I have a Motorola Surfboard SB5100 cable modem and Time-Warner cable modem service here in Fuquay-Varina, NC. This modem/service provides me with about 45 kB/s uplink and 3 to 5 MB/s downlink speeds. Why are the uplink and downlink speeds asymmetric? Here's how my thinking goes. Sure, we're transmitting AND receiving on one physical set of copper wires (the cable). One easy way to do that is to frequency-division multiplex, i.e., set the rx band to be different than the tx band. However, wouldn't this be a waste of bandwidth? That is, the tx interference can be cancelled in the receiver since it is known. Likewise at the other end. So the entire bandwidth available over the wire could/should be available for both uplink and downlink. Where am I going wrong, or why isn't this the case? -- % Randy Yates % "So now it's getting late, %% Fuquay-Varina, NC % and those who hesitate %%% 919-577-9882 % got no one..." %%%% <yates@ieee.org> % 'Waterfall', *Face The Music*, ELO http://home.earthlink.net/~yatescr
Asymmetric Data Transfer Rates of Cable Modems
Started by ●September 14, 2007
Reply by ●September 15, 20072007-09-15
On Fri, 14 Sep 2007 22:27:38 -0400, Randy Yates <yates@ieee.org> wrote:>I have a Motorola Surfboard SB5100 cable modem and Time-Warner cable >modem service here in Fuquay-Varina, NC. This modem/service provides >me with about 45 kB/s uplink and 3 to 5 MB/s downlink speeds. > >Why are the uplink and downlink speeds asymmetric?There are several reasons, but the most basic reason is that by and large almost always the download demand per user is far greater than the upload demand. So from a resource allocation perpective it is much more efficient to design a multiple-access scheme where the uplinks are sized according to expected demand. Some people use this as an argument for why TDD systems are "better" (whatever that means) than FDD, since TDD allows (potentially) a more dynamic allocation of resources. Cable modems are pretty much all FDD with some spectrum allocated for the narrow uplink channels. TDD has it's own set of problems, there's no free lunch either way.>Here's how my thinking goes. Sure, we're transmitting AND receiving >on one physical set of copper wires (the cable). One easy way >to do that is to frequency-division multiplex, i.e., set the rx >band to be different than the tx band. > >However, wouldn't this be a waste of bandwidth? That is, the >tx interference can be cancelled in the receiver since it is >known. Likewise at the other end. So the entire bandwidth >available over the wire could/should be available for both >uplink and downlink. > >Where am I going wrong, or why isn't this the case?You're not going wrong anywhere. Things are the way they are because it's practical and manageable to deploy a system that way. One thing that cable modems have to work around is the legacy spectral allocations for the cable TV channels. You don't want the downlink or the uplink interfering with the TV or vice versa, and in a multiple-access situation it's not trivial to manage access and cancel everybody as an interferer. You can keep the user terminal low-cost and simple by taking a more direct approach. The cancellation of each ends Tx signal to reuse the same channel works reasonably well when the links are stable (i.e., not so much multiple-access going on, continuous streams both directions, whatever), and if the medium is expensive this can make a lot of sense. This is done in satellite channels in some instances these days since it can halve the cost of spectrum on the transponder by reusing it. It's a simple, almost degenerate, form of Multi-User Detection, but there's still no free lunch and it's not trivial to implement. Advanced techniques like smart antennas(e.g., beamforming), spatial multiplexing, Multi-User Detection, etc., are all great and all have something to offer, but all have costs and limitations. Often the implementation cost is high and there's a corresponding increase in power consumption or diminishing returns happen more rapidly than expected or whatever. It'll be interesting to see just how well the 802.11n devices work in the field, since they include a number of advanced techniques (and lot more optional techniques that may never be implemented, but may also get shaken out at some point). Early reports are hit and miss, and I'm interested to see really how well low-cost spatial multiplexing can be made to work in general. I suspect it's not going to be as good as many have anticipated, but we should get a chance to see over the next few years. Eric Jacobsen Minister of Algorithms Abineau Communications http://www.ericjacobsen.org
Reply by ●September 15, 20072007-09-15
Eric Jacobsen <eric.jacobsen@ieee.org> writes:> [...] > The cancellation of each ends Tx signal to reuse the same channel > ... [is] a simple, almost degenerate, form of Multi-User Detection, > but there's still no free lunch and it's not trivial to implement.What's so hard about it? Ok, maybe it degrades the bandwidth both ways somewhat, but how much? I would think it's a matter of more dynamic range in your ADCs, so it's going to be a good bit expensive in hardware costs. Is this the key issue? -- % Randy Yates % "Remember the good old 1980's, when %% Fuquay-Varina, NC % things were so uncomplicated?" %%% 919-577-9882 % 'Ticket To The Moon' %%%% <yates@ieee.org> % *Time*, Electric Light Orchestra http://home.earthlink.net/~yatescr
Reply by ●September 15, 20072007-09-15
