 t10
join:2003-05-25 Woodbridge, ON
| [Fiber Pics] Max theoretical speed over single strand of fiber?
Just as the topic states, what is the maximum theoretical data speed/bandwidth over one single strand of fiber.
This is providing you can use any possible technology, say like DWDM, as long as all data flows over one physical strand?
The reason I am asking, is that some sites have mentioned "unlimited" speed, which is impossible, there are just X amount of practical wavelengths possible. |
|
  ronpin Imagine Reality
join:2002-12-06 Nirvana
·AT&T Southwest
| Re: [Fiber Pics] Max theoretical speed over single
We should say practically unlimited. uDWDM (u= "ultra") is currently projected at up to 1024 lambda's. Currently each of those lambda's can use standard equipment to get 10gbs on each lambda. You do the math.
uDWDM is not in common use -- yet. |
|
  Agent 86
from: tschmidt 
| reply to t10 Re: [Fiber Pics] Max theoretical speed over single strand of fib
The theoretical limit has been estimated to be 100 terabits/s over long distances. |
|
 t10
join:2003-05-25 Woodbridge, ON | reply to t10 Darn, hopefully we see this in our lifetime ;p |
|
  whizkid3 Premium,MVM join:2002-02-21 Queens, NY
·Earthlink Cable Mo..
| reply to t10 The maximum data rate of a digital signal that can be transmitted over any medium, including fiber optics, is based on the following factors:
- the amount of power that can be transmitted - the amount of noise (together with above - the signal to noise ratio) - the bandwidth of the channel(s) - the mathematical coding system - the modululation type
The theoretical maximum data rate (or capacity) is found using the Shannon-Hartley Capacity Theorem. Assuming that you could find an ideal coding scheme (which would also require a tremendous amount of latency), you could approach the limit described in the following theorem but never exceed it:
C = BW log2 (1 + (S/N)) where:
C = theoretical maximum capacity (or data rate) BW = the bandwidth of the medium S/N = the signal to noise ratio
Looking at this formula, if you were to assume there was no noise, S/N would approach infinity, and therefore so would the capacity (maximum data rate). However, you can't assume there is no noise - it is a physical impossibility. The other problem is that the bandwidth of the fiber and the light spectrum that can be transmitted over it is also limited. To top it all off, the amount of noise is proportional to the amount of bandwidth. The theoretical bps per Hz is determined by substituting a formula for noise in terms of bandwidth into this formula. With the theoretical possibility to transmit an unlimited amount of power making the signal to noise ratio a huge number, such as 50 or 100 dB for example, the capacity tops out at about 16 bits per second per Hz. To find the answer to your question, you merely need to multiply the theoretical maximum bandwidth of the fiber by 16.
We will assume the fiber can transmit all visible light without any breaks in the spectrum, and that we have imaginary wide-band transmitters and receivers that could accommodate this. The visible spectrum is about 320 Terra-Hz wide. This give us a theoretical maximum data rate of 5120 Tbps. (5,120,000 Gbps or 5,120,000,000 Mbps).
Since we know that fiber optics can transmit infra-red and ultra violet light, the question becomes where does one draw the line at the bandwidth? Lets assume we can transmit all ultra-violet, visible and infrared light. The spectrum (conservatively) is then from 10^12 to 10^16 Hz. Basically the bandwidth would be about 10000 THz. This gives us a theoretical maximum data rate of 160,000 Tbps (160,000,000 Gbps or 160,000,000,000 Mbps). That should be enough 'speed' to handle most gaming applications. |
|
  somitho JNCIE Premium join:2004-08-17 Richmond, VA
·Verizon BroadbandA..
| reply to ronpin Re: [Fiber Pics] Max theoretical speed over single
said by ronpin :We should say practically unlimited. uDWDM (u= "ultra") is currently projected at up to 1024 lambda's. Currently each of those lambda's can use standard equipment to get 10gbs on each lambda. You do the math. uDWDM is not in common use -- yet. It isn't? SouthernCross and Apollo are the two largest uDWDM networks for undersea cable, and Level(3)'s entire network is uDWDM. Its also at 40 Gbit per lambda. So effectivly: 40,960 Gb/s.
