Channel based vs Packet based ?

Discussion in 'Wireless & RF Design' started by Mathematics!, Mar 24, 2012.

  1. Mathematics!

    Thread Starter Senior Member

    Jul 21, 2008
    Ok , I pretty much know all the major ways of modulating digital and analog data. In principle there are only 3 ways to vary a sine/cosine carrier wave
    amp , phase , frequency ( digital are ASK , PSK ,FSK / analog versions are AM ,PM ,FM )

    Most of the ones used are a combination of them like QAM 16,64,256 ,...etc as well as others.

    I also know to have multiple people using a wireless or wired media at the same time you either have to do some sort of round robin or multiplexing technique.

    For the round robin you could use CSMA avoidance/detection techniques or just a time slot.

    If you want to transmit all people at the same time then you would have to do some multiplex technique like TDM ,FDM CDMA ,COCDMA , ...etc etc there are many choices but in principle they all use some form of time, amplitude , frequency , phase , use multiple carriers (i.e vary shape) and or combination of them to allow people to transmit all at the same time with no interference. (I know by doing the math which ones are better in which situations )

    Channel based transmissions ones that are used in mobile phone networks MUST USE MULTPLEXING or then we would have chaos if a ton of people wanted to use it forget round robin or waiting.

    But then we have Packet based based transmissions like the ones in 802.11 or ethernet home LANS. In these multiple people really cann't transmit all at once like the mobile phone network allows but they use CSMA detection/avoidance which usually a person wireless/wired card waits for the channel to be clear to send. Same thing when it comes to going out my cable internet modem.

    I am curious why don't we just all use the multiplex techniques and stop using the CSMA stuff and use the multiplexing technique that allow us to have many people sending there data all at once. (i,e with no waiting to see if the channel is clear or other things)

    Question 2
    Is this sort of CSMA avoidance/detection / time slot info sending just used in the computer LAN and modems to the ISP. Or is there more places it is used in today. Because I just cann't see the waiting being necessary when you have far superior multiplexing technique to allow everybody to transmit at once with out worrying about collisions or interference.

    Question 3
    Doing Fourier series and the math out it is easy to see how to mathematically extract the data from the carrier or demodulate /modulate the data. But this is just on a piece of paper I am more fascinated or curious on what we can do with actually physically building these modulators/demodulators from the math.
    I have built simple FM,AM transmitters / receivers over the years and I know the basics of how to create an oscillator , detector , tuner, ...etc circuit from simple resistors,caps,inductors...etc

    But I am amazed that we can physically build all those modulators/demodulators circuits that you can do out on paper.

    Curious from a hobby point of view is it possible to build detectors/modulators other then the simple ones with out special chips or hardware.
    I know in theory you can do anything with enough cap , resistors,transistors, ...etc and enough space/time but I would think it would get to the point where it would take a life time to build these from simple electronic components/from scratch.
    So is it true that other then the simple AM , FM transmitter/receivers the only reasonable way to create the more complex ones to send away for special QAM chips ,...other chips etc etc.

    Though correct me if I am wrong
    I know these questions are rather all over the place but my main one is if anybody knows more on why we use CSMA when we have all the multiplex things to never have collision issues. ( as well as if we use things like CDMA and other spread spectrum based multiplexing technique gives use are data security and prevents in most cases jamming signals)

    Thanks for an info I hope you understand.
    Last edited: Mar 24, 2012
  2. K7GUH


    Jan 28, 2011
    I urge you to read up on the various digital techniques employed by amateur radio operators, particularly on PSK-31. Don't know if it will answer your question(s), but it might be a good place to start.
  3. nsaspook

    AAC Fanatic!

    Aug 27, 2009
    We are using MIMO technology to multiplex channels on a single carrier. There are only so many channels that can be allocated with any multiplex scheme (far less than the number of users in most cases) so a resource allocation method will have to be used. Because mobile air/ground based radio communications link stability can vary by a huge amount in a short time it's usually better to use the full bandwidth of the link to send messages in the shortest time possible.

    I suggest you do a little research on the subject of radio based packet switching , queueing theory ,latency time and channel utilization by load.
    Chap 6: Packet
    Chap 7: CSMA
    in my old 1st edition textbook.
    Last edited: Mar 24, 2012
  4. Mathematics!

    Thread Starter Senior Member

    Jul 21, 2008
    Ok, thank you.

    CSMA CA is really a one person sends data at one time then another...It really is not like multiple people can send at the same time like multiplexing does. So I am getting confused between when we use CSMA and when we can just use a multiplexing scheme and have everybody sending at once instead of waiting for the channel to be clear to send like CSMA CA does. Seems like CSMA CA being used would bog down the network since there would be alot of wait time for each person/device. As well as if the wait time was to great you would get choppy calls or the device would run out of storage space for the waiting data in some cases.

    Looking at it from a cell phone network point of view.

    The high level view of things
    I understand any of the modulation/demodulation schemes out there
    I understand any of the multiplexing/demulitplexing schemes out there
    But what I cann't understand is when/where the CSMA stuff falls into.

