 So, we're moving, we're gradually moving away from the details of getting data across links and the next topic after this we'll move on to how do we get data across a network, across a set of links, okay? We're focused mainly on how to get signals across a link and those signals represent bits or some data and the previous topic about how to efficiently get that data as frames across links by doing flow control and error control. This topic is just to mention the concept, what do we mean by multiplexing? Before we go into it, just a reminder, remember when we send a signal we can think that when we transmit a signal containing data that that signal is transmitted at some center frequency but it covers a range of frequencies. Remember a signal, our earlier lectures we drew a plot of the spikes, the impulses at different frequencies and we calculated the bandwidth of a signal. So in general we can talk about a signal has some bandwidth and we transmit a signal across a range of frequencies but often we just talk about one frequency. For example when I transmit to the Wi-Fi access point the frequency used is approximately what? What frequency do we use for Wi-Fi? About 2.4 gigahertz, okay? It's more specific than that so the frequency that we transmit is about 2.4 gigahertz but in fact the signal is a range of frequencies so plus or minus a little bit. The bandwidth is about 20 megahertz so it's 2.4 gigahertz plus or minus about 10 megahertz that's the range of frequencies we send. What if two entities transmit using the same frequency? What happens at the receiver? For example two laptops transmit to the same access point, two people talk at the same time, two lines transmit, what happens at the receiver? Q, think of the signal received by the receiver, what do they receive? There are two signals transmitted, let's say they're received by the one receiver, what happens at that receiver with respect to those signals? You can think, remember if you remember back to noise we have, let's say from the perspective of one of those signals received the other is noise and they add together and what happens at the receiver is that it receives both of the signals at the same time and as a result it cannot understand either of them, it's like they interfere with each other. It's the same when two people talk, when I'm talking and someone else is talking someone else hears both of those signals and they cannot understand either because those signals interfere with each other. So that's a problem that we don't want to occur. With wired communications, does that problem occur? Why not? Some people are shaking their head, no, no, so okay, correct, why not? Why is there no interference? I'll remember back that the term for wired we said it was guided, we can think that the signal, say the electrical signal is contained within the cabling. It doesn't disperse out much, the way that it's designed with the twisted pair for example the outer coating, you can have one cable here, transmit a signal across it and another cable nearby a few centimetres away, transmit another signal and they will not interfere with each other. So with wireless communications and some systems if we transmit at the same time we can interfere but with cabling we can transmit two pieces of data at the same time, two signals at the same time across separate cables and there will be no interference, okay. So now let's say we have our users here at SIT, at Bankere on this campus and we have a hundred different users and they all want to send data to a hundred different users at the Rung SIT campus and do it at the same time, how can we do that? Well the basic way is if you think of the users having computing devices is to have a cable from each computing device here to each computing device at Rung SIT, okay. So my computer in the office has a cable going to someone else's computer in Rung SIT, your computer has a cable going there as well and everyone else does. Is that good? It's good from the perspective of interference that is again we said with cables everyone can transmit at the same time and they will not interfere with each other, what's bad about it. The cost and convenience is terrible that is everyone for every pair of users we need a separate cable, maybe we could do that today but now another user comes along then we need to spend a week laying the cable just so that they can communicate. So what should we do? We don't want to lay cables for every pair of users. We have one cable and share it amongst the users and that's what multiplexing allows us to do. It allows us to use one link but have many transmissions of different pieces of data across that one link sharing that link amongst multiple users. This set of slides has the concept of multiplexing and then a few on multiple access and they are very similar. Multiplexing is really relevant most times to point to point links. So cables for example, connecting two devices. Multiple access, same issues but relevant to point to multipoint links like Wi-Fi access. We will not touch upon multiple access here, just introduce multiplexing. So the idea. If we have multiple users at one location say these sources A, B, C and D at our campus and they all want to send data to some corresponding sources at another location then the basic way would be to give them a cable each, a link each. So I have four different cables connecting those two locations one per source destination pair. And they would transmit a signal and what I've tried to plot in this example is the signal and the frequency domain. Think of a signal has some bandwidth B and some frequency F1 where that's the centre frequency of that signal. So because we have cabling they can all transmit at the same time at the same frequency and not interfere with each other because the signal is contained inside. So that would work but we just said it's wasteful of resources because for every pair of users we need a separate cable, very costly, very inconvenient. If you want to add a new user you need to lay a new cable. So it's hard to expand. It's good because there's no interference and each