“All work and no play makes Jack a dull boy…”—James Howell (or Jack Nicholson in the Shining)

I don’t live to work. I work in order to live. Am I saying that I dislike what I do or where I do it? Really, it’s quite the contraire. However, I have a huge number of things that are more personally satisfying to do on my nights, weekends, holidays and vacations than being on a call or in the office because a production issue has occurred and our critical services are down.

As a DNS administrator, it seems like this state of being is potentially the norm. Availability of online services is a key component of today’s businesses; whether it’s a corporate website, commerce application or an email server—if any of these go offline, it costs the company money. Unfortunately, issues happen—servers go down, hardware fails, applications have defects. Additionally, even general maintenance needs to be performed in the black of night so that any impact is minimized.

How would a managed DNS service limit exposure and help provide notification and failover capabilities? What should you even be looking for in these types of services?

• Proactive Monitoring: With a DNS failover solution, it is necessary to be able to configure monitoring against each of your key assets. Being able to test the availability of your assets via one or multiple different methods (HTTP, DNS, TCP, Ping, etc.) at regular intervals (anywhere between 15 minutes to 30 seconds depending on how quickly you want to react to an issue) is invaluable. Additionally, even if you only use a single probe type, you want to make sure that you can monitor from multiple global locations and can set the number of locations that have to agree in order for a threshold to be considered violated. The strongest type of monitoring solutions will additional allow user to configure multiple thresholds. Non-failure thresholds allow for the potential of other performance actions to be taken prior to a failover event ever happening.
• Automated Response Rules: As mentioned above, looking for a solution that allows for the flexibility to configure multiple different monitoring thresholds and tie them to automated response behaviors is a huge advantage. Having the ability to receive advanced notice of declining performance or having an automated threshold that not only notifies, but begins to shift a percentage of Internet traffic away from poorly performing resources is very valuable and could allow you to avoid a failure altogether. Most importantly, the ability to remove a poorly performing resource and replace it with one that is performing well (be it locally or in another data center / region of the world) is the basis of a failover capability.
• Scheduling Events: Being able to manually change resource priorities and remove or add resources during maintenance are generally your only option with most DNS failover solutions. However, having the ability to schedule activities in advance is extremely convenient. An automated scheduling feature should ideally allow users to not only set a date / time to start and end the event, but change which resources are to be served, which are available as failovers during that time and even the monitoring thresholds that are used at that time.

The Neustar UltraDNS services for SiteBacker and Traffic Controller were specifically designed to not only help protect against downtime by having a failover capability, but also to proactively make management of resources easier. Will these services completely eliminate the occasional night or weekend work that can occur? Unfortunately, no. But…..they can extensively lessen the impact of those events and let you get back to living you life (and stop being a dull boy / girl).

Written by James Fink, Director of DNS Traffic Services, at NeuStar