable.

Test Your Response

Your annual disaster recovery test should include simulating a cloud outage. This will enable you to determine the effectiveness of your recovery plan and improve your response strategy.

How resilient is your cloud? Work with the professional services team at dcVAST to design and implement your cloud and disaster recovery strategies. Managed services ensure your cloud is monitored 24x7 with a team ready to respond to problems immediately. Contact dcVAST to learn more about making your cloud and your business more resilient.