This year marks the 30th anniversary of the birth of the Internet, which originally linked universities and laboratories around the United States as an outgrowth of a Department of Defense project begun in 1969. Eventually, the Internet will be as much a part of people’s lives as the radio, television and telephone — all of which will no doubt reside on the Internet in the not-too-distant future.
The Internet’s explosive growth has been fueled by a single factor: open standards consisting of an agreement upon a common set of protocols that enable computers to communicate with one another. Standards also brought Internet mail into existence, which has had an incalculable impact on the way we communicate with each other. And yet another set of standards created the Web, which has allowed everyone with a computer to trade information.
It did not take long for businesses to see the potential of the Internet. Today you’d be hard-pressed to find a company that does not use the Internet for e-mailing sales information to customers, displaying Web site showrooms or integrating sales and supply-chain activities to maintain optimal inventory levels.
As exciting as the Internet’s past 30 years have been, they’re only a shadow of what is to come. We are already catching glimpses of the future in projects sponsored by universities and pharmaceutical research labs.
On the Grid
By downloading a special screen saver, we can allow our personal computers to hook into a worldwide network when we’re not using the machines. This network processes information as part of the search to find cures for cancer, HIV and, most recently, smallpox — processing that would take an individual computer weeks or months but that can be done in days through this voluntary “grid.”
In a somewhat different form, grid computing — a Web-based operation that will allow companies to share computing resources on demand — is where we’re headed.
Simply put, grid computing uses more of a server’s computing power. Today’s computers, like human brains, typically operate at only a small percentage of their capacity; they often sit idle as a processor waits for data. On the grid, the idle time of hundreds — even thousands — of servers can be harnessed by any customer needing a massive infusion of processing power.
Just as the electrical grid provides power to consumers, the computing grid distributes all sorts of computing resources to solve problems. In the process, a grid can give rise to “virtual organizations” — ever-changing groups of individuals and institutions exploiting the resources of the grid for several purposes, in much the same way that individuals in a household exploit electricity for their own needs.
Logical Outcome
Grid computing is the logical — and desirable — outcome of several factors. First, the idle time inherent in computing means that by working on a grid, computers can share operations according to their abilities and unused capacities. Information thus will be processed with maximum efficiency.
Second, the advent of broadband enables networked computers to share data constantly at high speeds — critical to the functioning of any sophisticated network.
Of greatest importance, though, is the collaborative culture that the Internet gave us in the first place. Without common protocols, the Web would be a Tower of Babel. Businesses, and the companies that supply technology to business, are understanding that the virtues of open standards — flexibility, scalability and adaptability — are keys to the success of information technology systems and to the ability of those systems to contribute to an attractive bottom line.
Businesses thus are developing IT systems that embrace open standards, and are constructing those systems around middleware — software that enables different technologies to interact with one another smoothly and transparently.
Platform Internet
Capacity, speed and open systems are setting the stage for the leap to grid computing — for the Internet itself to become a computing platform. The resources we desire will be at our fingertips, but they will reside on the Internet rather than in single computers or on local servers.
We will engage these resources through secure channels, as we do now when we transact business on the Internet. But we will never again wish that we had more storage capacity, more data-manipulation capabilities or another suite of applications. The Internet’s virtual computer will be distributed worldwide rather than existing in one place.
It will provide us with everything we need, whenever we need it, and will enable us to coordinate our work with anyone, anywhere, without technology conflicts. Like powergrids, computer grids will become an essential element of our infrastructure, operating invisibly except in the rare cases when a problem arises.
Tech Community First
Like Linux and other new technologies, grid computing is emerging first in the scientific and technical communities, enabling scientists and engineers to collaborate in applications across institutions and around the world in disciplines such as high-energy physics, life sciences and engineering design.
For example, the Mayo Clinic is developing a system for linking its own medical database with vast external public and private data sources to develop more-effective patient treatments. The University of Pennsylvania is building a powerful computing grid that will make remote breast cancer diagnosis and screening a reality, putting the benefits ofsophisticated, high-tech healthcare delivery at the disposal of entire populations. And Indiana and Purdue universities have linked their systems and developed an application that helps assess the impact of catastrophic events on populations.
The next step is for grid computing to break into the mainline business world. Butterfly.net uses a grid for its massive, multiplayer online games. And grids are beginning to see use by financial institutions for applications such as risk-management analysis and portfolio optimization, by manufacturing companies for engineering simulation, and in the energyindustry for seismic analysis.
Bright Future
Of course, Utopia has yet to arrive, but these early trials hold the promise of the future. New standards, beyond interoperability, must emerge among different computers and networks. And if we are to make computing grids an essential, yet invisible, part of our society, we must focus on simplicity of design and create a self-healing, autonomic infrastructure.
This is an architectural style popularized by the Blue Gene supercomputer, which is being developed to give scientists and doctors greater insight into combating diseases. Blue Gene has eight million autonomous computing elements that cooperate among each other and self-manage the computations.
The promise of grid computing — of Web-based services providing universal access to information and computing in a collaborative environment — is as real as it is seductive. In fact, it’s only fitting that the Internet, born out of a vision of collaboration, should now come full circle to fulfill that vision.
Willy Chui is the vice president of the High-Volume Web Sites organization in the IBM Software Group. He and his team of experts have created a center on the scalability of high-volume Web sites. The center serves as a focal point for customers to address the challenges of excessive growth of Web site traffic and increasing complexity of e-business infrastructures.
Yes, Of course thanks for such a wonderful explanation on Grid Computing. that was a great article. By the way have you heard about Cloudslam 2010 conference which is the world’s largest conference on cloud Computing and its technologies. I AM very much interested to gather more information about the Cloud computing and its technologies. the complete details about the conference available through http://cloudslam10.com
Indeed seductive Mr. Chui. Thank you for this article. Here’s a comment and an idea:
Whenever technology demonstrates collaboration on such a large scale, it is this collaboration that is as exciting as the technology itself. The fact that the Internet is becoming THE computing platform for planet earth suggests increasingly hopeful and effective levels of collaboration and cooperation across business, political, social and economic structures.
But here’s an idea: how about this global computing platform using its underutilized resources like a giant virtual robot? In addition to undertaking massive computational tasks as you’ve mentioned in your article, consider a technology that can turn virtually any computer network into a grid, the purpose of which is to emulate physical human behavior.
As an example, imagine the actions of 5,000 people from all over the world simultaneously visiting the same eCommerce site (that may be even dynamically provisioned with the latest grid computing software), all performing a mix of different transactions: looking up information, requesting email support or ordering widgets. Or how about a "crisis" where many thousands of people rush to a site to view or order the latest lingerie during half time?
This type of collective human behavior can wreak havoc on most Web sites, especially if it is sudden or unplanned. But if a technology existed that used the spare computing resources of consumers’ Internet-connected computers (and paid them for doing so), in aggregate, to emulate this type of behavior, then we would have a giant virtual robot, capable of emulating human actions on a large scale. But as important, this "robot" will provide visibility all the way to the users’ desktop. Interestingly, this "robot" would be using the Internet as a platform, for testing the Internet as a platform, to carry out these actions against a target Web site…on the platform.
Combined with other human emulations, like the autonomic nervous system, it could readily self-manage Web infrastructure so that the "World Wide Wait" can finally become a comical historical footnote.
This technology does exist; this is what we do at Distributed Computing, Inc. (www.capcal.com), providing a "next step" for grid computing to move into the business mainstream. We continue to be excited with more and more applications being developed around this most interesting computing model.
Sincerely,
Joseph Giove