Smarter Data Handling Could Make Net Faster

Researchers at the Massachusetts Institute of Technology (MIT) say they've developed a system that could help speed up the Internet.

The research is designed to find an alternative to the Transmission Control Protocol (TCP) system. TCP is the set of rules that govern the way data moves around the Internet.

Whenever data travels online, it's broken down into small chunks known as 'packets' (in effect, info packets!). Often, these packets take different routes around the various connections that make up the net, then get reassembled at their destination.

Net Currently Favors Accuracy Over Speed

The big limitation of TCP is that it's a system which is designed to put more emphasis on making sure data gets to the destination accurately and completely. There's less emphasis on making sure that data gets there quickly.

Another problem is that the TCP system sometimes struggles to efficiently deal with data flow fluctuations.

The MIT researchers believe they've found a way to improve TCP. It means figuring out new ways of routing data -- a task that's so complex it's difficult for most humans to wrap their heads around.

The result: MIT researchers have developed Remy, an automated system that works out more efficient sets of rules for the flow of data across the Internet. (Source: infoworld.com)

New Computer System Improves Internet, Cellphone Networks

In early simulation testing, Remy was able to not only pass information through three times as quickly as a network using TCP, but it was able to handle twice as much data. Remy also appeared to work on cellphone networks, though with a less drastic speed improvement. (Source: popsci.com)

The big question is this: how will Remy's algorithms work on the real Internet?

Researchers say that's hard to predict without more extensive testing. So, it would seem we won't be using Remy or systems like it for some time.

But even if Remy is never used on the Internet, it could play an important role in helping speed up local networks. For example, companies that have huge data servers -- such as Google -- might be able to use it to improve performance on their internal systems.

And if Google's data moves faster, it's possible your everyday Google users will experience a boost in performance when carrying out searches.