Data centre infrastructure management, or DCIM, is growing in popularity as a logical way to manage data centres, offering improved energy efficiency, power availability and other benefits. In this article, Tan Yu Ming, general manager at Kohler Uninterruptible Power, looks at how UPSs with the right topology, communications and software can contribute to a data centre’s DCIM strategy.
While data centre operators endeavour to manage ICT environments that are ever more complex, diverse and dynamic, they are also under pressure to be as energy-efficient and green as possible. Data centre infrastructure management, or DCIM, is gaining in popularity as a solution, with Gartner viewing it as one of today’s most significant areas of green computing.
DCIM looks at data centres and their internal functions from a high level. These include data centre design, asset discovery, systems management functions, capacity planning and energy management; together, they provide an overall view of the data centre, ranging from the rack or cabinet level to the cooling infrastructure and energy utilisation.
DCIM encourages efficient energy use, optimised equipment layouts, supports virtualisation and consolidation, and improves data centre availability. Ultimately, it allows balancing of computing and power capacity against IT load even as it changes rapidly; holding computing costs down in a dynamic environment.
As cooling and power systems, including uninterruptible power supplies (UPSs), contribute significantly to availability, capital costs and energy use, facilities managers as well as IT specialists are stakeholders in DCIM strategies. In particular, bringing the UPS into the DCIM environment provides significant benefits in securing data centre availability and efficiency. This relationship works best if DCIM/UPS communication is two-way. The DCIM software monitors the UPS status, and can make tactical or strategic decisions that drive changes to the UPS’s operation, maintenance plan or configuration.
The UPS’s contribution depends on its topology, communications capabilities and software. With a Simple Network Management Protocol (SNMP) connection a UPS becomes an intelligent network device, which can be interrogated by the DCIM. SNMP is a vendor- and platform-independent standard protocol that allows many devices to be monitored and controlled from a single location. It also enables interfacing with other network management systems that may be used by different DCIM subsystems and components.
Exact functionality depends on the manufacturer and software, but remote control and rebooting of UPS-protected devices over the network or Internet should be supported, together with protection of information through automatic and graceful shutdown of multiple critical load devices when an extended power blackout threatens UPS battery autonomy. Activity and alarm logging should be implemented, so that operators are made aware of problems and can resolve them before they cause outages.
Other desirable features include security through SSL and SSH data encryption and authentication, and automatically serving web pages in the selected local language. Reaching engineers and managers away from the control desk with real-time email, mobile phone or SMS alarm messages is also useful, or in some cases essential. Regular status updates are important, as they can highlight trends such as frequency of power cuts that may warrant further investigation.
Battery monitoring and charging control are critical, as battery health is fundamental to the UPS’s role. Frequent measurement of cell internal resistance, temperature and voltage gives an indication of battery condition and developing problems. Faulty cells can be identified and exchanged; additionally, a satisfactory battery health report tells operators that the backup autonomy will actually be available as specified when an extended power blackout does occur. Access to this measured data is important, as a battery’s quoted design life cannot be relied upon. Operational and environmental factors, such as charge/discharge cycles, depth of discharge and elevated ambient temperatures, can compromise this figure significantly.
The best UPS topology today uses modular design, as this allows flexibility and scalability fit for modern dynamic data centre loads. One modular UPS system, for example, comprises a vertical racking enclosure that will accept from one to five 100 kVA modules. This eliminates wasted power in unnecessary hardware, as a system can start with just enough capacity for the load. Even if DCIM reporting forecasts a rapid increase in demand, this can be met in minutes – without interrupting protected power – by plugging in more ‘hotswap’ modules as required. Up to six enclosures can be paralleled to extend capacity right up to 3 MVA.
This modular approach ultimately contributes to UPS availability as well. As extra capacity can be added incrementally by plugging in another module, N+1 redundant configurations can easily be set up. For example four 100 kVA modules can share a 300 kVA load, so that if a single module fails the others can still fully support the load. The fast addition of a “hot-swap” module releases an extremely low MTTR, which contributes to higher UPS availability, with figures up to ‘six nines’ (99.9999%) being achievable.
As data centre loads become ever larger, more complex and changeable, DCIM offers a logical management solution while achieving greener, more highly-available data centres. UPSs with the right topology, software and communications capabilities can contribute significantly to these objectives.