The OSS canary

Life for an actual canary in a coal mine could be described in three words: “short but meaningful.” Early coal mines did not feature ventilation systems, so legend has it that miners would bring a caged canary into new coal seams. Canaries are especially sensitive to methane and carbon monoxide, which made them ideal for detecting any dangerous gas build-ups. As long as the bird kept singing, the miners knew their air supply was safe. A dead canary signaled an immediate evacuation.”

OSS are used for many things – improved efficiency, coordinated / repeatable activities, customer experience management, tracking system performance, etc. They can also be used as the canary in the coal mine, providing a leading indicator when something is wrong with the system and is about to break.

It often takes quite careful analysis to learn your specific system well enough to identify what the leading indicators are. Then if you’re then able to understand your system well enough, you will know which levers to pull to bring the system back into a controlled state, giving you the ideal closed-loop solution proposed by exponents of predictive analytics.

The predictive OSS industry has taken significant steps forward in recent years but unfortunately, most OSS still tend to act as trailing indicators, helping operators identify where problems have occurred and hopefully help with a resolution. Exacerbating this is the tendency for operators to have monitoring mechanisms and response processes that just aren’t fast enough to prevent failures from occurring.

Even older reactive alarm systems can be fine-tuned into being a canary for your network. Once you can identify many of your leading indicators, the next step is to ensure that you have an early warning system to monitor in real-time and know your levers that rapidly respond to those indicators. Easier said than done!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.