Slashdot Mirror
UK Research Aims For 100x Speedup In Fiber-Based Broadband
Mark.JUK writes "The UK governments Minister for Science, David Willetts, has awarded £7.2 million to help support the University of Southampton's newly rebuilt Optoelectronics Research Centre (ORC) and the development ('Photonics HyperHighway') of new technologies that would be capable of making broadband internet access over fibre optic cables 100 times faster than today."
What would you like to do with 100 times your own current network speed?
Linux iso's.
So much porn. I'd be downloading about one hundred times more than I do now.
There is no -1 Disagree.
As of today, the fastest fiber NICs run at 100 Gbps (not a typo). I'm pretty sure the big Network Equipment Manufacturers (Cisco, Juniper, Broadcom et al.) are investing more than £7.2M in R&D to develop this too, so I'm kind of mixed. Also, most Carriers and ISP core networks still run on aggregated 10 Gbps, sometimes 40 Gbps, and none (as far as I know) run at 100 Gbps. So the "capacity crunch" they talk about is mostly due to a lack of investment from the companies...
so shouldn't it be fibre? =^)
I would spread the good word that people can increase their endurance with huge savings on enlargment pills!
The technologies available for backbones already are fast enough for the next decade or so. The main problem is the 'last mile'. However once everybody has fibre to their homes, there might be some bottlenecks on the backbone. I estimate this to be the case once everybody has a gigabit connection at home. Today you can, with off the shelf parts, transmit about a terabit per second over a single fibre. A typical exchange would be connected to it's neighbours with hundreds of fibres, but serve only a few thousand households.
However it is important to do basic research. Eventually we are going to need that kind of technology. Just perhaps not within the next decade.
What would you like to do with 100 times your own current network speed?
That would give me >2Gbit/s actual. I could stream what like 40 blurays simultaniously? Don't need it. Can't really imagine anyone who does, really. And I'd probably still be downloading from torrents because the TV/movie execs won't offer it here, no netflix, no hulu, no TV shows or movies on iTunes.
And for most things like series I follow my computer could just download it encrypted the night before in maximum quality, then deliver the key at release time. Bandwidth is really not a problem, at least the pirates seem able to deliver so it's strange if a big company couldn't. Sure I'd still take more if I could but it's no longer a bit deal. Before this is I had 2 Mbit down and that was horrible.
Live today, because you never know what tomorrow brings
Just a few hours ago, /. had this story: http://hardware.slashdot.org/story/11/01/29/2222246/A-Kinect-Princess-Leia-Hologram-In-Realtime. If you follow a few links, you eventually arrive at http://www.media.mit.edu/spi/M2.html, where you will find these bits of information:
The resulting image is horizontal parallax only (HPO), with video resolution in the vertical direction, and holographic resolution in the horizontal direction.
and
The Holovideo Cheops system provides six synchronized frame buffers to drive our 256Kx144 display
I infer that holographic resolution takes 1,000 times the bandwidth of conventional video. So, yeah, I think I can think of ways to use this much bandwidth at home.
Nothing for 6-digit uids?
As of today, the fastest fiber NICs run at 100 Gbps (not a typo). I'm pretty sure the big Network Equipment Manufacturers (Cisco, Juniper, Broadcom et al.) are investing more than £7.2M in R&D to develop this too, so I'm kind of mixed. Also, most Carriers and ISP core networks still run on aggregated 10 Gbps, sometimes 40 Gbps, and none (as far as I know) run at 100 Gbps. So the "capacity crunch" they talk about is mostly due to a lack of investment from the companies...
I think L3 runs 100Gbps. The MLXe has 100Gbps line cards and aggregated 40Gbps is quite common on the backhaul networks. As most people have said, it's not the backhaul that's the issue, it's the last mile. Research into higher speeds over copper twinax is what we really need. I can't see any telco putting fibre to the door unless they are forced to, or it's a new build. Copper is going to be here for a very longtime, so let's get that sorted soon. Most Network analysts expect the traffic loads to explode in 2011/12 with double the amount of traffic in 2009/10, most of it driven by serving rich content (IPTV, streaming video etc.) so these very dense backbone are needed, but industry is already doing research in that dept.
What would you like to do with 100 times your own current network speed?
Shit man, I'd be able to watch videos off YouTube!
In nearly a year and half, my local BT exchange has been congested. "Virtual paths: red". I went from November to January last year at 300Kbits/s on an 8Mbit ADSL line. This month it's been 700Kbits/s. Yet if I wake up at 5am, I have 7.1Mbit/s and can watch two HD streams off iPlayer.
There is slightly more information in the grant overview from EPSRC http://gow.epsrc.ac.uk/ViewGrant.aspx?GrantRef=EP/I01196X/1 although it is quite light of specifics.
The proposal appears to be usual blend of new modulation techniques, all optical switching and the usual "green" nonsense which is required to get anything approved these days.
And the ISPs just want you to be able to hit your download usage limit as fast as possible.
More like:
Photonic
Optimised
Resolution
Networks
or PORN for short...
Laters Sol "Have you found the secrets of the universe? Asked Zebade "I'm sure I left them here somewhere"
There are a fair few unlimited, or at least "unlimited for practical purposes" ISPs available. Sky or Be, for example. I downloaded 200 GB one month, no problems.
The limited ones are generally the ones that use BT's backhaul from the exchange rather than doing their own (LLU), because BT charge a very high per-Mbps rate. Even then, it's enough for gaming.
Raw fiber to the home has enormous implications we are not capable of imagining.
Wiring efforts should accelerate and government regulation should be copied from countries like Sweden, Japan and South Korea to ensure maximum bandwidth and minimum latency, worldwide.
It just makes sense. There cannot be a long-term loss in this investment. Lay fiber everywhere. Construct a fractal grid-net over the planet and get as close as possible to the speed of light. between any 2 given points.
Everybody is in favor of it. What follows will be free sound/video calls and videoconferencing across computers and smart-phones. Inevitably.
What will also follow is distributed computing, as latencies grow lower. As reliability increases, more efficient ways to treat data will emerge, which will greatly increase efficiency. The positive pressure of multi-coring our way forward under the GHz limits will increase the importance of distributed code (but for how long?) so we basically need a very fast, very reliable internet to use our cpu cycles more efficiently.
Probably preaching to the choir here...
Fibre to the home is very rare here in the UK. The incumbent telco (who still own much of the infrastructure, despite no longer being a monopoly) is rolling out FTTC, though their first priority is towns that already have cable.
Widespread FTTH is something I don't see happening for at least 10 years here.
The problem is most home fiber isn't point to point on pairs, its shared with between 32 and 1000+ people on a bi-directional fiber. This is causing problems with the higher speed PON versions since you have to coordinate the talk times of all the end points and you have to leave a quiet time between talk and listen phases so that you don't blind the receivers. A 1kbit packet on a 10 gig fiber takes up about an inch which means if you want to fill up a 10 gig back channel form a number home homes, you have to know how many inches each link is, coordinate the time to about a 1/100 of a nanosecond and then make sure they all talk at just the right time. Our best cheapish consumer clocks are based on GPS and they only know their internal time in the range of about 50 to 90 nanoseconds. 90 ns is about 135k worth of data that can be stepped on. Off the shelf point to point fiber has increase about 20,000 times in 40 years. Shared Passive Optical networks are 40 times faster in the lab today than they were 20 years ago.