Professor Ian Bitterlin gives his tips on how to save energy in your data centre by identifying the biggest resource drains
The spotlight has been on data centre efficiency for several years now. Yet inefficient technologies dominate most facilities, with utilisation rates remaining surprisingly low.
Often, even the simplest decision or change takes three weeks and six people to make. Although the cost of electricity is becoming increasingly significant, the business impact of those inefficiencies goes well beyond this.
Reducing energy and operational costs by half in a midsize data centre – a realistic goal with current technologies – could save businesses millions annually. Yet the bigger cost may be in the area of agility. An inefficient IT infrastructure is less able to respond to the demands of the business it serves. Therefore, the first step in rectifying this situation is identifying the systems and processes that are draining the energy – both electrical and human – from the data centre.
Mapping Your Minutes
With 70 percent of organisations still working toward the highest tier of data centre performance efficiency, it’s time to ask yourself if this is true of your business. While innovation and new business is intended to be the main focus within an organisation, the reality is that day to day operations often restrict this. When there are constant operational issues to address, carving out planning time requires extraordinary measures. Without planning, you’ll never get to a better balance of “keeping the lights on” as opposed to innovating. Planning starts with knowing what you have. Assess both your people and your environment.
Assessing Your Assets
Key to this planning is a clear understanding of your IT estate. What do you have in place today? Think about the age of your servers, what applications they are running and how often they are used, along with what power and cooling assets support them. Do you have an up to date map of both your physical and virtual layouts, knowing who owns the assets and your workload efficiency?
You should also consider your human capital as one of your assets, including the skill sets, training and certification of your staff. Analyse their work style, including levels of communication between staff managing servers, networks, storage and power, since greater innovation is likely to come from a pooling of ideas.
Finding the Efficiency Drains
When assessing your various assets, be they physical, virtual or human, here are a few common drains on your efficiency that are important to look out for:
Drain 1: Buffer Capacity
If you are going by nameplate capacity estimations, you are most likely operating with a 20 percent buffer, and many organisations won’t encroach on this for fear of causing downtime. Harnessing that buffer could defer a major addition or even a new data centre build. The right data centre infrastructure management (DCIM) solution will give you capacity information in real time and let you confidently use that capacity.
Drain 2: Ghost Servers
Having some way to measure energy use will let you make informed decisions regarding capacity, availability and optimisation and DCIM is one way to get there. Knowing what you have can tell you what to turn off – an idle server wastes nearly double the amount of energy it draws and getting a handle on this wasted energy saves money and capacity.
Drain 3: Inefficient Servers
While many organisations have embraced virtualisation, they have overlooked other opportunities to cut energy costs. Energy efficient processors and power supplies have been shown to reduce data centre energy costs by 18 percent and using the server power management software that’s likely already a part of your servers can shave off another 9 percent.
Drain 4: Heterogeneous Environments
If one person is managing five servers when another is managing 50, you may need to look at standardisation of equipment, tools and processes to make people more productive. Your physical environment should have a high degree of standardisation, with one staff member managing one brand of servers, using one set of tools and procedures.
If you are not in a standardised environment now, use your server refresh cycle to kick-start the process. Put together a plan to get to a mostly homogeneous environment in three years, aiming to standardise processes as much as possible.
Drain 5: ‘Needlessly High SLAs’
If your IT organisation is fairly standard, you answer your business needs as they arise. Every business client wants the best available network bandwidth, responsiveness, disaster recovery, and storage level available – until they assess the cost, at which point they see the value of matching service level with need. Make sure the level of cover matches the equipment; differentiate between cover options which are “nice to haves” vs. business critical “must haves”.
Drain 6: ‘Limited Visibility’
Make sure you have the right tools to enable your staff to be as efficient as possible. Does your staff get alerts when an issue emerges or discover them later on? Do they need to walk the floor to find space for a server or can they look at a dashboard that maps available capacity and “what if” scenarios? Critically, if you move toward a, highly available, utilised and consolidated infrastructure displacing old applications with new ones as you evolve, you’ll increase operational and electrical efficiency without impacting your current applications. More importantly, you’ll position your IT organisation to support dynamic business demands in real time.
The needs and priorities of every organisation will of course vary. But by addressing some of these immediate resource drains, you will have a strong foundation for the efficient data centre environment that your business requires to run in an agile and profitable manner.
Professor Ian Bitterlin is EMEA chief technology officer at Emerson Network Power.
Do you know all about Green IT? Take our quiz!