Since the Industrial Age, nearly every piece of commercial equipment invented has had some form of associated telemetry (also known as instrumentation). In most cases this telemetry has been put in place to confirm, modify, and diagnose its operational status. Analog or digital, nearly every device can benefit from instrumentation.
The advent of electrically powered equipment brought the ability to use much more advanced technology to look deeper, in new and exciting ways and extend the range of this telemetry. In addition, the analysis of raw data was no longer limited to a single point in time by a single user, but could be correlated, viewed, and studied by any number of users at any point in time.
Instrumentation has become critical to the operation of any type of complex system. The cockpit of a modern jet airliner houses thousands and thousands of points of data that are being gathered in real time, sourced by a wide array of individual manufacturer’s components in a variety of measurement units. Some key discrete points are presented without interpretation to the crew as they are gathered, while other points may be grouped and considered relative to other discrete points before the crew is made aware of this condition.
A modern data center is also a complex system. For a data center to be operated efficiently, the operator must gather as much information as possible from a vast array of dissimilar vendor equipment, normalize the measurement units, and then analyze this instrumentation to allow highly confident business intelligence. A modern data center must be designed to take advantage of all of the discrete instrumentation points across IT systems, power chain components, HVAC systems, and building security systems. Aggregating and normalizing these data points (in real time) is critical to achieving previously unattainable efficiency levels.
Connectivity has always been the Achilles’ heel of instrumentation plans. Typically, while some level of sensing technology itself has been available in the devices themselves, the ability to transport the resulting data created by these distributed sensors has been difficult. In early mechanical systems, the presentation of many sensors was physically restricted to the sensors’ locations (such as a railroad track switch position mechanical flag), or connectivity was addressed with simple technologies such as pressure transfer lines and round pressure gauges that indicated status at some distance.