In Data centre best practices have always been top priority for most IT teams across industries. In fact, data centres strive to provide customers the most modern and efficient tools, both hardware and software, ensuring that the best practices are deployed within data centres to make them more green and efficient.
A testimonial to this effect is APC Schneider Electric, the data centre energy efficient solutions provider, which has associated itself with Lawrence Berkeley National Laboratory to leverage certain best practices in bringing about energy efficiency at data centres and helping customers to deploy the best tools and approach to enhance efficiency.
As part of his visit to Schneiders data centre, Dale Sartor, PE, Applications Team, Building Technologies, at Lawrence Berkeley National Laboratory (LBNL), Berkeley, caught up with ITNext along with APC Schneiders senior executives in a freewheeling chat. The discussion focused on the best practices that IT teams could follow to enhance data centre efficiency, cut down power cost, lower TCO and optimise design elements.
Data Centre Criticalities and Challenges
Sartor recommends that the first charter for any IT manager or data centre operator is to challenge the conventional wisdom or approach to the data centre. According to him, the conventional approach says: Data centres need to be cool and controlled to tight humidity ranges; they need raised floors for cold air distribution; they require highly redundant building infrastructure; and IT and facilities partnership are key. But there have been new game changers to enhance the efficiency of data centres, and the traditional philosophy of having high tech buildings no longer serves any purpose as they are energy hogs, says Sartor.
Dr Satish Kumar, Energy Efficiency Ambassador, Vice President, Schneider Electric India, affirms that the increased redundancy built in data centres seem to have increased the cost of maintenance as the loads of economies are different.
Aniket Patange, Director, Professional Services, APC by Schneider Electric, finds the most challenging aspect to be delivering value while taking account of the people, process and technological framework of the data centre lifecycle, even as there is no standard operating procedure to reduce cost.
The common challenge that these executives experience is that data centres are energy intensive facilities as the demand for storage surges; server rack designs require more power, and there are power and cooling constraints in existing facilities.
Dr Satish says, As the cost of power and power required for infrastructure increases, even surpassing the capital cost of IT equipment, the rise in total cost of ownership is alarming.
Create Differentiators
The key to addressing these challenges is creating differentiators which can help organisations see the power cost lowered, energy saved and efficiency enhanced.
A logical calculation to data centre energy efficiency according to Sartor is: Data center energy efficiency =15% (or less). (In other terms, Energy Efficiency = Useful Computation/Total Source Energy).
The task of defining the energy efficiency mechanism throws up an immense opportunity for IT teams and data centre solutions providers. Aniket Patange, DirectorDatacenter Lifecycle Services, Schneider Electric, says the solution lies in evolving strategy around power conversion and distribution, server load/computing operations using server innovation, virtualisation, high efficiency, power supplies and load management strategy, and deploying cooling equipment such as better air management, moving to liquid cooling, optimised chilled-water plants, use of free cooling and heat recover and so on.
Dr Satish recommends that IT teams can create value by engaging with multiple levels of teams and also bringing about certain policy regulations and standards to enhance energy efficiency.
It is critical to associate with industry associations, be part of the consortium of thought leaders and bring in ISO 50001 standards to create required differentiators, avers Dr Satish.
Best Practices for Benchmarking
The most important feature in the data centre efficiency mechanism is to spot the data centre maturity model based on industry standards. According to Sartor, it is critical to benchmark for energy performance improvement and pick up a few learnings from peer comparison to identify best practices.
Sartor advocates that it is difficult to manage that which cannot be measured; and hence every initiative needs to be measured.
The key energy metrics, according to Patange, is calculating PUE and partial PUEs, utilisation and energy reuse, while the future revolves around computational metrics, for example, and peak flops per watt, and transactions per watt.
The ideal way is to go by Moores Law of providing miniaturisation: smaller, more energy-efficient transistors resulting in 1 million times reduction in energy/transistor size over 30+ years.
- It is important to go for a refresh of IT equipment for performance, as old servers consume 60 per cent of energy, but deliver only 4 per cent of performance capability.
- Performing IT system energy assessments is compelling to evolve IT energy usage patterns
- Decommissioning of unused servers is mandatory
- Virtualising and consolidating server and storage
- Cloud computing most ideal, as it can help in dynamically scaling resources over the Internet, balance different application peak loads and typically achieve high utilisation rates.
Besides, using the IT team effectively to manage IT energy saves energy and sets the goals to provide the same level of monitoring and visualisation of the physical space that exists for monitoring the IT environment. Last but not the least, deploying visualisation tools is critical for debugging, benchmarking, capacity planning and forensics.
N Geetha
Add new comment