IT pros and consultants working at energy utilities say smart-meter and smart-grid projects to modernize U.S. energy delivery will be a boon for the data storage industry, as the digitized grid is projected to generate massive amounts of data.
By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
Smart grid is in the news after the passage of the American Recovery and Reinvestment Act of 2009, also known as the economic stimulus bill. The bill set aside close to $5 billion to get digitized energy grid projects under way.
Major consultancies and integrators like Accenture, IBM Business Services and Capgemini have already established energy grid practices and are deep into city-wide projects both abroad and in the United States. But efforts to transform these systems nationwide will go on for the next decade or more. Even the larger integrators acknowledge there will be room for their smaller counterparts.
"I think this is an area that will develop an entire ecosystem with large and small players," said Jeffrey Taft, global architect for smart grid at Accenture. "Some companies are going to be best suited to work with really big players. But in North America, we have 3,000 small utilities, and a lot of those will turn to smaller players to get what they need."
Smart-grid rollouts have two phases. First comes the deployment of Advanced Metering Infrastructures (AMI), or smart meters, followed by the rollout of true smart grids that can absorb and redeploy distributed sources of alternative energy. True smart grids would also balance energy delivery loads through IP networking and data analysis.
The earliest adopters in the energy space so far have gotten to the smart-metering stage. Austin Energy of Austin, Texas, is one of the furthest along in its deployment, with the Pecan Street Project announced last December. Austin Energy CIO Andres Carvallo says Phase 1 of the smart-grid project is expected to be complete in July, when a half-million smart-meter devices will have been rolled out across the company's entire service territory.
Even in this first phase, the data storage requirements are huge.
"Our total information online was 20 TB before we began this project. With the smart grid, we're projecting 100 TB of data a year," Carvallo said. The data won't be kept long -- probably only 20 TB will be retained for compliance purposes for longer than a year -- but the requirements probably won't slow down.
"Right now we're capturing data from the smart meters every 15 minutes," Carvallo said. That requires 200 TB of storage space when disaster recovery redundancy is factored in. "If we move to five minutes, that 200 TB would become 800," he said. "If we moved to one minute, it would be 1.5 PB."
This doesn't factor in Phase 2 of the smart-grid project, in which Austin hopes its customers will be able to plug in local sources of alternative energy to the grid, like solar panels or fuel cells. "We are entering the petabyte age," Carvallo said. "Other hydroelectric utilities will eventually be where I am today."
Meanwhile, California's Pacific Gas and Electric Co. (PG&E) has also been working to roll out smart meters since 2006, and has some 700,000 smart-meter devices deployed so far. When all is said and done, said David Ping, PG&E data center storage team lead, there will probably be about a million devices in the field.
The company began in 2006 with between 600 and 800 TB of storage to support open systems within the company. As it's rolled out smart meters, it's brought in new storage subsystems, adding six IBM DS8300 storage arrays to support the initial rollout -- approximately 1.2 PB of storage. Today, that number has grown to 12 DS8300s, and the company is managing a total of 3.2 PB of storage between open systems and the smart-meter project, and there's talk of boosting that to 4 PB, Ping said.
"The bottom line is that once utilities go to the smart-meter concept, constantly reading and sending information, the storage requirements within the company grow quite a bit," he said. As with Austin Energy, the frequency of meter readings at PG&E could also increase. Right now PG&E's meters are being read twice a day, once every 12 hours, but there's talk of moving to hourly readings to improve the company's forecasts about how much energy it needs to purchase for distribution.
A challenge for the industry
The sheer size of the data set combined with performance requirements to do data warehousing and analytics on it in real time present a challenge to today's storage technology, Ping said. "We have terabyte drives now, but do they allow the performance to support polling meters every five minutes? If so, what does that system look like?" he asked. As an IBM user, Ping says he's learned about the SATA-based parallel-processing XIV system, but discussions about long-term data center design at the company are still at a very early stage.
Not every company will be able to reduce its long-term retention of smart-meter and billing data, according to Capgemini Vice President of North American Utilities Andy Roehr. Capgemini has been consulting with San Diego Gas and Electric on its smart-meter project. "Depending on where they are in the country, some users will have to keep 36 months of reads," Roehr said. "People are using the word 'exabyte' in casual conversation now."
Just storing the data is also only the first step in gleaning benefits from smart meters and smart grids, he added. "The industry as a whole has a challenge in front of it as we learn how to use the data. Right now, we're trying to figure out what the right technologies are and appropriate data collection intervals. People are going to learn as they go."
Given that power and cooling are already concerns in the data center, at what point does the massive data storage and processing required to run a smart grid outweigh the energy savings of the grid project itself? "That's a fascinating question, but I don't know that anyone's actually done the ROI analysis," Roehr said. "The potential savings in conservation [on the energy grid] are fairly huge -- I don't really think the power consumed by data storage will be a significant factor in offsetting the gains, but it's something users will have to think about."
"Existing storage systems at utilities will be stressed and taxed -- they will certainly need to be evaluated and probably upgraded," said Paul Williams, an IBM Business Services Group energy and utilities software solutions architect who has worked with Austin Energy on the Pecan Street Project. "But it's great news for the storage industry -- there will be tons of data that needs to be stored, analyzed and turned into operational information."
SearchNetworkingChannel.com Site Editor Rivka Gewirtz Little contributed to this story.