Collaborative Internet Applications
Frustrated by Slowness
If you've ever found your fingers drumming on your desk, frustrated with the slowness of a database in churning out your results, you'll appreciate the work of Jonghun Park, assistant professor at Penn State's School of Information Sciences and Technology (IST) who has been working to speed things up. The Penn State researcher has found a faster, more efficient method for sharing internet resources of wide distribution, such as those of high performance computers, various services available online, and databases. Park's protocol revs up the allocation of internet resources times ten.
"In the near future, the demand for collaborative Internet applications will grow," says Prof. Park. "Better coordination will be required to meet that demand, and this protocol provides that."
The results of Park's work were described in a paper called, "A Scalable Protocol for Deadlock and Livelock Free Co-Allocation of Resources in Internet Computing," and presented at IEEE's Symposium on Applications and the Internet in Orlando, Fla. Park's algorithm enhances the collaboration of internet applications so they have the flexibility to support large-scale computing. The proposed protocol employs parallel methods as opposed to the serial methods in current use for processing user requests. Not only does this help distribute resources in a more efficient manner, it also prevents both deadlock and livelock which may occur when making multiple requests by internet programs that run at the same time and that compete for available internet resources.
The algorithm also makes it possible for internet applications to pick and choose from the resources at their disposal. Until this time, technology did not support such choice and that imposed limits on the user. Park's protocol is also independent and decentralized so that it can function on its own, allowing for collaboration with many different organizations throughout the world wide net. Until now, the protocols in existence depended on communication between applications, slowing things down to a crawl.
The integration of information and computing resources into one network should broaden the exchange between more people making the internet more user-friendly for governments, businesses, and the military, for instance. One such collaborative effort is known as Grid Computing, which is geared to the performance of large tasks and works much like electrical grids. Grid Computing takes internet resources and harnesses them to feed super-computing, large-scale intensive computational efforts, such as for scientific pursuits like seismic analysis and economic forecasting.