Slashdot Mirror
Intel Boosts Optical Communication Speeds
An anonymous reader writes "Intel has developed a device, the Avalanche Photodetector, that senses light pulses and amplifies output signals for faster data transfer over long distances. Researchers claim this is a big advancement in the field of silicon photonics, in which silicon is used to transfer light pulses for data exchange between chips and devices. APD can detect light at higher frequencies and moves data at rates of 40Gbps, making it more sensitive and quicker than earlier photodetectors, at a tenth the cost."
"Intel has developed a device, the Avalanche Photodetector" is a bit of an overstatement, since these have been in use as nuclear and particle physics detectors for almost 30 years now...
I'd say the real news here is "tenth of the price". The 4 times the previous speed improvement is nice, mind you, but it's the price drop that means it's likely to find it's way to consumer-level stuff eventually.
Also interesting is DARPA being involved (again) which seems to happen more often than our patenting overlords would like us to know.
There's another article about it here. Interesting snippet: "Earlier this year, Sun received a US$44 million contract from DARPA to boost computing performance by enabling chip communication using lasers over silicon and to reduce power consumption by placing chips close to each other."
yawn, it's not like it was too slow before
how about spending time and money on more practicle things
like actually start running their own ISP connectig the unused dark fibers so comcast stops ripping us off with their "unlimited" capped bs.
This is the craziest period in history ever. Everything technological is getting way faster and way cheaper ALL the time. I look forward to the day where I'll have a free, unthrottled and provided-by-the-city 1 terabit internet connection. Must...avoid...typical 2008 cloud computing joke......ahhhh
First, APD's have been around for years; their vacuum cousins are called photo-multipliers.
Second, the article states electron pairs; that should be amended to electron-hole pairs. When a photon whacks into a lattice, it ejects an electron, leaving a hole. This absence of an electron can be filled by a nearby electron, so the hole moves, rather like a hole in a crowd.
The important parts (pointed out by other posters, also), are that:
In all, life just got more interesting; we just have to wait about five years for this to be in regular production
This is progress?
I would like to take the time to point out that avalanche photodetectors are exceptionally old news, but most people (including subby) don't seem to realize this. The original incarnation was called a photomultiplier tube (yes, vacuum tube technology). For decades we've used these things in industry and primary research. For example, a Geiger counter is basically a photomultiplier hooked up to a voltage source and a speaker.
That said, I'm very curious to know the exact academic/industry/government research relationship which most likely gave rise to Intel's advances. It'd be a good way to wake people up to the fact that truly risky innovation simply does not happen in private industry. There's no reason to take on that risk to stable profits.
Sure, some will argue Intel needs to innovate or die, but I'm relatively sure all of this technology is an outgrowth of high-sensitivity avalanche photodetectors used in conjunction with scintillators to detect very rare particle events like neutron-boson interactions. (Hint: Extremely non-profit primary research funded entirely by collaborative efforts which used to be largely funded by U.S. government grants but which have slowly but surely shifted to European and Asian government grants.)
Long story short, when we abandoned the SSC and the scaled down European project (LHC) took its place we lost primary research opportunities which, due to the nature of particle detection technology, could probably have helped American companies. The advancement of avalanche photodetector technology is directly related to this kind of high-cost, government funded primary research.
On a personal note, I left physics in large part because I would have been an experimentalist and for the past decade or more I saw that such a life would be half making due with shoddy equipment and half writing grant proposals. I'm sure I'm not the only one. Then I read about the LHC and feel sad that EVERYONE was thinking small the past 20 years or so.
I could with my remaining eye, but now I can't anymore.
I think we can keep recursing like this until someone returns 1
Can these avalanche photodetectors be fabricated into a matrix and used as a high-sensitivity imaging chip? I could imagine a low-cost night vision / low-light digital camera without the expensive vacuum photomultiplier tube. ISO 100,000 sensitivity here we come!
Speaking as someone who does university research in the field of silicon photonics:
Whenever Intel makes a marginal improvement on something, they fire up their massive PR department and get 100 news stories about it.
This is not a major development. This is barely a minor development. This paper is a 6% improvement (six percent!) over something that was published in 1997 (their Ref. 19 in the paper) by one of the co-authors.
As far as I can tell, the Wall Street Journal is the only place that did some background research and critical thinking instead of just copy-and-pasting the Intel press release:
http://online.wsj.com/article/SB122867271550286035.html?mod=googlenews_wsj
Speaking as someone who does university research in the field of silicon photonics:
Whenever Intel makes a marginal improvement on something, they fire up their massive PR department and get 100 news stories about it. This is not a major development. This is barely a minor development. This paper is a 6% improvement (six percent!) over something that was published in 1997 (their Ref. 19 in the paper) by one of the co-authors.
As far as I can tell, the Wall Street Journal is the only place that did some background research and critical thinking instead of just copy-and-pasting the Intel press release: http://online.wsj.com/article/SB122867271550286035.html?mod=googlenews_wsj
Would this be "Super-Infiniband", or "Super-Duper-Infiband..."... kinda tough to come up with a nickname that tops infinity. Who would have thought that Intel marketing might trap itself in the child's game of saying.. "Well I have 2x infinity bandwidth..." "I have infinity infinity bandwidth"...
Maybe they should just go straight to "Stay out of my airspace. Quit touching me! I'm not touching you! You are in my airspace.... hahah, my hand is just at the edge of your airspace but its in mine..."
This is my sig.
The real speed can be seen when we switch completely to photons and photonic circuits. No more electrons. Blazing Tera Hertz Transfer Speeds. Just a few decades away.