"Eric Jacobsen" <eric.jacobsen@ieee.org> wrote in message news:qrjme3930g6dqqop6eoajnhpa6e8e1mfrg@4ax.com...> Some people use this as an argument for why TDD systems are "better" > (whatever that means) than FDD, since TDD allows (potentially) a more > dynamic allocation of resources. Cable modems are pretty much all > FDD with some spectrum allocated for the narrow uplink channels. TDD > has it's own set of problems, there's no free lunch either way.The difficulty of the TDD over the copper is that the trunk has to be synchronized despite of the different delays in the different branches. The mutual interference between the pairs due to the lack of synchronization is a heck of a problem.> The cancellation of each ends Tx signal to reuse the same channel > works reasonably well when the links are stable (i.e., not so much > multiple-access going on, continuous streams both directions, > whatever), and if the medium is expensive this can make a lot of > sense.It works for the conventional dialup modems but not for the DSL. It is not feasible to extract the Rx signal from ~60dB depth below the local Tx in the same band. Especially if the bandwidth is 10+ MHz.> Advanced techniques like smart antennas(e.g., beamforming), spatial > multiplexing, Multi-User Detection, etc., are all great and all haveIn the copper? :-) Vladimir Vassilevsky DSP and Mixed Signal Consultant www.abvolt.com
Reply by ●September 15, 20072007-09-15
"Vladimir Vassilevsky" <antispam_bogus@hotmail.com> writes:> [...] > It is not feasible to extract the Rx signal from ~60dB depth below the local > Tx in the same band. Especially if the bandwidth is 10+ MHz.Why not? -- % Randy Yates % "Remember the good old 1980's, when %% Fuquay-Varina, NC % things were so uncomplicated?" %%% 919-577-9882 % 'Ticket To The Moon' %%%% <yates@ieee.org> % *Time*, Electric Light Orchestra http://home.earthlink.net/~yatescr
Reply by ●September 15, 20072007-09-15
"Randy Yates" <yates@ieee.org> wrote in message news:m38x783bol.fsf@ieee.org...> I have a Motorola Surfboard SB5100 cable modem and Time-Warner cable > modem service here in Fuquay-Varina, NC. This modem/service provides > me with about 45 kB/s uplink and 3 to 5 MB/s downlink speeds. >If the "B" in both of your numbers stands for "bits", then your service is broken. Time-Warner advertises 384kbps (bits-per-second) upstream and 5Mbps downstream. How did you arrive at your numbers?
Reply by ●September 15, 20072007-09-15
Michael K. O'Neill wrote:> "Randy Yates" <yates@ieee.org> wrote in message > news:m38x783bol.fsf@ieee.org... > >>I have a Motorola Surfboard SB5100 cable modem and Time-Warner cable >>modem service here in Fuquay-Varina, NC. This modem/service provides >>me with about 45 kB/s uplink and 3 to 5 MB/s downlink speeds. >> > > > If the "B" in both of your numbers stands for "bits", then your service is > broken. Time-Warner advertises 384kbps (bits-per-second) upstream and 5Mbps > downstream. > > How did you arrive at your numbers? > >Usually B --> bits b --> BYTES ;)
Reply by ●September 15, 20072007-09-15
"Michael K. O'Neill" <MikeAThon2000@nospam.hotmail.com> writes:> "Randy Yates" <yates@ieee.org> wrote in message > news:m38x783bol.fsf@ieee.org... >> I have a Motorola Surfboard SB5100 cable modem and Time-Warner cable >> modem service here in Fuquay-Varina, NC. This modem/service provides >> me with about 45 kB/s uplink and 3 to 5 MB/s downlink speeds. >> > > If the "B" in both of your numbers stands for "bits","b" represents bits, "B" bytes. -- % Randy Yates % "I met someone who looks alot like you, %% Fuquay-Varina, NC % she does the things you do, %%% 919-577-9882 % but she is an IBM." %%%% <yates@ieee.org> % 'Yours Truly, 2095', *Time*, ELO http://home.earthlink.net/~yatescr
Reply by ●September 15, 20072007-09-15
"Randy Yates" <yates@ieee.org> wrote in message news:m34phvlhvn.fsf@ieee.org...> "Michael K. O'Neill" <MikeAThon2000@nospam.hotmail.com> writes: > > > "Randy Yates" <yates@ieee.org> wrote in message > > news:m38x783bol.fsf@ieee.org... > >> I have a Motorola Surfboard SB5100 cable modem and Time-Warner cable > >> modem service here in Fuquay-Varina, NC. This modem/service provides > >> me with about 45 kB/s uplink and 3 to 5 MB/s downlink speeds. > >> > > > > If the "B" in both of your numbers stands for "bits", > > "b" represents bits, "B" bytes.Then you're not paying attention to both of your numbers. A 45 kbytes/sec download speed is reasonable, but a 5 megabyte/sec upload speed equates to 40 megabits/sec and is not a typical number that can be achieved through Time-Warner.
Reply by ●September 16, 20072007-09-16
"Michael K. O'Neill" <mikeathon2000@nospam.hotmail.com> writes:> "Randy Yates" <yates@ieee.org> wrote in message > news:m34phvlhvn.fsf@ieee.org... >> "Michael K. O'Neill" <MikeAThon2000@nospam.hotmail.com> writes: >> >> > "Randy Yates" <yates@ieee.org> wrote in message >> > news:m38x783bol.fsf@ieee.org... >> >> I have a Motorola Surfboard SB5100 cable modem and Time-Warner cable >> >> modem service here in Fuquay-Varina, NC. This modem/service provides >> >> me with about 45 kB/s uplink and 3 to 5 MB/s downlink speeds. >> >> >> > >> > If the "B" in both of your numbers stands for "bits", >> >> "b" represents bits, "B" bytes. > > Then you're not paying attention to both of your numbers. A 45 kbytes/sec > download speed is reasonable, but a 5 megabyte/sec upload speed equates to > 40 megabits/sec and is not a typical number that can be achieved through > Time-Warner.Ayup - I meant 3-5 Mb/s. -- % Randy Yates % "And all that I can do %% Fuquay-Varina, NC % is say I'm sorry, %%% 919-577-9882 % that's the way it goes..." %%%% <yates@ieee.org> % Getting To The Point', *Balance of Power*, ELO http://home.earthlink.net/~yatescr
