Multiplexing, Framing, and Data Link Basics

such as frequency division multiplexing FDN which is used for analog multiplexing for wireless or wired medium which is the typical medium and then the time division multiplexing TDN which is used for the digital multiplexing for the typical wireless and wired medium and lastly the wave-range division multiplexing WDN which

which is used for analog multiplexing or the typical wired fiber optical medium. The application of the WDM is very limited compared to the TDM and FDM. Today we are going to take a look at the data lake management first. Let me briefly talk about the data mismatch issue.

So here, not all input links may have the same data in general. So some links may be slower or faster than other links. So to overcome this issue, there exist three different ways to solve this question. The first method is multilevel, and the second method is multiple slots, and the last one is first stop. So first of all, the multilevel is used when the data rate of the input links are multiples of each other.

So when the data rate of input links are slower than others, and then the multi-level can be used. Secondly, the multiple slots is when there is the greatest common divisor or CCD, the highest common divisor between the data rates. So the higher mid-rate channels are located more slots per frame,

Then the output frame rate is multiple of each input link. And the last one is pulse stopping, which is used when there is no GCD, "최대gun" between the links. Then the slowest speed link will be brought up to the speed of the other links by bit insertion. These are the three different methods to overcome the data mismatch issue.

Okay, then first of all, let's take a look at the multi-level multi-flexing by giving an example. So this is the example of the multi-level multi-flexing. So if there is the slower data rings compared to the other rings, say 20 kbps, however, the 40 kbps is two times higher than 20 kbps, so we can integrate this 20 kbps into a single thing.

which is 40kbps. Then the data range for four links are matched with each other. Therefore we can integrate the entire four links into a single link using the multiplexing. Secondly, there exists multiple slot multiplexing. So let me give you an example for this case. In this case, the number of the

the one link has the higher data rate than other links. But this link, 50 kbps, is two times greater than the other link's data link. And then we can divide the 50 kbps into 225 kbps. And then we can make the five links into the...

the single ring with the 125 kbps of it. So here suppose that the first data has the blue packet or blue data and the second data is represented by yellow color and third data is represented by blue.

orange color and the last one can be represented by the green color. And then first blue data can be divided by the two data in a single frame. So in each frame there exists two blue data and one yellow data, one orange data and one green data.

So lastly, let's take a look at the first topic. Actually, in this example, the specific one, 46kbps, is much lower than the 50kbps. But this 46kbps does not have any GCD compared to the other links. So that's why we can insert some random lead.

or the meaningless bit into this 46 kbps to make the 50 kbps. This is the process of the pulse subbing. And then finally we can make the 150 kbps by multiplexing. Okay, now let's move on to the next concept. In the multiplexing we need the frame synchronization step. Thank you for the registration.

Here, the one or more synchronization bits or frame bits are included in the beginning of each frame to distinguish between the different frames. It allows the D-Volt flexor to synchronize with the incoming steam so that it can separate the time slots accurately. Then, the synchronization formation consists of one bit per frame alternating between the zero and one.

Let me give you an example here. If we take a look at the synchronization pattern like 1 or 1, and then the 1 is used for the first frame, and 0 can be used for the second frame, and 1 can be used for the third frame, and so on, and so on, and so on, to distinguish between each frame. So anyway, this synchronization pattern should be shared between sender and receiver.

Okay, so then let's take a look at some practical application of the multi-flexing. So this is one of practical application which is digital signal service. So digital signal or DS service or digital hierarchy system can be shown by this. So for example, SK Telecom or KT which is mobile operator.

implement time division multiplexing system through a hierarchy of a digital signal. As I mentioned to you before, there exist a lot of connections or a lot of links in the world, or in South Korea as well. That's why we need to integrate 24 numbers of the link into one single link. And this one single link can be also integrated by the other links.

