Researchers at Aalborg University, MIT and Caltech are working on software and mathematics that could make online communications via computer, mobile phones or satellites much faster and more secure than they are today.
This particular study uses a four-minute mobile video as an example: the method used by the Danish and American researchers resulted in its download being five times faster than through current state-of-the-art technology, while at the same time its streaming was uninterrupted (whereas through the "conventional" method it got stuck 13 times).
This technique could bring dramatic changes to the market and the Internet in general. "In experiments with our network coding on Internet traffic, equipment manufacturers have seen speeds that were five to ten times faster than usual. And this technology can be used in satellite communications, mobile communications and regular Internet computer communications," says Frank Fitschek, professor at the Department of Electronic Systems at Aalborg University.
Internet communication formats data into “packets.” Error checks ensure that the signal arrives in the correct form, but they often also mean that some “packets” have to be resent, which slows down a network. Danish and American researchers are solving this problem with a special type of network coding that uses innovative mathematics to store and send the signal in a different way. The advantage is that errors do not require a “packet” to be resent - instead, upstream and downstream data are used to reconstruct the lost data, using a mathematical equation.
With conventional systems, it is possible to send a “packet” 1, a “packet” 2, and so on. In this case, this is replaced by a mathematical equation. As Fitzek explains, a good comparison would be with cars on a road: “we don’t need red lights. We send cars into the intersection from all directions, without having to stop for each other, which results in a faster flow of traffic.
Network coding is suitable for use in the areas of the Internet of Things, 5G systems, software defined networks, content centric networks and possibly in distributed storage solutions. However, for it to work, data needs to be encoded and decoded using special, patented technology. Fitchek and two of his former students from Aalborg University (Janus Heide and Morten Videbeck Pedersen), together with their colleagues from the US, founded the software company Steinwurf, which makes the RLNC (Random Linear Network Coding) technology available to hardware manufacturers - while, as noted in a relevant announcement from the university, secret negotiations are underway to eventually bring this technology to consumers. As part of this effort, the company has opened offices in Silicon Valley (however, its headquarters are still in Aalborg).
Source: naftemporiki.gr
