What You Need to Know
Although the tide of organisations signing up to the EU Code of Conduct for data centre energy efficiency is growing, up-take in the main has been slow. There are several factors that have contributed to this ranging from a less than impressive website, with no case studies or examples of organisations that have benefited from embracing the best practices, to a lack of understanding by the general data centre community regarding the ongoing commitment involved in signing up to the Code.
This is a shame as there are many organisations that have seen huge energy savings by adopting the best practices set out within the document. Add to this, the best practice principles have the proviso that they should only be adopted providing there is no negative impact on the delivery of business services. This gives data centre owners and operators the opportunity to adopt the principles that work best for their organisation, making signing up the Code a real “no brainer”.
What you need to know about the Code:
The main point, it’s all about energy efficiency, saving money and CO2
The second point, it’s good proven advice that really works
The Code of Conduct Document provides good advice and a thorough checklist covering all aspects of the data centre facility, from data storage to environmental set points. If you decide something is not right for your organisation then that’s fine. Nothing in the Code of Conduct expects you to compromise your ability to deliver your service.
So what does the Code of Conduct cover?
Data Centre Management – there are three distinct disciplines; software, hardware and physical infrastructure, that need to work together to ensure a data centre operates to is full potential. The first section of the best practice guide looks at creating a common Management Team made up of representatives from each discipline. This team should meet regularly and discuss all planned changes to the facility and the impact any changes may have to the data centre as a whole.
Data Centre Audit – an internal audit of the data centre hardware and software should be carried out. This will give an accurate picture of what is in the data centre and a starting point for the Data Centre Management Team. This then avoids buying anything new without first evaluating whether you could meet the requirement within your existing estate.
Levels of Resilience – look at the levels of software and hardware resilience. Ask the questions; are all of our services business critical? Do they all require the same level of resilience? If the answer is all services require 2N software and hardware to support the business needs, fine. Equally if your organisation allows staff members to store photos or itunes in a designated file area are these deemed business critical?
Hardware – when the decision is made to purchase new hardware make part of the process based on the energy efficiency and operating ranges of the unit required. The broader the environmental range of the hardware the less energy your physical infrastructure will use keeping the environment within those set points.
Software – review each process and ask what levels of resilience do I really need? What data needs to be stored and for how long? Can data be stored on passive devices? Look at how efficiently the applications work i.e. how much processing power is required to carry out a particular task? If your organisation writes its own software investigate how this could be done more energy efficiently.
Infrastructure Resilience – traditionally a data centre’s supporting infrastructure resilience level is decided and delivered to all the services within the facility, this could be 2N. Ask the question do all of our services require this level of resilience? Allowing some systems to be UPS and generator backed while others can carryout a graceful shut down, this will save energy in your data centre.
Power – having an uninterruptible, clean supply of power is essential to most data centres however many units, particularly if they are over three years, can have high levels of inefficiency. Ask the questions; how old is our UPS and what level of efficiency is it operating at? Swapping to a 99% efficient UPS may have a return on investment of less than a year depending on the size of your data centre.
Cooling – after the hardware itself the cooling system is generally the biggest consumer of power within a data centre. By extending the environmental ranges, such as temperature and relative humidity, your cooling infrastructure will have to do less work, in turn using less energy. Often airflow issues will need to be addressed before raising air-on temperatures to negate the risk of hot spots. The basic rule for this is to segregate the supply and return air as best as possible. Aisle containment can often provide a cost effective solution. (for more information on aisle containment follow this link)
Commissioning – make sure everything is set up correctly to your specific design and environmental requirements, i.e. it’s no good buying hardware that operates at an air-on temperature of 28°C if then the commissioning engineer sets the air-on temperature of the air handling units to 21°C. This sounds painfully obvious but if the Hardware and Mechanical Teams don’t have an opportunity to work together it happens, and we see it regularly. Commissioning a system correctly, and especially a system with free cooling, can make a huge difference to its operational running costs.
Maintenance – a regular, well structured maintenance schedule offers three major benefits; first any potential problems with equipment can often be identified before they become a problem, reducing risk and the interruption of services. Second, upgrades to equipment software become available and these often allow units to operate more efficiently, this is particularly relevant for UPSs where a simple software upgrade may make a 10% difference in energy efficiency. Thirdly, people working in the data centre may “fiddle” with settings on units, particularly air handling units, because they feel a “hot” spot or believe the system is operating incorrectly. These changes may cause problems that are counter intuitive, such as turning the fan speed up on an air handling unit to reduce a hot spot in front of a particular cabinet may actually make the hot spot worse.
SLA’s for hardware support – if you have hardware that runs at wide temperature and humidity ranges make sure your support SLA’s accurately reflect this. Some support SLA’s require environmental ranges far tighter than the hardware itself requires.
In conclusion the EU Code of Conduct for Data Centre Energy Efficiency is an excellent well thought out document. It will help data centre owners and operators with existing facilities as well as help those who are planning a new build or refurbishment project.
If you would like to sign up to the Code or improve your data centres energy efficiency by embracing its best practice principle contact info@future-tech.co.uk or complete this form.