user has a dedicated link. There's no sharing of resources but very inefficient. So multiplexing tries to achieve the same thing by using one cable between the two locations and having the users share that cable. So having the signals from each user be transmitted across that cable so that the corresponding destinations receive it. So there's a single line that connects the two devices. By now we introduced some special devices and they are called a multiplexer, a MUX and a demultiplexer. And what happens is that the sources transmit their signal the same as the original case so they all transmit their signal at a particular frequency with some bandwidth. The multiplexer combines it somehow and sends a single signal containing the data of all of those users to the demultiplexer which splits it up and sends the corresponding signal to the intended destination. So that's what we want to achieve, allow all users at one location to send their data to the corresponding users via a single line. And we do multiplexing to do that. And there's two basic ways which are used to combine the signals. First one is called frequency division multiplexing, FDM. Think that the users have their original signal that they transmit to the multiplexer. So this source A through to D and the MUX are at one location. This is just a special device that they connect to. They transmit their signal and it's the same as, if I just jump back, as this one. So A transmits a signal at frequency F1 with bandwidth B. So does B, same frequency, same bandwidth to carry their data. They can do that because from A to the MUX and B to the MUX, there are separate links there, the separate cables. What the multiplexer does is combines those signals and with frequency division multiplexing, it combines them and shifts them to be transmitted at different frequencies. So if we think of the signal sent from the multiplexer across our link, that signal contains the signals from each of the four sources, but shifted to different frequencies so that they don't interfere with each other. So in this example, the data from source A is sent at frequency F1, still has a bandwidth of B, but the data of B is shifted to a different frequency, F2. And C is at F3 and D at F4. So now be careful, this is a signal versus frequency. So the signal transmitted by the multiplexer contains frequencies from here, the lower part of the signal from A, up until the higher part from the signal from D to here. A bandwidth of at least four times B, B four times. And as is the normal practice, there's always some spacing in between the signal. So that's why I write greater than four times the bandwidth of the individual components. So here, each user wants to transmit a signal with bandwidth B. The multiplexer combines them together, but shifts them such that they're not overlapping and transmits a signal at least four times the bandwidth of those individual users. So sends one signal in that case. The idea is that the D multiplexer that receives this signal, it knows the signal components at this lower end must be sent on the destination A. The signal components within this, the frequency around F2 must be sent to B and around F4 to D. So it splits up. And what A receives is the same as what source A, or destination receives the same as what source A sent. So the idea is to transmit the individual user signals at different frequencies all at the same time across that shared line. Because they're at different frequencies, there's no interference because the D multiplexer can detect and receive all of that and then splits it up based upon the known frequencies. So the signals from each user are transmitted at the same time, but different frequencies. We say that we can divide the sharing of the line up by frequency, and we get frequency division multiplexing. Any questions? On FDM, this will be a short topic, but let's make sure it's clear. The signals from multiple users at the same time, but each user is allocated a different frequency by the multiplexer. Another way to visualize that is to think, this example, think there are six users. So each user is allocated a different frequency, F1 through to F6, but they send a signal at the same time. So they use different frequencies, but they all transmit at the same time. We'll see if this makes sense when we compare it to the other approach. And the transmissions of each user, and when I say user, it doesn't need to be a human user. It's some allocation to that we allocate to a particular purpose, an application, a computing device, for example. And we name really those users as a channel. So we talk about channel one is transmitted at frequency F1, channel six at F6. So we sometimes just refer to channels. It's possible when the medium of our link has a bandwidth which is larger than the bandwidth of the data that we need to send. In this case, the data that the users need to send was a bandwidth of B. The bandwidth of this link in this example must be at least four times B. If it's not, this won't work. We will not be able to separate the signals from the individual users. And in practice, usually because the signals are not perfect, it's not a perfect accuracy in terms of bandwidth. We usually separate them, have some spacing between them. This is what's used, for example, in cable TV, or even in normal TV with terrestrial TV. With cable TV, for example, you have a cable coming from the cable provider, the company, coming into your home and connects into your TV. And what the cable company does is effectively transmits the video content from all those different stations over the same cable at the same time. So there are many different data sources. All the different TV stations are the different users in that perspective. It's all transmitted at the same time. One signal across your cable, and when it comes to your home, your TV tuner, selects one frequency to listen on to receive. So the way that that works is that each TV station, the video content is transmitted at different frequencies so that they don't need to feel with each other. So that's why you call them TV channels. What