What is the biggest challenge in the fight to make data centres more energy efficient? It’s a question of finding a more balanced approach to power, energy and environmental management. Ultimately, if this ‘balance of power’ is not as optimised and accurate as it should be, there’s a risk of compromised resilience and business continuity.
The unparalleled growth in data processing and the increasing power and cooling demands of new technologies has seen many data centres stretched to their limits. It is not unusual for older (legacy) data centres to be operating ‘blindly’ with power data limited to the building, room or switchgear.
As well as spiralling operational costs — where cooling systems alone can account for as much as 45 per cent of an energy bill — there are environmental issues to consider. A fundamental step is to understand what the facility’s overall power and energy consumption is. As well as the cooling, computer-room air conditioners (CRACs), chillers, heating and ventilation, key IT and network infrastructure factors should be taken into consideration.
While new-build data centres can achieve end-to-end DCIM relatively easily, legacy facilities face a bigger challenge as a result of site restrictions, existing legacy management and control systems, and limited or no monitoring within the FM and IT space.
In effect, DCIM calls for the full integration of traditional building management systems (BMS) with IT network management systems (NMS) and payload monitoring.
This provides a detailed and granular picture of data centre performance: from the point of entry into the facility to plant equipment (CRACs, chillers, UPS, HVAC), individual data racks and server payloads.
While a BMS will offer many high-level building monitoring and control capabilities, in many cases it may have limitations in the accuracy of monitoring and metering. Building management systems typically do not offer flexible energy management (and reporting or access by energy stakeholders), meter provisioning tools, meter management tools and visibility outside of the building or local IT environment.
BMS architectures also reflect a wide diversity of vendors and equipment. Rarely are they vendor homogenous and even a single vendor typically has to support a myriad of communication standards that have developed over the years. As a result, legacy systems have problems in addressing the more specific requirements of enhanced energy monitoring required today and may not have a high level of open communications or integration functionality. This poses the problem of sharing data with other third-party management platforms to meet DCIM requirements.
BMS systems are gradually evolving towards Internet Protocol (IP), making it easier for the integration of third-party management solutions. There is now, more than ever, the necessity for FM and IT to work together in gaining control of data centre management.
In addition to areas monitored and reported on by the network management system (NMS) — fault, configuration, performance, security and accounting management — these should also include data room and rack layout, hot and cold aisle configuration and containment, the recirculation of hot air back into IT equipment, and other hotspots that challenge cooling.
What is needed is a more holistic view of the entire data centre estate, from the building’s point of entry through to the individual server payload. This would unite the traditional domains of the FM department with those of IT, enabling all concerned to work smarter.
The necessary approach should factor in real-time as well as historic information capture. These must include the NMS, BMS, the point-of-entry fiscal metering systems, facilities asset sub-metering, individual IT payload load-monitoring and the integration of existing third-party management platforms.
The latest generation of intelligent remote power management solutions will incorporate: monitor only; monitor and control; per outlet monitored; and, controlled power strip variants. These allow the specification of the number of outlets, type of outlets, maximum power throughput, individual socket fusing and individual neons, as well as provide readings for RMS amps, RMS volts, total kVA, total KWHr, power factor and frequency.
Defined upper and lower thresholds allow user control over differentiation between warning and critical status, generating various alerts if exceeded. In addition, aggregation of configuration information across data centre equipment provides real-time identification of changes, scheduled polling, and the ability to capture detailed data to produce highly sophisticated and graphical management information reports.
Status information can be gathered either locally or remotely via secure authenticated IP access for monitoring RMS volts, amps, kVA, kWh, power factor and frequency.
Detailed values and management information across device location, power and cooling, network connectivity, asset application and customer information will enable analysis on current and historical trending that can be further used to inform decisions on data centre capacity planning, risk analysis, performance benchmarking and policing service level agreements.
Apart from ensuring adequate energy capacity for existing and future needs, a key benefit of intelligent power management systems is the ability to take accurate and dynamic power usage effectiveness (PUE) readings and gain the information necessary to take proactive measures in driving down the PUE still further.
To consolidate all of the power and energy data being monitored, a scalable enterprise software management platform is necessary. This must be capable of generating highly granular real-time and historical management reporting to measure, control and deploy energy optimisation initiatives, including: energy consumption; alarm status; PUE and data centre infrastructure efficiency (DCiE) monitoring; CO2 tonnage; temperature/humidity monitoring; capacity management; asset management; and CRC compliancy.
Organisations have begun to recognise the value and strategic importance of a more integrated approach between FM and IT within their data centres. This is a key driver in shaping the design of next generation power management and monitoring technologies, as well as the software platforms necessary for providing the all-important holistic management and reporting intelligence.
