If you have a medical background or have had the misfortune to have spent time in a hospital, then the term telemetry may not be new to you. If this isn’t the case, then you may be wondering where the recent excitement about Network Telemetry comes from.
Medical telemetry refers to the machines that patients are hooked up to when they are at risk and need to be continuously monitored. There are many variations which measure important biological metrics such as cardiac rhythms, blood pressure and SpO2 (blood oxygen level).
Medical telemetry is an important part of critical care in any medical environment. It seems obvious that medical telemetry is useful but what properties make this the case? I suggest there are a few important properties:
- Utility of each metric that measured
- Sample rate
- Alerts when the metrics indicate a problem
Accuracy and the utility of each measured metric are the simplest of these to understand. If what is being measured doesn’t provide a strong insight into the problem, then there is little point in measuring it. Similarly, if the measurement isn’t accurate, the metric provides little value even if the underlying value is useful.
Granularity refers to tracking the metric at a size or unit that makes sense for the intended purpose. For example, it would make little sense to track the average heart rate in a hospital. This is still a measure of an important metric, but it doesn’t enable the doctors to make any useful decisions beyond knowing if there is at least one patient alive.
Not all metrics require a high sample rate but many do. A heart rate measurement that takes a sample every 10 minutes isn’t very useful but blood pressure readings on an hourly basis are.
Finally, telemetry data can be large and complex due to the number of metrics, their granularity and the sample rate. It’s not reasonable to have a doctor or nurse staring at the telemetry machines all the time – real time detection and notification of problems is required. Everyone knows the sound of a blaring telemetry machine from medical TV shows.
Bringing this back to networking, traditional network reporting and monitoring solutions do not meet the requirements for a good telemetry solution in several ways, the most important of which are the granularity and the sample rate.
Network telemetry is the extension of network reporting to higher granularities and sample rates combined with actionable metrics and alerting. One could argue that this is simply a difference in degree not kind. While there is an element truth in this, we believe the value goes up exponentially with the granularity and sample rate and therefore it is useful to use another term to communicate this difference.
You wouldn’t accept your doctor monitoring the average heart rate in the hospital and only taking that reading every five minutes. It’s time network monitoring evolves to network telemetry to deliver a similar level of insight and value.
If you like this post, please help us by sharing it with your network by using the icons here