TV channels are there? What are the popular ones in Thailand? The free-to-air ones, three, no longer free-to-air, is it? Three, five, seven, nine, I think are the four main ones, but there are many others, of course. Well, those channels actually specify a particular frequency at which they are allowed to use. And I don't have it for Thailand, but you can look it up and find the exact numbers. So I have a plot which maybe recovers the world, but different regions. There are theories across different countries as to what channel corresponds to what particular frequency. I'll zoom in in a moment, but what we'll see here is just regions or countries in the world, and here is the range of frequencies, and we'll see the TV channels. And it's very similar. If we zoom in, we'll see, say, the top one, Western Europe. So, for example, in Europe, the yellow ones at the top, the frequencies here are 40 MHz, 50 up to about 70, channel one, channel two, channel three, channel four, for example. So channel three is transmitted with a frequency of around 59 MHz, sorry, 54 MHz up to 61 MHz. So that's the one particular channel. The bandwidth is six or seven MHz. That's the typical bandwidth for a TV channel. Actually, the TV channel contains both audio and video content. So it's split up within that one part for video, one part for the audio. And it's similar through different parts of the world, and it's separated that up here you get to the higher channel. So channel seven is around 190 MHz, plus or minus a few MHz. So the different channels are specified so that you can transmit all at the same time via a cable or even via the air and not interfere with each other. So this is an example of using frequency division multiplexing. Any questions? What is the difference between NTSC and PAL? Anyone want to answer? What are they? Is it a better question than what is the difference? What are they? I cannot give a correct answer because I don't know the details, but it's the way you encode video. So what is video? Pictures which are moving, but the way that TV encodes it is that you think, so digital video you can think of pixels and frames changing. But analog, the way that it worked is that you think of a frame of the image and to encode it what would they do is scan line by line. So scan, so the image would be sent to your TV and the way your TV displays it is on an old analog TV is display line by line. So create really the pixels at the top and then come back and scan and go through the whole frame and then come back up to the top to display the image. There are different algorithms for how to scan line by line. Sometimes you scan I think each line, sometimes you go I think every second line and then come back and fill in them in between lines. So you get interlaced and progressive scanning. So you see things like even with HDTV, well 1080p and 1080i referring to progressive scanning and interlaced scanning. NTSC and PAL from my knowledge are just two different ways to represent the video for TV transmissions. One NTSC for US and other regions and PAL came from Europe and spread in other regions. So that's how to encode the video and also parts of the audio I think. Go read up about TV to find out more about the encodings and even the channels. But just one example of FDM. Any other questions on frequency division multiplexing? What's the alternative? So here we have multiple users send at the same time using different frequencies. The alternative, come back, send at the same frequency but using at different times. Take in turns. So we get time division multiplexing. So in this case the multiplexer receives the signals from each of the sources and transmits a signal out which has the same bandwidth as the original sources. So it's no longer 4 times B, it's back to B. But instead of transmitting the signal from A through to the D multiplexer and same with B, C and D at the same time it alternates or switches between them. And what I've tried to show here is across time the multiplexer would transmit a signal for some period of time representing the data from A so starting at time t1 let's say for some fixed time period, maybe 1 millisecond then it transmits the signal for t2 for 1 millisecond, t3, t4 then back to t1 and keep going like that. So we transmit a signal with the same frequency as all the users but we transmit the signal carrying the data from the users at different times. The D multiplexer knows the timing so when it receives this signal for the first say 1 millisecond it knows that corresponds to the user A and sends that corresponding signal to destination A then the next portion of time it sends that to B, C and D and then back to A and it keeps switching through those. So this is time division multiplexing taking turn, which one's better? Which one performs better? Or maybe easier, what's the difference with respect to performance of FDM and TDM? Maybe go back to FDM What's the problem with FDM here? It's the bandwidth. So the data from A requires a bandwidth of B but the link here we need a bandwidth of 4 times B so that we need at least 4 times the bandwidth of the individual user bandwidth so that's the disadvantage here we need to use more bandwidth and we know more bandwidth is more costly What about with TDM? We use the same bandwidth B What's the problem? We send less data in the same amount of time or we need more time to send the same amount of data that is from user A's perspective it's only getting its data sent one fourth of the time here and then it waits for the other three users and then the user A again so we get less time per user in this case which turns out that they are about equivalent one uses more bandwidth or the other way you need more bandwidth to get the same performance this one you need to send faster higher data rate to get the same performance so that we can get the same end-to-end performance from source through to destination the other differences are really the difference of implementation and which ones are easier to implement so there's no one better than the other with respect to performance of sharing bandwidth