And then also the integrated link can be integrated as well with the other links and so on and so on and so on. So I can say this system as the hierarchy system. Gaechingjogin system in Korea, right? So that's why in the end, finally the SK Telephone will make the 274.176 Mbps of the speed.

with a single link. But this number doesn't matter to you. So you can... You don't have to remember such kind of number. But you can understand what meaning is. So then using this T1 line for SK Telecom or multi-flexing Telecom lines. So if there exists 24 voice channels

And then each channel has the corresponding signal pattern. This is the original signal pattern between 0 to 4 kHz. And then the original signal will be sampled at 8,000 samples per second using 8 bits per sample. And then we can generate the 8,000 samples per second multiplied by 8 bits. For example,

Then sample sample can be cancelled to generate the 64 kbps per second. So we can see 64 kbps here. And there exist 24 voice channels so that's why we need to multiply the 24 in the 64 kbps. With additional overhead, the same 8 kbps sensor.

And then the integrated some channel can be bundled into a single frame like this. And per second the SK Telecom will generate 8000 frames per second. So that's why we need to multiply 8000 into a single frame BPS to generate the final 1.544 Mbps sensor.

In addition to this telephone system in SK Telecom, there exists another application to use the digital signal service. Another application can be the second generation cellular telephone, which is digital transmission services with the wireless connection. In the 1990s.

The world wide world, even South Korea used the second generation of the cellular network in 1990s. So we can send the text message to others with each other because they use the digital connection with each other. For your information, in the first generation of the cellular network, we can't send the text message to others. because this is the analog transmission

not the digital transmission. So we cannot generate the text message at that time. Here, oh, okay.

So in the second generation of cellular network, each band has 30 kHz of bandwidth. So six users can share the band. So each band consists of six times range. Which means that this second generation of cellular telephone system has the six times greater capacity compared with the first generation of FDM cellular telephone. Okay, so actually this is the end of the lecture seven.

So let me summarize the lecture 7. In this lecture we have addressed the various multiplexing schemes. So what we are wondering about here is that how to share the common resources with the multi-producers. The answer is to use the multiplexing. So let me show you a few of these questions. So the first discussion point is that which multiplexing scheme has advantage or disadvantage compared with the other multiplexing schemes?

in each environment and so on. So actually, let's compare FDM and TDM. As I mentioned before, the FDM was used for the first generation of cellular network, whereas the TDM was used for the second generation of the digital cellular network. So because of this, the FDM system is much more easier.

much easier compared with the TDM system. So in the TDM system, so in the TDM system, it's more complicated to implement compared to the FDM system. However, as I mentioned before, the TDM system is more efficient in terms of transmitting the data compared to the FDM system.

Also, the TDS system can use the digital transmission, not the analog transmission. So there is some trade-off between that. Okay, so, and then the second discussion point is that how to enhance the efficiency of sharing of resources. So here, in terms of the efficiency, we can think about the network capacity.

Suppose that in some environment, this environment will support the 1Mbps of the network capacity, but the design system, the design multiplexing system can support the 80Mbps. Then we can calculate the utilization of this system can be 80%. and then um

Suppose that someone developed the system can support up to 90 Mbps. And then the utilization of such kind of system can be 90%. Then compared to the 80% system, the 90% system can show the more efficient performance compared to the previous system. So our goal is to design such a kind of system to increase the efficiency of the system.

system. Efficiency of the system or efficiency of the resources. And then in addition to the FDM, TDM, or WDM, there exists more advanced multiplexing techniques in the third generation or LTE or 5G and so on, such as OFTN N-O-M-A, something like that.

but it will be as a more advanced communication system courses you can maybe learn such a kind of state of the art technique in the advanced communication system ok so this is the teaser for the next lecture so until now we have run some upper physical layer

Let me remind the OSI layers again. So as I mentioned before, there exist seven OSI layers from the physical layer, the data link layer, the network layer, transport layer, some other layers and application layers. So among these seven layers, OSI seven layers, Until now, we've learned the higher physical layer.

