Posted by i-cio.com staff | 28 Mar 2011
Internet connection speeds of 100Gbps or more are likely to become a commercial reality by 2015.
Imagine a world where users routinely demand high-definition streaming video to the desktop and the home. Already, many of the world’s Internet providers are offering consumers fiber broadband connections of up to 100Mbps, which can support this kind of activity, with South Korea and Japan benefiting from the fastest speeds.
For business, the opportunities are potentially huge — and not just in terms of content provision. Among other things, a superfast fiber Internet could give organizations much more flexibility to use cloud-based storage and processing services, or tie together remote data centers and private clouds without affecting data transfer speeds.
It could also usher in universal “telepresence”-style videoconferencing, which could drastically cut the need for corporate travel and facilitate mobile and remote working.
But there’s a problem. While much of the coverage has focused on the need to implement the “last mile” fiber connections to homes and businesses, it’s the Internet’s backbone that’s really holding things up. The 10Gbps lines most commonly used to link together trunk lines and major data centers would present a huge bottleneck if the volume of data being carried across them grows as predicted.
Yet replacing them with 10 times faster 100Gbps lines has been a thorny issue, since signals carrying data at speeds of 100+ Gbps over very long distances suffer from what’s known as “nonlinear waveform distortion.” It was thought the vast, complex “nonlinear compensation” circuits required to get around this would not become feasible for mass production until about 2020.
Fortunately, research scientists now look like they’re solving this problem ahead of schedule. Fujitsu, for example, recently announced it had developed a digital signal-processing algorithm that compensates for the distortion, reducing the size of the required circuitry by approximately 70%.
The company says this will make long-haul fiber-optic transmission at 100+ Gbps (at a lower cost than conventional 10Gbps systems) a commercial reality by 2015. In addition, it has developed a high-speed optical switch based on silicon-germanium (rather than traditional silicon) that can operate across a wide range of wavelengths, while requiring less power than ever before. So it seems previously unimaginable speeds may be here sooner than we thought — and it won’t only be faster, but greener.
• Another development, “flow control,” has been demonstrated by MIT researchers. By preventing the need for routers to convert optical signals into electronic ones for processing, this could make the Internet 100-1,000 times faster, while reducing the amount of energy it consumes.
• In July 2010, 29% of broadband lines in Europe ran at 10Mbps or faster, compared to 15% in 2009, according to EU figures.
• The EU expects more than half of European households to have 100Mbps broadband by 2020.
: Mon, 21 May 2012 16:23:23 +0000
: Mon, 21 May 2012 16:19:10 +0000
: Mon, 21 May 2012 16:09:50 +0000