Chris Smith highlights areas that need additional consideration.
Late last year a high-street financial institution experienced a power failure at an IT centre in Yorkshire, shutting down cash machines for a few hours, as well as undermining retail transactions and online banking. This incident was only the latest in a telling series of power failures affecting UK organizations including ISPs, hospitals and financial trading firms, highlighting the growing need for effective assessment of risks and disaster scenarios. This encompasses effective IT infrastructure planning, provision of power supplies and environmental concerns influencing organizations’ daily operations.
The scope for knock-on effects such as system failure and critical data loss has been intensified in recent years by the expansion, complexity and power constraints on ICT infrastructures as ‘UK plc’ migrates an increasing proportion of business processes and systems online. In 2008 Gartner made a global prediction that half of data centres will start to run out of effective power supplies. However, local conditions often present more immediate difficulties.
A number of global and local developments are complicating the management and operation of data centres. The Copenhagen summit has set the tone for more careful stewardship of resources by Western economies in the future, regardless of binding targets being set or not. The prospect of financial penalties in the years ahead for large organizations under the mandatory Carbon Reduction Commitment is forcing IT and facility managers to examine more energy-efficient operations, server virtualization, consolidation of IT estates and more comprehensive Green IT and Corporate Social Responsibility plans.
Constraints are also coming from one-off sources. The lack of data centre sites inside the M25 with effective power supplies and connectivity is already understood by business, but the ‘ring-fencing’ of power allocations in favour of the 2012 Olympics infrastructure has accentuated this trend for the mid-term.
All of these factors have contributed towards a rush of interest in high density computing infrastructures using more energy-efficient servers and cooling equipment. However, this tactical, or even haphazard, approach to cost and energy efficiency is threatening to put extra strains on data centres’ existing infrastructure and power supplies. It also highlights the split of responsibilities found in many data centre operations between owner and user. A more joined-up approach is vital and a number of key areas need to be addressed:
The IT infrastructure’s power supplies must be subject to regular reviews from IT managers. In the boom years, the usual practice was to build up high density server estates in an ad hoc fashion, without examining the wider impact of power demands on a data centre’s power and utility requirements. An understanding of the business’ expectations of its IT, allied to regular assessment of that infrastructure by expert providers can help identify current power supply issues, potential cost savings through system fine tuning, and future power supply issues, all of which will greatly reduce the potential of a power outage occurring.
High density IT infrastructures
The increase in uptake of high density computing (HDC) through technological advances such as blade arrays, has heightened the need for more sophisticated power management. HDC requires continuous planning and adjustment of associated cooling platforms. Virtualised computing environments for example require more sophisticated power management techniques to support the flexible movement of applications within the infrastructure.
The approach of simply increasing the number of air conditioning units within a data centre is a false economy, leading to increased heat output, and significantly higher heating bills. Often simpler heat management techniques such as configuring cabling to maximise free air cooling around units is more suitable and cheaper in the long term.
The adoption of so-called ‘green’ data centre technologies has created further enterprise risk. The energy demands of the efficient blade servers during start-up are far greater than that whilst idling or in use, causing a far greater strain on existing uninterruptable power supplies (UPS).
As the IT sector moves further towards being environmentally sustainable, regulation around IT infrastructures power usage and carbon footprint will undoubtedly add to businesses’ operational costs. The EU Code of Compliance was an important step forward in encouraging organizations to ensure that power costs and ensuring power delivery systems are both efficient and cost effective.
New data centre construction
UK government legislation such as the Carbon Reduction Commitment (CRC) and local planning guidelines could complicate and potentially add to organizations’ running costs, and also reduce the number of firms constructing new data centres. For example, in London, the Greater London Assembly (GLA) now specifies a proportion of recovered heat for new buildings. Organizations that are considering expansion of operations or strategic consolidation to a smaller number of larger data centres will need to assess demand and plan ahead on computing capacity more carefully. This will be an issue for some years unless a wider, more sustainable building development for the capital’s industrial and commercial properties is developed over the next few years.
The continual development of business continuity planning by CIOs is vital, given the increased risk of power outages. A broad understanding of power demands across the organization is needed, appreciating the purposes and likely timings of such requirements. Where critical IT infrastructures exist that must run continuously, the CIO needs to consider systems such as UPS that operate in conjunction with a generator to cover any gap (seconds, minutes or none at all as required) between a power interruption and the backup generator being able to provide full power.
It is essential that CIOs are actively communicating with both the data centre facility owner, facility management company and power utilities provider, to be able to successfully manage and assess the provision of power and develop plans to mitigate further risks.
Author: Chris Smith, sales and marketing director, on365. http://www.on365.co.uk/
•Date: 8th Jan 2010 • Region: UK/World •Type: Article •Topic: Power management
Rate this article or make a comment - click here