Textbooks, encyclopedias and newspapers are educational resources of the past. Today, students seek answers from the sky -- specifically, the clouds.
A new way of using the Internet is making data limitless. It's called "cloud computing," and it's allowing educational institutions, businesses and individuals to keep more information at their fingertips than is possible with even a room full of computers.
Companies such as Google are delivering Web applications via cloud -- the Internet -- from an estimated one million servers, allowing users to run these programs on the Web cost-effectively instead of buying and installing proprietary software. By using Google's search function, creating a Gmail account or using the Google Docs application, teachers and students are participating in cloud computing.
Another function of the cloud is data storage. Research institutions store lab data, student information and course applications on their servers, adding annual financial costs to universities to increase storage space. By keeping this information in the cloud, these institutions are saving money. For example, an information technology research and advisory company estimates it costs $10 to $100 for each e-mail account monthly, plus $10 per gigabyte of storage. By switching to the premier edition of Gmail -- and leaving e-mail in the clouds -- each account costs $50 per year.
Christophe Bisciglia, a senior software engineer at Google, said cloud computing is analogous to the evolution of banking. At first, people were hesitant to give money to banks, but later realized that these institutions provide a more beneficial service than keeping their money themselves. In the same way, cloud computing is revolutionizing the way students store -- and access -- data by offering its machinery to consolidate billions of bits of information.
"What if you had 1,000 times more data?"
Cloud computing is cutting-edge material for computer engineers because it introduces a scale that is almost unimaginable. While interviewing Google job applicants, Bisciglia realized that the engineers could come up with complex algorithms -- procedures for solving problems -- but they were consistently stumped by one question.
"I would say, 'What if you had 1,000 times more data?'" he said. "And the students would go blank."
Google created a revolution in computer science that the students were not familiar with, inspiring Bisciglia to spearhead a program at the University of Washington (UW) -- his alma mater -- teaching students to write software on many machines in a cloud environment.
Ed Lazowska, a computer science and engineering professor at UW, helped Bisciglia pioneer "Google 101" or "Problem Solving on Large-Scale Clusters." Students read Google research papers, perform programming exercises and present projects demonstrating how to analyze massive amounts of data.
"It teaches students to analyze amounts of data that have never occurred to them before," Lazowska said. "You don't do that with a single machine because the amount of time it would take is years. Rather, you do it with hundreds of thousands of computers in parallel."
Last year, Google partnered with IBM to donate hardware, software and services to ebb financial barriers to educating engineering students about this change to computing. Google and IBM dedicated a cluster of several hundred computers from a data center to the academic community.
"We have no idea where this data center is located -- we're just using it," Lazowska said. "The machine in our building is dozens of processors. The machine at our data center is thousands."
Bisciglia said the same program is being introduced at other U.S. universities and internationally.
"This is not just a niche," he said. "This is a fundamentally new computer science. Engineers graduating today need to understand how to write software that runs on many machines in a cloud environment."
Moving to the clouds
As more research institutions and individuals take advantage of the cloud, businesses are beginning to see the cost-effectiveness as well. Bisciglia said that cloud computing is almost universally adopted by small businesses.
"Why would you set up e-mail servers when Google will do this for you for free?" he said.
However, Lazowska said that this adoption isn't always straightforward. For example, UW medical centers and neighborhood clinics can't store patient information in remote data centers because of privacy regulations. These facilities uphold the rules within the Health Insurance Probability and Accountability Act, which includes electronic security standards.
Cloud computing brings about concerns for security such as data integrity, data recovery and privacy, just to name a few. And just as banks in their infancy couldn't give customers a 100 percent guarantee, cloud computing is still working out the logistics.
The official Google blog states: "It's much like securing your house. You put your most private information in a safe. You secure the safe in your house, which is protected with locks and possibly an alarm system. And then you have the neighborhood watch program monitoring your neighborhood. It's very similar at Google. Our most sensitive information is difficult to find or access (the safe). Our network and facilities (the house) are protected in both high- and low-tech ways: encryption, alarms and other technology and strong physical security at our facilities. And finally, we've learned that when security is done right, it's done best as a community (the neighborhood); we encourage everyone to help us identify potential problems and solutions."
Although cloud computing cannot guarantee security, people also run a high risk by storing data on their personal computer. Since people don't always back up their data, a problem can become detrimental -- all information could be lost. When companies outsource this material, they give data to organizations that are committed to data security.
Not only are they keeping data secure, but they also allow users to make use of this information faster. Lazowska said cloud computing is using "a thousand computers for one hour" instead of "one computer at a thousand hours."
The technology has barely hit the surface of its capabilities. Still in its infancy, its future may include more than data storage and application services.
"Dealing with enormous volumes of data is what the future is going to be about," Lazowska said. "The challenge is going to be to figure out how much of what we think can't be done can be done."
*This story is from Converge magazine's Mixed & Mashed 2008 special issue.