Let's move the feds into the cloud
By David Kralik
It's all too common for people to criticize government inefficiency, but rare to hear suggestions that the solution is for government to get its head into the clouds.
I'm talking about cloud computing, a transformational technology being embraced by the private sector because of its promise for enormous productivity gains and reduced costs.
Cloud computing has three basic characteristics: capabilities are accessed over the Internet, housed in an off-site data center, and paid for on a subscription basis. This new model delivers computing applications as a utility, similar to electricity or telephone service. Many applications including e-mail, office document productivity, data storage, and customer databases are moving in this direction because of the opportunity to eliminate the need to buy, maintain, or upgrade information technology systems. But sadly, outdated bureaucratic rules and regulations prevent the federal government from fully being able to embrace and reap the benefits of this technology.
The first benefit is cost. Three separate independent studies conducted in 2004 (Gartner Group, the Yankee Group, and Morgan Stanley Research), all suggested that the cost of cloud computing over three to five years is almost half the cost of similar non-cloud solutions. The federal government spent $64.4 billion on information technology in FY 2008, much of which could be reduced had it adopted a model that replaces significant capital expenses for hardware and upfront license fees for a system of renewable per-user subscriptions. This also brings more predictability and stability in future cost outlays and allows for scaling on demand.
The second benefit is speed. Again, because cloud applications are delivered via the Internet, deployment can be done instantly and simultaneously to thousands of users in different locations around the world. Cloud applications are also regularly updated, which can alleviate the constant challenge that government institutions face in being behind on the latest security or upgrade.
A final benefit is ease of use. If you use applications developed by Amazon, Google, or Salesforce.com you are already using applications delivered over the cloud. These applications are easy and intuitive and hold a lot promise for streamlining government if its services operated on similar efficiency. If the government of the District of Columbia can find a way to incorporate things like Google Apps into their operations, shouldn't the federal government be able to do the same?
The lack of those three benefits is what often results in many classic failed government IT projects like the FBI's 2003 decision to terminate its Virtual Case File (VCF) management system. After spending $170 million--a number itself way over budget--the FBI still doesn't have a system to track criminal activity that could prevent terrorist attacks at home, although a replacement is expected this year at a cost of over $425 million.
The VCF failure was more than just a failure of contractors; it was a failure of the type of technology (proprietary, software-based systems) that is now past its prime. In addition to helping solve homeland security challenges, cloud computing could also improve congressional constituent management systems and help reduce voter registration fraud.
With enormous benefits like these, one has to wonder: why isn't the government fully embracing cloud computing now? Two key reasons can be suggested.
Congress spends millions to support a proprietary in-house data infrastructure system at the Ford House Office Building. They do this on the theory that so long as the data is physically housed at a certain location, it can be protected from search and seizure. The same theory holds for data protection from congressional subpoenas when an administration evokes executive privilege. But if Congress can update age-old rules like the 1775 Franking Privilege for a specific technology (YouTube, as they did in October of 2008) why can't it pass a law to allow for greater business efficiency while ensuring that government data remains protected?
The second concern is security. The theory here is that because one cannot physically see where data are stored remotely and the fact that the applications are accessed over the Internet, they must be insecure. But economies of scale are allowing for more sophisticated, state-of-the-art security, disaster recovery, and service reliability features than any individual institution can deploy on its own. Specifically on disaster recovery, there is significant risk in the event of natural disaster of data being housed in one location.
But cloud computing allows for safeguards so that only authorized users can access remotely stored data. This could have prevented the serious breach of privacy such as in October of 2008 when the U.S. Department of Veterans Affairs reported that personal data on over 26 million veterans was compromised from a stolen laptop.
The move from mainframe computing to microprocessor was a major transformational change in information technology, as was the transition from punch cards to software and the invention of the Internet. A fourth major transformational shift is occurring right now as software is replaced with "software as a service," which can significantly improve government operations, lower cost, and move government into the 21st century. As we begin a new session of Congress and new presidential administration, it's time to give this technology serious consideration.
As the single largest purchaser of information technology, government could benefit from this if only it would get its head out of the sand and into the clouds.
David Kralik is director of Internet strategy for American Solutions for Winning the Future, an advocacy organization founded by former Speaker of the House Newt Gingrich.