Last June I started writing a technical paper on Green Capacity Planning. I felt satisfied with my work and was able to cover the topic in a 30-pages document. Then unfortunately the hard disk of my laptop met the failure and I lost all the work. This costs me more than that. Situation was like attending the same course twice and I couldn’t start it until recently.
In these 6-8 months the awareness about Green Capacity Planning has improved a lot and everyone is talking about it. Some questions from my professional network triggered me to once again start writing about it. I assume (forgive me If I am wrong!) that for 90% of the readers it is a new road to travel and those are my very target audience. As the topic is comparatively stale now, which means some of you already knew it; still I want you to assure that you will get something out of it. The reason for blog posts instead of a paper is that backups are automatically taken and who knows that at the end I will release a paper with more details. Another advantage is that you don’t have to wait for fully-fledged paper to be released.
So let’s step up a gear and discuss what is Green Capacity Planning. With the increasing electricity prices and tougher business conditions businesses are scrutinizing the power consumption and other operational expenses too closely and IT people have started concerning about the inflationary spiral operational expenditures. In the recent years a new market and practices have emerged which consider energy, cooling, space etc aspects during capacity review process for a data center. Because these practices are related to a much bigger “Green IT” initiative, it is commonly known as Green Capacity Planning or Intelligent Capacity Planning.
Why Green Capacity Planning?
Frankly speaking eye-popping electricity prices and increasing operational costs of a data center are the major driving forces pushing towards adopting Green Capacity Planning discipline. Today’s data center infrastructure is beyond the traditional IT equipments which include servers, network devices, and storage sub-systems. Now it also includes cooling systems, uninterrupted power supplies (UPS), lighting etc. The operational cost of these equipments is a significant part of total data center operational expenditures (Opex).
In the recent years due to the tough economic conditions there is an increasing pressure on IT teams to implement cost cutting measures. IT teams are already cashing in on the technologies like virtualization, server consolidation, and cloud computing. The awareness about green initiatives had given data center managers a food for thought, a way to go beyond the traditional Infrastructure performance management, to further cut their operational expenditures and thus improve the efficiency. As the data center managers continue to be challenged by the business to increase the efficiency in a cost-effective way, companies are beginning to realize that being “green” isn’t just good from a PR perspective; it can also make good financial sense.
Technology advancement is another reason for the need to go green. CPU in data center servers has been truly sophisticated in terms of power distribution. It dynamically switches to low-power mode and turn off cores in the system when there is low volume of work at hand. Despite of these improvements supporting infrastructure is still the same which had been made for older servers without dynamic power. As a result anticipating the power requirements for a data center has become challenging and results in either overloading or over-provisioning the electrical infrastructure.
Apart from cost and technology improvement there is an increasing pressure from regulatory authorities. Various agencies like US EPA (Environment Protection Agency), DOE (Department of Energy), EEA (European Environment Agency), LEED (Leadership in Energy and Environmental Design), The Green Grid, IEA (International Energy Agency) have been constantly working to encourage organizations to use sustainable energy. Similarly organizations are advised to report and put efforts to reduce carbon emission levels. It is not surprising that many organizations have already started measuring and reporting these metrics under their Corporate Social Responsibility (CSR) programs. After 2011 Japan’s devastating nuclear disaster world is looking for alternative green energy resources such as wind power, solar power and geothermal power. Governments are encouraging companies to source environmentally friendly electricity by means of tax reliefs, recognition. After climate change conference held in Durban in Dec 2011 world seems to agree on legal-bounded deal to limit carbon emissions. It means there would be more strict laws to reduce the carbon emissions by data centers.
Green Capacity Planning (What it is about?)
Typically IT Capacity planning includes collection of relevant workload and resource utilization metrics for IT components and analyzing it with the business demand to see its impact on the IT components and give business a view of when capacity upgrade/downgrade is required and most cost-effective way to achieve this without affecting the agreed SLAs.
The scope of traditional capacity planning includes:
- Computer power or CPU
- Memory- physical and secondary
- Space: Internal and external like SAN
On the other hand a parallel stream of specialized people take care of power, lighting and cooling aspects of a data center and they are known as facilities management or building management. However being IT as their customers they need to know the IT demand and impact of any changes in IT Infrastructure and demand on the facilities infrastructure. This missing link results in under-provisioning or over-provisioning of facilities infrastructure, poor time to market of IT services because of slow change management process.
Green Capacity Planning is a coordination effort of IT and facility teams to enable informed decisions about data center capacity and is a natural extension of IT Capacity Planning. This synergy brings optimal infrastructure sizing, cost-savings, and effort to save our natural resource – energy. The aim of Green Capacity Planning is to extend the scope of traditional capacity planning practice and include power consumption of the individual IT equipment and overall energy usage profile of the site. It also includes carbon emission reporting and analyzing cascaded effects of any IT capacity upgradation on underlying Infrastructure. Because under newer approach the scope is wider and more intelligent thus it is also known as Intelligent Capacity Planning and specialized professionals skilled to apply these practices are referred to as Intelligent Capacity Planners.
The benefits of Green Capacity Planning lie in the collection of power consumption metrics. Facilities management is solely responsible for collection of power consumption data at individual component level and data center as a whole using specialized DCIM (Data Center Infrastructure Management) tool or using power strips. Together with configuration items data (through Configuration Management database CMDB) and resource management data, workload data, performance data (through Performance Management database PMDB and/or monitoring tools) and business demand data (from business); IT Capacity Planner would be able to forecast future capacity requirements in terms of IT infrastructure and energy. Having understood all facets of a data center, it will become possible to perform predictive analysis. For example for a new project Intelligent Capacity Planner will be able to predict how much IT Infrastructure, energy and cooling would be required. Also periodic capacity reports including energy consumption, energy efficiency metrics, and carbon emission will enable IT and facilities teams to monitor their efficiency and feed back the data to external regulatory authorities.
So in this post we have discussed the concept of Green Capacity Planning, background and it’s evolution. That lays the foundation for the various other aspects of the topic. In the next post we will discuss various authorities working towards energy efficient data centers.