    Seems to me modulation/demodulation allow you to encode/decode the data from the carrier
    And multiplexing/demultiplexing allows you to transmit multiple peoples data all at once on the carrier
    So with those 2 facts why would you need CSMA CA CD since they are only there to make sure the channel is clear / no errors.
    The whole point of multiplexing is to avoid the need for a person to send at a time so I see know point in CSMA CA
    Last edited: Mar 24, 2012
  5. nsaspook

    AAC Fanatic!

    Aug 27, 2009
    The problem is slot queueing on the multiplex with more (random) users than channels. You can dynamically (uses bandwidth , needs state and has timeout problems with mobile links) hand out tokens for each user for a channel or allocate a static channel (inefficient) and drop users. The actual CSMA wait times are short due to good exponential backoff.

    It seems counter-intuitive at first that CSMA would be better but in real-life, random user, high utilization, high contention, lossy network (cell phone) links it is with limited resources.
  6. Mathematics!

    Thread Starter Senior Member

    Jul 21, 2008
    O, I think I get it.
    Correct me if I am wrong CSMA CA only comes into play/or is need if the users exceed the channel capacity /(the multiplexing capacity )

    But if you had always enough channel capacity / bandwidth to support all the users. Then in theory CSMA CA would not be needed. And know wait time would ever have to exist in sending anything. Correct me if I am wrong.

    If that is true then we use CSMA CA in conjunction with a multiplexing scheme only when we are not sure that we will have enough channel capacity to support all users all the time at any instance.

    If that is true then in most home lans and stuff that only a few users every use at once CSMA CA is really rarely a major factor with these types of small networks provided the channel capacity is always > the users there would be really no use.

    So really CSMD CA is really only comes in to aid multiplexing when the user capacity gets to large for multiplexing to handle all at once. Users simple have to wait for a turn/space on the multiplexing signal.

    Please let me know if I am miss understanding it.
    Last edited: Mar 24, 2012
  7. nsaspook

    AAC Fanatic!

    Aug 27, 2009
    That's correct, any time there is demand above the max link capacity it starts to restrict access and constrain the combined bandwidth of all users (of equal priority) to match the link capacity. In the almost 30 years since I studied networking (long before modern cell networks) nothing really has changed except they now pack even more users into limited channels and charge you more for it.
  8. Mathematics!

    Thread Starter Senior Member

    Jul 21, 2008
    agreed, we had the theory all mapped out when Fourier came around in the 1800's. And in theory it will never change other then getting better hardware and modulating/multiplexing schemes. So we can pack more data and modulate more users faster with less space/bandwidth ...etc

    Things like OFDM , COCDMA ,...on word. The only change is improving on the hardware to do more harder schemes...etc

    That is why it should really be 1g to 2g was analog to digital switch but 2g onward is just improving on the modulation/multiplexing ,...etc as such I really don't think it should be called a totally new generation 2g onward.

    But I agree some one like you could pick up another system in no time.

    So CSMA is more of a protocol built for the devices to not collied with one another where as modulation/demodulation and multiplexing/demultiplexing are circuits you have to build. Once you have those circuits CSMA is just a protocol that runs on top to make sure when modulating data it won't screw up any other data...etc

    Makes total sense now.
    Appreciate the help. That had always bothered me and I could never understand why we had CSMA when we had multiplexing. Now I do thanks again.
  9. WBahn


    Mar 31, 2012
    I may have missed it, but one important point that has to be made is that even if I have more than enough channels available, there still has to be some way for the two people that want to communicate with each other at this moment in time of agreeing on which specific channel to use. Let's consider a simple example in which we have a hundred different radios in service but, at any given time, only twenty ever want to be on the air in one-to-one communication at a given time, hence ten different channels is all I ever need. But if person A wants to talk to person B, which channel do they use? In theory, we could assign a unique channel for each possible combination, but this would require nearly ten thousand channels, only ten of which would be in use at a given time, and would require that every user update their list any time a new user was added. Instead, we can designate one channel, call it Channel 0, and specify that anytime anyone wants to initiate a conversation, they make a call on Channel 0 to establish contact with the other person (assuming they are listening) and to agree upon one of the other channels to shift to (assuming it is open). But this means multiple people will want to use Channel 0 at the same time and so we need some kind of scheme to make that possible.

    This is how truckers still use Channel 19 on the CB, for instance. A lot of people think that truckers ONLY use Channel 19. They will use it for their entire conversation if the channel isn't in use otherwise, but if multiple people are using it, then most will limit their use to agreeing on another channel to try.

    Now, you can make the simple scenario described above as complex and as sophisticated as you want, but at the end of the day (in anything but the smallest, most tightly controlled networks) you will always be left with the need to have a common channel that allows multiple users to share it for the purposes of dealing with network joiners.

    In turn, this means you need to have one of two things - some form of medium access control (MAC) protocol to limit use to one user at a time (within a given range) or some form of modulation that permits multiple transmissions to exist in the same channel without interferring with each other and while allowing everyone listening to receive all of the transmissions. To the best of my knowledge, there is only one such modulation scheme in existence today, based on concurrent coding theory, and it has not been commercialized yet (and probably won't for some time).