And then from now on, from the next lecture, we're going to take a look at the data link layer, the second layer in the OSSM layer. So always you need to understand, now I'm studying the data link layer or now I'm studying the physical layer and so on. So from the next lecture, we're going to take a look at the concept of the data link layer.

which include the framing, hardware address or the MAC address and the multiprocess protocols such as FDMA, TDMA and CDMA FDMA is frequency division multiprocess, TDMA is time division multiprocess CDMA is cost division multiprocess so in this lecture we already learned about the multiplexing but this is the multiprocess protocol

Actually these two terminologies are very much confused to students because their roles are very similar with each other. But in the perspective of the system designer, they should design the multiplexing system. But in the perspective of the users, they should consider the multiple access protocol.

How I can access to the system, which is called the multiple access. But in the perspective of the system designer, how I can design the system to distribute the integrated rig into some other sources, or how I can integrate the multiple sources to all signaling and so on. So that's the main difference between multiplexing and multiprocessing protocol.

Okay, so this is the end of the lecture seven. So let me give you some questions as a QR code. And please solve this problem for five minutes. And then I'm going to move on to the next lecture.

Thank you.

Did you solve all the review questions? This is lecture A, the first part of the data link layer. In this lecture, I am going to talk about some concepts, hardware address and protocols of the data link layer. Before we get to the details of this lecture, let me remind OSI 7 layer again. So here, let me go.

to think about what's the difference between physical layer and data link layer. In the previous lecture, we already addressed about the differences between physical layer and data link layer. Please remind the exact concept or differences between them.

Basically, these two layers connected with each other in physical connection between two devices. So if there exists device A and device B, these two layers address the exact connection between two devices. These two devices should be physically connected with each other.

using the wired medium or wireless medium. So this is the common point of the physical layer and the link layer. But what's the difference between them? In the physical layer, they will address a single transmission, say, "How I can transmit zero bit? "How I can transmit one bit? "How I can transmit zero bit? "One bit, zero bit, one bit, one one one zero zero zero each other."

and so on and so on and so on. So the physical layer will focus on the specific transmission of a single bit. How I can transmit this single bit over the air or over the cable. On the other hand, the data link layer will transmit frame of the series of which is the same. Which means that the data link layer will address the multiple transmission of the

of the series of beats. That's the main difference between them. The second difference between physical layer and data link layer is that in physical layer, each device should transmit some packet to the other device. This is the one-to-one connection. Okay.

A and B are connected with each other directly one-to-one. On the other hand, the data link layer will address the multiple devices, say the multiple source and single destination or single source multiple destination and so on. So that's why in the data layer, it's very important concept to consider.

situation that how multiple devices should access to this link which is called as a schedule. In a perspective of each user how I can access to this link, while not. Because at the same time these four users cannot access to this link simultaneously. So that's why we need to classify

So at this time slot, the user 1 should be scheduled, or at this time slot 2 should be scheduled, at this time slot 3 should be scheduled, and so on and so on and so on. Which is called as the multiplexes. So such kind of multiplexes concept is the most important concept in the data decrypt. So you need to memorize or you need to understand that fact.

Ok, so then let's move on to today's lecture. Because already we covered the review of the lecture 7, that means keep this page. FDM, TDM, WDM, data array, measurement and so on. So today here, what we will learn in this lecture 8. So in the lecture 8, we are going to build up from the framing. What is the framing?

And then we're going to take a look at the link layer addressing, which is hardware addressing or MAC address. So this MAC address is definitely different from the IP address in network layer. And then we're going to take a look at some multiprocess protocols such as FDMA, TDMA, and CDMA. FDMA means Frequency Division Multiprocess, TDMA means Time Division Multiprocess, and CDMA means Quote Division Multiprocess.

