As “Green IT” continues to hold centre stage, the data centre, with its escalating demand for power, cooling, and space is increasingly seen as one of the major culprits generating an estimated 170 tonnes of carbon dioxide emissions, or 0.3% of the global total.
As corporations continue to grapple with escalating storage requirements, Disaster Recovery protocols and globalization, it is set to only get worse as a recent McKinsey report has predicted. According to the study, data centre emissions will overtake those of the world’s airlines by 2020. Not to mention that data centers now account for 25% of the IT budget with energy bills growing at 16% a year – so as someone famous undoubtedly said, “something’s got to give!”
What is largely unrecognized are the options available today to manage this, starting with a basic understanding of the influence and impact that IT purchasing decisions have on data center power, cooling, and space consumption from the very outset. As a result, very few companies have formally incorporated facilities considerations into their vendor selection processes and this premise was supported through a recent study conducted by Oliver Wyman of the data centre industry.
Respondents to the survey identified three primary pain points: data center space, power capacity, and heating/cooling requirements.
First, because of the rapid increases in computing and storage requirements, many data centers are running out of space, both in terms of square feet and rack units. To delay expanding their data centers as long as possible in order to save costs, companies are deploying denser drives and more highly utilized systems, which in turn leads to higher power consumption per square foot. As a result many companies are finding power capacity to be their primary concern while others are struggling with how to manage heat dissipation in order to keep the data centre cool.
Storage is another major contributor to the problems data center managers are experiencing and there are real and measurable differences across storage vendor platforms on the critical power, cooling and space metrics. According to the data center managers interviewed for the study, the primary drivers of power, cooling, and space in enterprise storage systems have little to do with disk drive technology differences and everything to do with product feature differences provided by storage vendors.
Data Center Constraints
A significant number of companies specified data center space as their primary pain point. The study participants identified that storage requirements are growing at accelerated rates within their data centers with some experiencing 50% or higher data growth year over year. To accommodate this demand for storage space, the majority of companies are replacing older, bulkier equipment with smaller, denser equipment.
As one study participant said, “We go through a 20% tech refresh every year to ensure that we are always using the most efficient storage systems out there.”
Some data centers have also been able to save tremendous amounts of floor space by utilizing virtualization technology. One large data center which converted a significant portion of its physical server environment to virtual, managed to reduce its server requirements for those applications from 250 servers to just 18. Leveraging on storage virtualization with technologies such as thin provisioning and RAID, having fewer and larger storage systems will provide more capacity and better utilization, also resulting in less space, power and cooling.
For the companies that have alleviated space issues, power capacity becomes top of mind as denser systems free up floor space but require more power and give off more heat per unit of space. Customers who had reached the maximum power capacity of their data centers discussed the options of either upgrading the Uninterruptible Power Supply (UPS) to increase the power capacity or expanding to a new data center. Upgrading a UPS can cost anywhere from US ,000 to US ,000, whereas expanding into a new data center can cost several millions of dollars and take multiple years to deploy.
To alleviate this power capacity stress, facility managers are using clever, but simple strategies to reduce the power needed for cooling equipment. Instead of the traditional methods of cooling, with raised flooring and forcing cold air up through the racks, new data centre cooling options drop a curtain of cold air down the front of the machines. This is then drawn into the machines and exits from the back as hot air, which naturally rises to the ceiling, where it is vented to the outside.
To prevent one machine’s exhaust from heating another machine, the layout places the racks front to front and back to back in an arrangement, called “hot aisle/cold aisle”.
Today, this simple application of a basic law of thermodynamics has become a best practice for data centre design because of its energy efficiency.
Financial Impact
The increase in power consumption is carving a bigger chunk out of IT budgets and therefore attracting lots of attention. The Oliver Wyman study found that this driver was more painful for data centers located in densely populated urban areas where utility costs can be up to three times higher than costs in remote locations. Moving to lower cost environments also offers advantages in terms of support staffing requirements and data security in the post-9/11 environment. Many organizations will be looking towards this solution in the longer term and for some, colder geographies also offer an alternative option.
Regulatory and Reputation Considerations
Rapid data growth has also caught the attention of public interest groups and government organizations and in some cases this attention has turned into regulatory action. One customer mentioned a levy that the UK government adds to the data center utility bill for every MW of electricity consumed. Other customers in the U.S. and different European countries have suggested that they also face the pressure of being “good citizens” and note regulatory efforts as being a key factor. In the near future, we will see regulatory action in Asia and this will lead the push towards green IT.
Although different types of companies have been afflicted by these pain points to varying degrees, the simple solution is to reduce storage requirements. Higher storage efficiencies allow companies to extend data center longevity, alleviate power and cooling issues, and reduce operational expenses (i.e., the utility bill). Storage and facilities managers have an opportunity to address these data center constraints and the resulting financial impact by choosing the storage vendor that provides the lowest number of raw TB for the usable capacity required. With each vendor offering varying levels of storage efficiency, customers who are able to reduce the amount of raw storage needed for each effective or “usable” TB will see direct savings in space and power consumption.
The author is Managing Director of NetApp, ASEAN |