whizkid did an EXCELLENT job of explaining above. -- Fiber Optics is the future of high-speed internet access. Stop by the BBR Fiber Optic Forum. |
|
  ronpin Imagine Reality
join:2002-12-06 Nirvana
·AT&T Southwest
2 edits | reply to whizkid3 said by whizkid3 :We will assume the fiber can transmit all visible light without any breaks in the spectrum Wrong. there are only 2 windows from 1300nm to 1600nm in which IR light will not be attenuated drastically. We must also have breaks in the spectrum -- between channels. Currently I think 25 GHZ is the current projected minimum channel separation (remember we're dealing with "Thz" bandwidth)
Since we know that fiber optics can transmit infra-red and ultra violet light, the question becomes where does one draw the line at the bandwidth? Lets assume we can transmit all ultra-violet, visible and infrared light. UV light? visible light? sorry, not for WAN communications -- maybe for short length office and medical applications.
(GREAT uDWDM PAPER -> »www.ee.ucr.edu/~ilyubomi/PTL_paper_2002.pdf |
|
  ronpin Imagine Reality
join:2002-12-06 Nirvana
·AT&T Southwest
| reply to somitho said by somitho :It isn't? SouthernCross and Apollo are the two largest uDWDM networks for undersea cable Sorry but uDWDM just isn't commercially viable yet. You mean DWDM (which ain't too shabby. I don't see anything about Level3 and uDWDM. Here's the lowdown on SouthernCross's 16 ch. DWDM:
The contract is for Alcatel to equip the third fiber pair in the Southern Cross network with 10 gigabits per second (Gbps) capacity per wavelength. Alcatel will supply undersea line Dense Wavelength Division Multiplexing (DWDM) terminals and terrestrial Synchronous Digital Hierarchy (SDH) equipment - both operating at 10 Gbps.
The upgrade of the Southern Cross Cable Network is made possible by using Alcatel super FEC (Forward Error Correction) technology for undersea terminals, officials said. Super FEC can supervise and correct detected errors of the signal transmission along the system for larger transmission bands than 2.5 Gbps. This makes it possible to send wavelengths operating at 10 Gbps over submersible cable and repeaters designed to operate at 2.5 Gbps.
The network also will benefit from multiplexers dedicated for high-speed and fully secured core transport systems with the Alcatel 10 Gbps multi-service node. Due to its non-blocking and powerful architecture, this system allows the choice of the most suitable network configuration for evolving network requirements as well as cost reduction and enhanced service quality.
The contract covers two phases. The first involves adding 50 Gbps (5x10 Gbps) to the initial total network capacity of 80 Gbps. The second phase will add a further 110 Gbps of capacity (11x10 Gbps) to reach a total of 240 Gbps of fully protected capacity. The two phases are expected to be completed by early 2003. (From 1999! )
Bell Labs Scientists Push the Boundaries of Optical Networking Capabilities Scientists from Bell Labs, the research and development arm of Lucent Technologies, have demonstrated two optical networking feats at this years analyst conference at its Murray Hill headquarters. On the display floor, Bell Labs researchers demonstrated a method whereby information at the rate of 160 Gbps over optical fiber using a single wavelength and plans to expand the system by using DWDM The 160 Gbps single-channel transmission demonstration was done over 300 km of Lucent's TrueWave RS fiber. Second, the researchers demonstrated the transmission of 1,022 wavelengths over a single fiber. Perhaps the most compelling aspect of the demonstration was the experimental transmitter used a single ultra-high-speed laser to generate signals over all 1,022 wavelengths, instead of using a separate laser for each, as is done in conventional multi-wavelength systems. Today's commercial optical systems combine up to 100 wavelengths on an optical fiber, using DWDM. Conventionally, DWDM signals are separate by 50 GHz frequency spacing, but the 1,022-channel system operates at a record high density of 10 GHz channel spacing. Each channel occupies only 10 GHz of frequency bandwidth, making this an ultra-dense WDM (UDWDM) transmitter. The 1,022-channel transmitter carries information at the rate of 37 Mbps of information, for a total system capacity of more than 37 Gbps. The researchers believe the system can be scaled up to OC-48 data rates, for a capacity of several Tbps. The team broke its own world record, the 1997 demonstration of a similar transmitter with 206 independent channels. Researchers plan further expansion of system capacity and density, and expect to use the new transmitter in novel data-networking applications. |
|
  Ignite Premium,VIP join:2004-03-18 UK clubs:
·BlueYonder Interne..