or share time we can see it here we can think of the time as time slots each user gets a time slot and again it's called a channel so here the channel doesn't refer to frequency it refers to the time slot each user gets I think the signal is sent using the same range of frequencies but at different times we switch between different channels channel one through to channel six one through to six and so on it's primarily used for digital signals whereas FDM it can be used for analog or digital signals or we deal with analog signals and with FDM in the main cases synchronous TDM is really you take in turns it's always one two three four five six one two three four five six asynchronous or statistical TDM is that you can do it on demand maybe it's one two three and then it comes back to one because user one needs more time slots to transmit more complex but can be more efficient for example of TDM there are many systems right so back to FDM broadcast or cable TV radio uses FDM your AM and FM radio send at different frequencies at the same time some long distance telecommunication systems in older telecom networks they used FDM a lot still do in some cases optical fiber we have one fiber with optical fiber we send light signals and what we can do with optical fiber is send different light signals at different frequencies at the same time and the receiver can detect so the light signals with different frequencies sometimes we think of different colors so that uses frequency division multiplexing but it's often called wavelength division multiplexing remember lambda the wavelength equals the speed of light divided by frequency so the same concept you either give them a different frequency or give them a different wavelength it just turns out with light and optics we often talk about the wavelength not the frequency at home internet access with dial up modems not ADSL dial up modems pre ADSL could you access if you ever use them could you access the internet and call at the same time no the way that they worked is that the data that you send and receive use the same range of frequencies that your voice would use when you talk so of course you couldn't do both because if you talk while you're sending data the risk that would interfere with each other what ADSL does is uses a wider range of frequencies of that copper wire from your home and divides it up for the different purposes the different users really and it differs in different variants of ADSL one example is the voice the voice range of frequencies are sent from frequencies of 0 up to about 20 kilohertz it's called POTS plain old telephone service so just the voice the old voice communications with a telephone are sent with this range of frequencies one cable from your home to the ISP your upload data when you're sending up is transmitted at a different range of frequencies the upstream and when you're downloading data because you can upload and download at the same time with ADSL you're receiving at a different range of frequencies so think of these are the three users that use for ADSL one for voice, one for upload, one for download so that's an example of FDM and it doesn't have to split evenly as you see here downlink gets a much larger bandwidth than uplink and voice because often that's more convenient for the users people want to download much more than they upload in many cases 20 kilohertz to cover the human voice TDM some common examples are used in wide area networks connections across cities between cities and between countries and usually using digital so they use digital systems and they replace some of the older FDM systems so the telecom companies which connect between cities and between countries use technologies which use TDM and the names of some of them are PDH and which use electrical cabling and it's been really improved become SDH you can SDH is synchronous digital hierarchy PDH is plesioncranous digital hierarchy used by large telecom companies and ISPs to connect between cities and so on we don't need to know much more about them this just with SDH some of the names or terminology used and some of the data rates for example links if you look at the map of the internet for Thailand I think you may have had it on front of your handouts and you see that some of the links there are in the order of 10 gigabits per second especially the international links they are most likely using SDH and different variations so 10 gigabits 2.5 gigabits, 40 gigabits per second these two are quite common for international links nowadays they use optical fibre and the OC it means it's an optical carrier just some details just as an example questions about multiplexing FDM and TDM what if we transmit TV using TDM? we can but usually with TDM we're transmitting digital signals so we do transmit TV with TDM but usually not to the end home user so think of maybe maybe satellite TV there's the TV stations that generate the content the audio and video they send that eventually to some satellite gateway that transmits up to the satellite and then down to the homes so that down to the homes may be using FDM different frequencies for the different channels whereas the content from the TV stations to the gateway may be transmitted as digital data using TDM and then it's just converted so it would work fine it wouldn't degrade quality as long as the right data rates were used for cable systems TDM is mainly used today the wireless systems FDM is commonly still used in the past FDM was a bit more common in wired systems but no longer multiple access have a look through there if you're interested but a lot of it's relevant to wireless systems like your mobile phone and wifi you've got a single link but now it's a point to multi-point link multiple users computers want to transmit up to this access point if two users transmit at the same time they'll interfere so multiple access is about ways to make sure that the two users share that link amongst the users efficiently some like FDM and TDM with variations some polling based techniques and some random based techniques but we will not touch upon them in this course