The rapid evolution of data centres has meant that critical energy monitoring technology hasn’t always kept pace with requirements of stakeholders. Over the last decade Data Centres have moved from a position of power-on at any cost, to one of driving operational efficiencies and the need for technology harmonisation.
The legacy approach to energy monitoring has often led to data centres critical energy being monitored, and managed using Building Management Systems (BMS). This is despite the BMS not being designed or equipped to carry out this purpose.
BMS are typically used to control things such as cooling, heating and lighting in a building, so they can monitor slow data and then slowly react to events. This system is not at all suitable when it comes to monitoring critical infrastructure within a mission critical environment, for which even a few seconds downtime can be detrimental to the data centres operation.
And while BMS have been around for a long time, meaning people are comfortable with them and the functionality on offer it can result in reluctance to change. BMS technology is now being supplanted by Event and Power Management Systems (EPMS), a system design to meet the current and future needs of data centre operations.
Measure, monitor, manage
An EPMS is a software solution which monitors electrical power, and provides end-users with up-to-date information on power usage across a data centre cluster, a site, a hall or a room, offering a host of business management benefits in the process including global compare and contrast of performance.
Oliver Farrell, EPMS Business Owner at Hanley Energy, explains: “An EPMS is a high-end power monitoring system allowing end-users to see all the electrical usage of innumerable assets distributed all across a data centre. The system reports on efficiencies and failures and, based on trends in the data, the software can predict potential failures so the user can avoid them happening.”
The main difference between a BMS and an EPMS is that the EPMS is designed specifically to monitor electrical usage and so offers a range of advantages.
Crucially, it measures, records and reports on all events to the millisecond, allowing for greater diagnostics of any incident. It also allows operational users to easily simulate different electrical loadings across a data centre, allowing for capacity planning. All achieved through a a simple on-screen site visualisation, enhancing end-user experience and accelerating decision making. These functions create a platform which provides greater reliability and efficiency, which contributes to reduced operating costs, and ultimately the mitigation of business risk in the form of expensive downtime.
Timing is everything
“BMS systems, whilst designed to be steady can sometimes be slow in event response” says Oliver. “Data which is 60 minutes, 10 minutes or 5 minutes old, is acceptable and that’s all a BMS is tasked to achieve. A BMS is designed for slow reactions to events. An EPMS is designed for high-speed results. In the world of Hyperscale data centres a second is both too slow and costly, millisecond data the entry level requirement.”
The cadence of reporting in milliseconds is important because of the advantage it gives in sequence-of-event recording. “A BMS will just report that something happened,” explains Oliver, “but it won’t tell you exactly what time it happened. An EPMS organises all events by millisecond timestamping and uses that as part of its report so you can see exactly what has happened and in what order”. This allows operational teams to better detect or analyse the cause of an event or a problem.
This, to-the-millisecond precision, coupled with site visualization allows end-users to monitor in real-time how the electrical system is operating across the data centre, and to identify problems immediately.
“With the site visualisation of the EPMS, you’re looking at the power distribution in its entirety, and it’s essentially like seeing fork lightning coming down from the top and branching out – ‘power on’ is green, ‘power off’ is red,” says Oliver. “You don’t have to drill down through 20 web pages to identify a problem, you simply click on the red section, within the visualisation and it’ll bring you straight to the incident. As a result the end-user can instantaneously see where a fault is and they are never more than two clicks, on any screen, from getting to the root-cause.”
Sharing the load
An EPMS allows for capacity planning – end-users can simulate extra load at any points throughout their site, and the software will calculate the stress that might put on the system. Thus allowing immediate understanding if tolerances or thresholds will be surpassed and gain a rapid assessment of the electrical system limitations. This effectively means you can plan to use your site to its peak efficiency – in the full knowledge and operational confidence that there is no danger of overloading the system and causing an incident.
Oliver says: “Think of plugging four electric kettles into the one socket on an extension lead, you’re going to blow the fuse. Capacity planning allows you to say ‘Can I put extra IT racks into this building’ or ‘no, because I will exceed the load, and I’m going to cause a problem by doing this’. You can put all the numbers into a hypothetical situation, and the system will come back and tell you where you are going to face a problem, it’ll go red on the screen to say if you put this capacity in, this breaker upstream is going to trip.”
This allows you to get the optimum efficiency out of your site, without the risk of pushing things too far. “If you’ve plugged in the server and you’ve gone live and suddenly you overload the circuit and you trip 10 racks, it’s too late. So, the idea of capacity planning is you prevent that from happening, you can pre-empt and avoid failures caused by adding extra load.”
Using data to solve problems before they happen
Due to the sheer amount of granular data recorded within the EPMS, there are various insights to be gained from its analysis – which is aided by the software reporting in a manner that is user-friendly and customisable to the user’s specific requirements. As data is accumulated over minutes, hours, days, months and years, you can compare and contrast energy consumption and use it to drive preventive maintenance operations.
Oliver gives an example: “If you had five generators running in parallel, and one of them was using more power than the others, you’d can isolate the problem. This is where the predictive function and the analytics comes into play.
“By analysing historical data, you can observe the lifecycle of the asset. For example, it allows end-users to determine if one of the generators is using more diesel than the others. So this could be an indication of a problem such as a failed bearing resulting in extra friction – analysing the data allows you to identify trends, and solve it before it becomes a big problem. This is essence of preventative maintenance, the identification and resolution of risks before they become a business issue.”
It’s clear that the future of critical power management is EPMS as they furnish end-users with the right tools to measure, monitor and manage complex electrical systems across a site ensuring optimum efficiency.