·Be There
| Worth mentioning that the kit either end is getting progressively smarter. At the moment the light has to be turned back into electrical signals when it reaches its' destination, however experiments are underway using magnetic fields to bend light and eliminate the need for repeaters etc to perform an electrical conversion. |
|
 bond787
join:2002-06-25 Guy | reply to t10 Re: [Fiber Pics] Max theoretical speed over single strand of fib
The instrument gauge that we use for testing goes from 0 to 9000 Mb/s. in my area we test these connections at a distance of 85 mi in between and it goes to 9000 Mb compare to T3 at 44 Mb/s Cable at 10 Mb and T1 at 1.54 Mb |
|
  Agent 86
| reply to whizkid3 whiz, that formula is not good for "any medium, including fiber optics". It is used for linear channels - which fiber is not. |
|
  Agent 86
| reply to somitho Re: [Fiber Pics] Max theoretical speed over single
"It isn't? SouthernCross and Apollo are the two largest uDWDM networks for undersea cable, and Level(3)'s entire network is uDWDM. Its also at 40 Gbit per lambda. So effectivly: 40,960 Gb/s."
I call B.S. - show me a link. |
|
  somitho JNCIE Premium join:2004-08-17 Richmond, VA | Call Alcatel for SouthernCross and ask for them to fax or email you the development study on SCCN's upgrade. |
|
  whizkid3 Premium,MVM join:2002-02-21 Queens, NY
·Earthlink Cable Mo..
| reply to t10 Re: [Fiber Pics] Max theoretical speed over single strand of fib
Ronpin wrote: quote: Wrong. there are only 2 windows from 1300nm to 1600nm in which IR light will not be attenuated drastically....UV light? visible light? sorry, not for WAN communications -- maybe for short length office and medical applications.
What part about this being a theoretical exercise didn't you understand? My post was not based on current technology, current silica fiber limitations, existing coding theory, etc. It was based purely on theoretical possibilities within the bounds of physics that can't be changed, even with new discoveries.
As far as the two windows you talk about, outside of which the light will be 'drastically attenuated', you should revise your knowledge. Historically there were three windows, including one around 850nm. I took a good look at the second edition of the 'Fiber Optic Communications Handbook', by the Technical Staff of CSELT, copyright 1990. Obviously, fiber production qualities have improved since then. From a chart of fundamental loss mechanisms in silica fibers, it is apparent that even if we limit our loss to 1 dB per km or less, this gives us a full window of transmission from 900nm to 1800nm. If we extend our acceptable loss to 10 dB per Km or less, we have a window from about 700nm to 1900nm. At the time, 3dB of loss per km was considered acceptable. In both single & multi-mode production-grade silica fibers there is a clear transmission window within these specifications larger than 800nm to over 1800nm. This was in 1990.
While 1300nm and 1550 nm certainly represent the 'sweet-spots' for fiber optic transmission, saying that outside this range drastic attenuation occurs, is not correct. These three windows are used today primarily because of both the historical inability to utilize the poorer quality fibers produced before this time at any other frequencies, and because of the very low loss that these bands provide. My exercise, of course, did not have to rely on even silica fiber and its properties, and it didn't. As well, attenuation is easily overcome with more power, especially when you rely on (like I did) 'future', super low-noise, wideband devices.