Okay, so let's get started from the framing. So first, let's define the framing. What is the framing? The data recreator will need to pack bits into frames. As I said before, data recreator should address a series of bits. So the series of bits should be packed into frames.

corresponding frame and each frame is distinguishable from another okay so in each frame he adds a sender's address and a destination's address this address is make address and then the frame can have fixed or variable size so there exists a fixed size frame in this fixed size frame there is no need to define boundaries of the frame

the frame because a fender and receiver both aware of the size of the frame. For example, suppose that the size of frame is 5, then 0, 1, 1, 0, 0 bit can be packed as a single frame and 1, 1, 1, 0, 0 can be packed as a single frame. Then both sender and receiver are aware of the size of frame is 5.

and then they don't need to define boundaries of the frame. Boundary means between this frame and this frame. Okay. That means that the size of the frame is the size of the frame. And then let's take a look at the variable size framing. In the variable size framing, there is a need to define the end of the frame and the beginning of the next frame. Okay. Because the size can be variable. So here, in this scenario,

Who defines the end of the frame and the beginning of the frame? Who defines this fact? The sender should define the end of the frame and the beginning of the frame. And then the sender should let the receiver know about this information. So there exist two approaches in variable size framing. The first approach is

character-oriented protocols and the second approach is bit-oriented protocols. First, let's take a look at the character-oriented protocols first. Here, the data to be carried are 8-bit characters from a coding system such as ASCII code. Here, the header will carry the source and destination address or the control information. and trainer will carry

detection or error correction returned to the bits. And then the character-oriented protocol should include the frag which separates one frame from the next frame. An 8-bit or one-byte frag is added at the beginning and the end of the frame to distinguish between the different frames. Okay so let me give you this figure.

here this is a single frame representation. So suppose that the data is received from the upper layer, the network layer, right? And then the character number can be variable between here to here. In addition to the data from upper layer, this framing will include the header, which includes the source and destination address and control information, and the trailer.

which includes error detection or error fraction redundant bits and then at the beginning of the entire frame and at the end of the entire frame we need this break okay? break

Now let's take a look at the byte stopping. The frame could be selected to be any character not used for text communication in the keyboard. The receiver thinks it has reached the end of the frame. However, if there exists this character inside of the data,

has a problem. To solve this problem, a byte stuffing strategy is applied. Byte stuffing is a process of adding one extra byte whenever there is a flag or escape character or ESC in the keyboard in the text. Let me give you an example about this. The data received from the upper layer, say network layer,

looks like this there is the flag and esc and so on okay so inside of the data they use the esc but unfortunately this system use the esc as a flag okay and then the receiver will recognize that oh because there is just esc so this is the end of the data or frame end of the frame okay

it is not correct. So in order to avoid such a kind of the mistake, within the frame, if there exists ESC in the frame, and then the framing will add additional ESC in this ESC. And then if the receiver encounter two executives, the

ESG 메세지, 그리고 그 레시이버는 '이건 프레임이 아니라, 이는 오리지널 데이터입니다' 이는 그 레시이버는 프레임을 수행할 수 있습니다. 좀 더 자세히 설명을 드리면 원래 데이터 안에 프레임으로 ESG가 들어있을 수도 있잖아요.

그런 것들을 막기 위해서 만약에 Sender 측에서 이 프레임을 만들 때 ESG가 있으면 ESG를 한 개 더 두는 거예요. ESG가 두 번 연속으로 둬서 리시버가 만약 ESG를 두 번 연속으로 만나면 그러면 얘는 이제 끝나는 메시지가 아니라 데이터 안에 들어가 있는 메시지라고 판단을 하고 얘를 하나 없애고 원래 데이터를 만드는 거죠. 여기서부터 여기까지가 전체 프레임이라고 생각할 수가 있는 거죠.

Okay, so now let's move on to the beat-oriented protocols. Okay, so here in the beat-oriented protocol, they don't use the character as a flag, rather than they use the beat for the flag, okay? So the data selection of a frame is a sequence of beats to be interpreted by the upper layer as text, graphic, audio, video, and so on. So the MOS protocols will use a special 8-bit pattern

say 0111110 as a delimiter say to define the beginning and the end of the frame. So let me give you this figure. So in this figure at the beginning of the header and at the end of the trailer they will insert this frame.

011111 011111 011111 0

the process of adding one extra zero whenever five consecutive ones will follow zero in the data so that the receiver does not mistake the parent 011111 0 or black. What does that mean? It means that if the original data contains the five consecutive one and then the transmitter will insert

zero by force. And then at the receiver if they encounter five consecutive one, then they recognize that oh this zero is not for the flag but it is included in the original data. By doing so they do not do the mistake at the receiver side. 이것도 비슷한 거죠.

