If Thursday’s Amazon outage issue has taught us anything, it’s that server outages can happen to anyone, whether it be in your own building where your in-house servers lose connectivity or when one of the most respected and industry leading cloud hosting companies has troubles somewhere within their buildings.
These types of problems are completely unavoidable — no matter what precautions you take to keep the servers up and running. There is no way to account for every eventuality. As a business, you obviously can’t afford any downtime since this affects both your bottom line directly in sales and customers but also in confidence in your service and ability to be available when needed.
The only solution is redundancy in your network. While you can’t protect your servers from every possible problem, you can be prepared to handle the fallout from a problem by having several servers available in different locations. So when an unexpected hiccup strikes, your backup servers will be sitting there, ready to handle the load.
Of course, having multiple servers is a great idea – not just from a backup point of view but also from a load point of view. If a spike in traffic develops on your site with multiple servers, the load is shared across them and your site will keep performing just as well as ever.
Now, no matter if you are running multiple servers in house at different locations, multiple cloud servers all with the same provider or multiple cloud servers across several providers, the need to load balance them and automatically fail over to your backups when an outage occurs is vital.
The Dynect platform provides these functions as a core of its service and can handle any combination of cloud servers and in-house servers. Dynect provides basic monitoring itself but also makes it easy to interface with any monitoring service or application your company chooses via a robust set of APIs. In fact, a couple of sample scripts and scenarios have already been tackled in case studies with Gomez and Zenoss .
Many of our largest clients that use Dynect even go so far as to use redundancy in their DNS services — either in house or with a second DNS provider. Though with Dynect as the primary, those have yet to be called into action. (Earlier this year, we showed you how you can accomplish this with Route 53.)
Unfortunately, many companies don’t account for the innate need for redundancy until it is too late and it costs them business, like the Amazon outage. Take that proactive step and be ready for the unexpected now.
Kevin Gray is a technical integrator for Dyn Inc., an IaaS (Infrastructure-as-a-Service) company that features a full suite of DNS and email delivery services. Follow at Twitter: @tuftsmoose and @dyninc.