When Vivek Kundra was chief information officer for the District of Columbia, the municipal government planted sensors on its snowplows, enabling live data feeds, over the Web, to citizens to let them know when plows were approaching their streets. "We even let them know whether the shovel on a plow was up or down," he relates. The result was a "lot of good feedback from citizens, which resulted in the department of transportation in DC redeploying its assets," and ultimately, saving money.
Now, as CIO of the United States government, Kundra sees such grassroots involvement as the core of efforts to employ technology to make the US government more agile, responsive, and cheaper to run. Kundra shared his vision for the digitally savvy federal government at The Economist’s Intelligent Infrastructure conference held at Pace University.
As part of this, federal government data is being opened up to enable new discoveries and innovations with the information that was formerly locked away in siloed agency systems. The data.gov site, for example, “started out with 47 data sets when we launched in May 2009. We now have 300,000 data sets online.” Such information includes aircraft schedules from the Federal Aviation Administration and Medicare/Medicaid information. “We’ve just seen an app which couples the FAA data with airline data to show if flights are arriving on time,” he relates. “We want to encourage people to work with the raw data to make discoveries.”
“The American people are becoming the co-creators of the next generation of innovations,” Kundra says. The success stories in the economy, he says, are companies that create platforms, leaving it up to users to fill in the content. “It’s not YouTube creating content; it’s you and I creating the content,” he points out. “Same with Facebook and Twitter.”
On many levels, Kundra wants the US government to operate its technology in much the same fashion as the technology industry. “I just spent some time in California talking with innovative startup companies and venture capitalists, and all are moving toward zero asset footprints,” Kundra says. “Their capital requirements are very low.”
There are important lessons for an organization that spends $80 billion for IT annually, he continues. “The government grew from 430 data centers to 2,094 over the last 10 years. There’s a lot of waste, a lot of duplication.”
For example, he says, the Internal Revenue Service essentially needed to build out a massive infrastructure to handle processing requirements for one day out of the year — tax day. “The rest of the year, those systems are 35% underutilized,” he relates. In another example, the systems handling federal government’s 2009 “Cash for Clunkers” program were overwhelmed when demand spiked toward the program’s deadline. “We anticipated demand on the infrastructure of the agency handling Cash for Clunkers would multiply three times — instead, demand multiplied twenty times, and the system started crashing.” As a result, consumers were left waiting at car dealerships because their vouchers could not be processed.
In contrast, Kundra points out, a site called Animoto, which enables users to build MTV-like videos, handily managed its own huge spike in demand because it was able to dynamically provision server power from the cloud. “They went from 40 to 4,000 servers in a few days, without losing a single user,” he says.
There is some progress, however. “We now have 4,000 systems that are not owned by the government,” Kundra says. “The government wants to move away from active ownership to provisioning to the cloud. Already, he says, there are examples of savings. The General Services Administration, for example, moved its email to Google Apps, and saved $20 million. The United States Department of Agriculture, “with 120,000 employees, moved to the Azure platform, and is seeing $25 million in savings.”
This post was originally published on Smartplanet.com