Will we ever be able to achieve the fantastic rates that I came up with? I certainly doubt it, nor do I think we will ever come anywhere near close. Even 1000 years from now.
quote: whiz, that formula is not good for "any medium, including fiber optics". It is used for linear channels - which fiber is not.
Really? In all my years, I have never seen any discussion of the Shannon-Hartley Capacity Theorem referenced to only 'linear' channels. In fact, I am not quite sure what you mean by a 'linear' or 'non-linear' channel - please elaborate. I certainly know what the difference is between linear and non-linear power transmission, but not exactly what you mean when referencing this to a 'channel' or medium. Honestly, I have never worked with wideband communications that were linear, so I fail to see how you have determined that Shannon-Hartley now doesn't apply to satellite communications, coaxial transmission, and fiber optics; all of which rely on 'non-linear' amplifiers to transmit power. Please do elaborate. I feel that we could both become very wealthy by proving them wrong, and becoming the next biggest names in communications history. ??? |
|
  Agent 86
| whiz,
I am surprised that someone familiar with the Shannon limit would not know what linear means. Informally, linear means free of distortion. In other words, noise is the only impediment to communication. The Shannon equation deals only with noise; there is nothing that relates to distortion.
A medium may be linear or not. Fiber happens to be a non-linear medium. |
|
  ronpin Imagine Reality
join:2002-12-06 Nirvana | reply to whizkid3 Re: [Fiber Pics] Max theoretical speed over single
OK -- I wasn't clear on the theoretical stance. A physics forum would've not even blinked  |
|
  whizkid3 Premium,MVM join:2002-02-21 Queens, NY
·Earthlink Cable Mo..
| reply to Agent 86 Re: [Fiber Pics] Max theoretical speed over single strand of fib
Quit wasting my time, Agent86, with your limited technobabble. Of course I know what linear means. You are jerking my chain. I suspect you have very limited knowledge with regards to communications systems, as the only things you managed to say in two posts so far:
- Shannon-Hartley theorem (which is true for all mediums) "deals with" noise. No kidding.
- Fiber-optics is not a linear medium.
I asked you to elaborate - you haven't and probably can't. |
|
  Agent 86
| whiz,
Are you Google-challenged? It's amazing that the conversation has gone on this long without you looking up the information for yourself.
Any information source on the Shannon limit will detail the precise conditions under which it holds true. And that is: a channel characterized by additive, white, Gaussian-distribution noise interference.
You can't simply pluck an equation out of context and apply it wherever you please. You have to understand the assumptions behind it. |
|
  whizkid3 Premium,MVM join:2002-02-21 Queens, NY
·Earthlink Cable Mo..
| reply to t10 quote: Any information source on the Shannon limit will detail the precise conditions under which it holds true. And that is: a channel characterized by additive, white, Gaussian-distribution noise interference.
Yes, that is exactly true. And this applies to fiber optics, as well as the lasers, amplifiers, and receivers used for fiber transmission. Interesting how you didn't mention anything about 'linear' this time around. I see you finally decided to look up what you profess to know on Google, no less. I don't need Google. I have enough experience, education, and quality books on my shelf about the subject.
Agent86, I'm not going to continue to go around in circles with you about this. You have yet to post anything relevant about the subject, with exception of the last blurb which you so conveniently plagerized, word for word. You are an amateur, and a poor one. See ya. |
|
  Agent 86
| "Yes, that is exactly true. And this applies to fiber optics, as well as the lasers, amplifiers, and receivers used for fiber transmission."
Wrong. It is clear from your previous posts that you don't know jack about fiber. You didn't know what range of spectrum was used for communication.
"Interesting how you didn't mention anything about 'linear' this time around."
You didn't understand linear: it was "technobabble" to you, remember? I keep trying different words, hoping something will register. No luck so far.
"I see you finally decided to look up what you profess to know on Google, no less. I don't need Google. I have enough experience, education, and quality books on my shelf about the subject."
Your posts speak for themselves. You haven't demonstrated any understanding whatsoever. Copying an equation from a book is not understanding. |
|