• PAGE ONE
• INDEX
• HOME

Peering into the future

N Nagaraj

SO, what is this peer-to-peer computing (P2P to intimates) craze all about? If you did not know already, and you are an avid surfer, you have probably been exposed to it without knowing about it. If you have ever used Napster to download your favourite m usic... well, you have used P2P.

Now, on to what it really is all about. As the name suggests, it connects two machines to one another, and at some level of interpretation, both the machines are equal. We are not talking in terms of processing power here, but in terms of hierarchy. This means that both machines are client and server at the same time.

For instance, today, when you visit a database heavy site, let us say a shopping mall, and search for information, all that your machine (the ``client'') has to do is to send the request, and at a later stage, display the result. The host computer (the ` `server'') actually does all the heavy work like sifting through its databases. This concept of client and server is under assault now.

How does P2P work? P2P is loosely based on a very old concept of just connecting two computers through a modem/dial-up line combination, just that instead of connecting the two computers on a dedicated line, it is now possible to connect two specific com puters over the Internet. The advantage in connecting this way is that you can use the universally accepted IP (which is easier to use) rather than mess around with different protocols and connections.

So what happens to the huge Web servers that already exist? Simple. They act as a ``switch'' (like in a telephone exchange) and forward the call to the right number.

Is it going to change the way people compute, surf and shop? Yes and no. It is not going to change the way people compute, but it is going to change the way computers compute and connect to each other. In all probability, P2P will assure a better quality of service on certain services. New business opportunities outside of increasing efficiency are going to be more in the creative and arts fields rather than any other.

What are the applications? The base application, of course, is that of global data sharing. You can connect your computer to any other computer in the world and share documents, spreadsheets, music and maybe even the schedule. The application, that is th e ultimate ``killer-app'' for the next one year is the one that will enable computers to share not just data, but processing power as well.

This is not new. The technology to share processing power (distributed computing) already exists, and has already been used in two famous projects successfully: The Search for Extra terrestrial Intelligence programme (SETI) and the Great Internet Mersenn e Prime Number Search (GIMPS). The problem is to make an application that can be used in corporate environments and business, and whether it will be successful.

For instance, in a corporate environment, large batch processing of database records can take place across the intranet during the night when the computers are free, or if less critical, can take place during the weekends. What happens is that the whole task is split into distinctly identifiable tasks that are ``bite-sized'' and each computer on the network will pick up a task, complete it, deliver it, and ask for the next task. This method also allows flexibility: The computer will not pick up the next task till the previous one is completed and delivered, and therefore, the user of the workstation can use it whenever he or she wants it.

What is even more of a challenge is the use of this technology across the Internet. Some start-ups are already offering services to companies to harness the idle computer power lying around the Internet. One such company is New York-based Data Synapse th at is targeting financial services companies as these require a lot of computing power that would otherwise be available only from mainframe, and therefore, legacy systems. Other companies planning to capture a part of this space include Entropia, United Devices, and Popular Power.

As of now, users who participate in the P2P arrangement are rewarded on a sweepstakes basis rather than paid for the work their computers do. Naturally, it is very likely that further down the road, participants will want their share of a pie of their ma king. No business models seem to have emerged yet, and the only successful demonstrations have been socially motivated projects.

Other questions remain to be answered: Can all corporate data processing be divided into itsy-bitsy pieces? Who is going to bear the infrastructure costs -- the service providers or the client companies? How secure is the data that is distributed for pro cessing? How will service providers handle viruses and worms? Is it possible that the service provider may fail due to ``denial of service'' attacks?

While these questions may remain unanswered for some time, it is also obvious that companies that see potential will work on removing the problems or at worst, work around some of these problems. Finally, it looks as though Gates' and Jobs' visions of in dividual workstations ruling cyber-space, and Larry Ellison's vision of the Network Computer has fused into something even bigger... The network is the computer. QED.