In challenging economic times, every organisation must be seen to be evaluating their operations with a view to making them as economical as possible.
This is particularly true of data centres, which by their nature consume a lot of energy, with implications for cost and environmental performance. However, with the correct measures, data centres can be optimised to save energy and money. The most important question to ask is, "How do I control my energy costs without impacting the delivery of the critical IT services the organisation demands?"
It is easy to pass the buck and believe that someone else is taking responsibility for the data centre. It is important to realise that this isn't the role of a single person. In many organisations, IT procures the computers, storage, and network equipment while the operating expenses are absorbed by other functions in the organisation. As such, to assess the current state and future possibilities of the data centre there must be an alignment of the numerous people involved in its construction and maintenance.
Both IT and facilities departments must work together and share responsibility for the day-to-day operation of the data centre. However, it is also essential that there is visibility to the other levels of the organisation. The CFO should also be engaged with the data centre operational costs as well as costs for procuring IT equipment.
Become aware- don't hide your head in the sand
It is startling how many managers don't know what equipment is in their data centre or how much electricity it consumes. Before any steps can be taken to improve efficiency, you must assess consumption and how much it is costing. As well as understanding how much you are spending it is also imperative to identify exactly where this is being spent.
One of the key challenges is the sheer size of your data centre.
Identify and analyse all existing IT equipment within the data centre, determine their business purpose, measure their power consumption and compare the current performance efficiency with a new server platform. The latest generation of servers feature built-in power monitoring via their out-of-band management capabilities, where this can be monitored. However, the vast majority of older servers do not have this ability. In this case other methods can be used such as the implementation of 'smart' power strips, or calculations made on a server's CPU utilisation.
This is the prime opportunity to remove any redundant servers that may be in the data centre. Anecdotal evidence suggests that a significant number of installed servers are not used at all by anyone but are kept in the data centre, taking up space and resources. By optimising the physical components of the data centre to current load and future predictions, there is the potential to eliminate up to 50% of electricity costs in real-world installations.
Evaluate the physical elements of the data centre
One of the key challenges is the sheer size of your data centre. While size can be reduced through the removal of redundant servers, also consider the possibilities of consolidating your most inefficient servers into a single new unit which will enhance performance, improve system utilisation and give you more space. This can also be financially advantages through immediately lower energy consumption.
When buying this new equipment, look for a higher-specification with newly developed energy saving capabilities. The design choices made for the selection of equipment for a new, expanded or retrofit data centre will affect not only capital and operational costs, but also the efficiency of the cooling system and data centre as a whole. Therefore, it is vital that those decisions are made by the entire data centre team, including facility owners, IT owners and those individuals responsible for capital and operational budgets.
It is also important that total cost of ownership (TCO) and return on investment (ROI) are factored into the decision-making process.
Alterations in the layout of the data centre alone can have significant implications on energy use and the efficiency of the air conditioning system. The ideal layout would involve hot-aisle/cold-aisle configurations with suitable air conditioner locations. The primary design goal of this floor layout approach is cool air and warm air segregation.
The location of vented floor tiles is also important. In an average data centre, many vented tiles are either placed in incorrect locations or an insufficient or excessive number of vented tiles are installed. The ideal design would optimise cool air flow by "tuning" floor tiles through varying locations and by regulating the proportion of vents that are open at any given time.
Maximise settings and software
Many data centres do not enable power management features within the installed servers. Power-management features should always be enabled to ensure most efficient operation of the server.
The installation of more efficient power equipment triggers is also recommended. For example, lighting can become even more energy efficient through the implementation of timers or motion sensors.
Power down servers when they are not in use
Certain types of servers will regularly go unused for random, lengthy periods of time - these should be targeted for powering down. Typical examples are servers found in test and development environments. The test team will know when a test run has finished and the server is no longer in use.
A large factor in data centre efficiency is managing data centre cooling. The cooling systems of today's typical legacy data centre are often highly inefficient, primarily due to the fact that cooling high-density equipment was not originally a requirement. However, today there are numerous methods that can implemented, both within existing data centres and new set-ups.
Continue to monitor and measure data centre performance
Once an efficient data centre is established, it is imperative to continue to monitor and measure its performance. The EU has launched a Data Centre Code of Conduct using The Green Grid Data Centre Infrastructure Efficiency (DCiE) metric to measure how much of total energy usage is consumed by IT equipment and how much is consumed by the facility itself.
Whilst this is a voluntary incentive from the EU, through continuous evaluation the data centre manager can identify opportunities to improve a data centre's operational efficiency, to compare efficiencies with competitive data centres and to determine opportunities to repurpose energy for additional IT equipment.
An efficient data centre has implications throughout the organisation, not least on the bottom line. Through systematic and careful evaluation, efficiency can be significantly improved without upfront investment.
As a global consortium of companies dedicated to advancing energy efficiency in data centres and computing ecosystems, The Green Grid provides neutral recommendations on best practices, metrics and technologies that will improve overall data centre energy efficiencies.
Vic Smith is Chair of EMEA Technical Work Group for The Green Grid and a Contributor to SearchVirtualDataCentre.co.uk.