It's no secret that, with storage growth exploding and data centers still needing additional hardware to meet business needs, data centers are steadily consuming more power and have ever-high demands for cooling.
At the same time, however, data centers are running out of power even as governmental environmental regulations are increasing and cost-per-watt is rising -- with incremental power (when available) being even more expensive. Consequently, a growing number of companies are finding that by using software to better manage data storage and growth, they can reduce their hardware power consumption and cooling needs.
Powering Up
The power and cooling challenges data centers face are very real. In a November 2006 Gartner study, 50 percent of data centers reported that they will have insufficient power and cooling capacity by 2008.
Worse yet, power is not cheap -- nor is it getting any cheaper. For example, cost-per-watt in the United Kingdom went up 35 percent just from 2005 to 2006. During the same period, the European Union dealt with a 15 percent cost-per-watt increase.
What's more, data centers use a lot of power. In the U.S., server power consumption alone totaled 45 billion kilowatt hours in 2005 -- that's more than Mississippi and 19 other states, according to a Lawrence Berkeley National Laboratory study. The world's server power consumption for 2005 totaled 123 billion kilowatt hours.
With data centers demanding more and more power, costs can soon skyrocket. In fact, a recent BusinessWeek article found that large organizations spent between four and eight percent, and sometimes as much as 10 percent, of their IT budgets on energy. Gartner predicted that this will rise by up to four times within the next five years.
Further, a 2006 data center survey by power supply vendor Liebert Corp. reported that 33 percent of respondents actually expect to be out of power and cooling capacity by the end of this year, and 96 percent expect to be out of capacity by 2011.
Of course, governmental regulations put additional pressure on data centers to become more energy efficient. Regulations such as Japan's Energy Law and the European Union's Directive for Energy-using Products (EuP) have companies around the globe looking for effective strategies for addressing these challenges. So do developments in the United States, which is seeing increased scrutiny by the Environmental Protection Agency and is dealing with the U.S. Supreme Court's ruling that carbon dioxide and other greenhouse gases are now considered air pollutants under the Clean Air Act.
To address these issues, companies are utilizing low power servers and standby power management software, and as a result they're improving power supply efficiency. Many are also leveraging virtualization to address a small portion of the problem. Indeed, using more energy-efficient servers is an effective approach to energy conservation.
Yet, as enterprises buy these servers, they are faced with another challenge: migrating enterprise data agilely to newer devices and subsystems while maintaining service level agreements (SLAs).
And that's where the good news begins. By using software that enables organizations to improve storage utilization, deploying storage tiering, and using hardware more efficiently, enterprises can create energy efficiencies in the data center and meet established response time levels.
Provisioning Storage
IT organizations support their enterprises according to pre-determined SLAs. These SLAs may specify how accessible specific data must be, how long it must be retained, and how quickly it must be made available during peak periods. Data that experiences high read, write, and update activity -- sometimes referred to as transaction data -- often requires high performance storage. Data that does not experience high activity -- sometimes referred to as non-transaction data -- typically can reside on lower performance and more power efficient storage.
Yet, in practice, the majority of data is non-transaction data stored on expensive, high-performance, power-consuming disks that have relatively modest capacities. By identifying and then migrating non-transaction data to the appropriate lower performance disks, organizations can save power and energy. The lower consumed watts per gigabyte ratio of these high capacity, lower performance disks and disk systems can save substantial energy use.
Today's dynamic storing tiering solutions provide continuous migration support for file system data using automated methods that respond to changing conditions. Using such a solution in combination with a scheduled data migration service for both file system data and data contained within database structures gives organizations effective enterprise data migration capabilities between tiers. Better yet, analysis tools can monitor file system activity and determine which files are active and inactive and, therefore, are eligible for archiving or migration to more energy efficient storage devices.
Businesses can also reduce power consumption by identifying duplicate data. Enterprises often retain too many data copies. In fact, a typical organization may have between 10 and 30 copies of each production data byte. Needless to say, many of these copies are no longer need, are misplaced, or even worse, the enterprise doesn't even know they still exist. Clearly, reducing the number of data copies reduces storage capacity requirements and storage power consumption. Once reduced, snapshots and other copies from high performance disks can be moved to lower performance disks.
Admittedly, because of continuing data growth, any existing data storage subsystem will eventually exceed available capacity unless more capacity becomes available, and the data center may not have enough space or electric power to support expanded storage capacity. At this point, many enterprises begin to consider building a new data center that can easily cost tens or hundreds of millions of dollars over time.
A more economical solution may be to leverage MAID (massive array of inactive disks) technology, which consumes a small fraction of the electric power used by lower performance storage systems that house non-transactional data. MAID storage subsystems power-off idle disks and power them back on when an application needs access to dormant data. This gives enterprises a third energy-efficient data storage tier for migrating data.
While hardware power and cooling requirements are the source of many data center energy problems, software can help remediate them. In fact, some vendors contend that their software can help customers reduce data storage energy consumption by as much as 50 percent and total data center energy consumption by as much as 25 percent.
Reducing electrical power consumption will likely remain an enterprise imperative. By improving storage utilization, deploying storage tiering, and using hardware more efficiently, organizations can significantly reduce their data center power consumption and cooling requirements.
Bruce Naegel is senior product manager for Symantec Corporation, and is responsible for Symantec's SIGSEC software infrastructure. He has over 20 years of storage product management experience covering both hardware and software.