비트 스토핑 할 때도 ESC가 원래 오리지널 데이터에 들어가 있으면 걔를 막기 위해서 ESC 두 개를 묻히는 것처럼 비트 스토핑에서도 원래 6개가 되면은 이게 플래그로 쓰이니까 5개 이후에 강제로 내용을 삽입을 해가지고 얘는 그냥 플래그가 아니라 원래 데이터라고 하는 걸 알려줌으로써 그걸 이제 미스테이크로 피할 수 없는 거예요

Until now we learn about the framing, let's move on to the next role of the data recreator. There exists flow control in data recreator as well. Let's define the flow control. This flow control will coordinate the amount of data that can be sent before receiving an acknowledgement. So the flow of data must not be allowed to overwhelm.

receiver. Overrail means overflow. Say suppose that there exists the Q at the receiver side. This is Q. But because the size of this Q is limited, right? So the size of the Q is limited. So if the sender transmit their data

fast. So the receiver cannot accommodate the transmitted data because of the limited queue size at the receiver. And then the receiver should tell to the sender that your transmission speed is too fast so please slow down the speed of the transmission. That is the flow control. So here the each receiver has a block of memory.

called the buffer. Buffer is the queue at the receiver side. Okay, buffer. Reserved for storing incoming data until they process. Okay. So if the buffer begins to fit up, then the receiver must be able to tell the sender to stop the transmission or the free slowdown the transmission speed until it is once again able to receive the data. So in summary, the flow control refers to a set of procedures

to restrict the amount of data that the sender can send before waiting for the close agent. Now let's move on to the error correction or error detection which are the error control. There exists error detection and error correction in the data link layer. So any time an error is detected in an exchange specified frames are retransmitted.

Meaning that if there exists an error, then the receiver will say to the sender that there is an error between the transmission. So please retransmit the packet. Retransmit the packet. So this process is called as the automatic repeat request or AIQ in the data in the creator. So in summary, the error control in the data in the creator is based on the automatic repeat request, which is the error.

which is the retransmission process of the data. Ok, so now let's take a look at some types of the errors. First of all, why error happened? Think about it. Because we already learned the physical layer in the WSSM layer, right? So that's why we are already aware of the vision, why the error happened during the transmission.

It's because of the thermal noise. In the entire world, we are living in the world, right? So, the entire world, naturally there exists the thermal noise between the communication medium. So, that's why there is a possibility to occur the error between transmission. So, because of interference or thermal noise,

the error can be happened. So there exist two types of errors. One type is single bit error and the other type is burst error. So in single bit error, only one bit of a given data you need is changed from 1 to 0 or 0 to 1. On the other hand, the burst error happens for the multiple types of errors

the transmission. This is the single bit error example and this is the burst error example. So the correction of the single bit error is much easier compared to the burst error. However, if there exists a burst error, it's very difficult to detect or correct the error during the transmission. Okay, now in order to address the error detection and correction, let's take a look at the redundancy concept.

What is the redundancy? So, redundancy of the bit is required to be able to detect or correct errors. Redundant bit means the extra bit in addition to the original data bit. So, in order to detect error or in order to correct the error, we need additional bits in addition to the original data. So, in the error detection process,

The answer is very simple. Error happened? Yes. Or error happened? No. Yes or no. For one single bit. However, about the error correction process, we need to know the specific location of the error. Because we address the digital data 0 or 1, if there exists an error at 0, then the error is 0.

the receiver will recognize that the corrected data should be 1 and if the received data is 1 but there exists an error then the receiver will recognize that the correct one is 0 and so on however, in order to correct the error, we need to know the specific location of the error if we take a look at the cities of the beach

The receiver can recognize that there exists an error and at the same time the receiver should know where the error happened among the cities of the beach. Say the fourth location or sixth location to correct the error from 0 to 1 or 1 to 0. In order to do...

this error detection or error correction process we need to code the original data which is called coding so there exists block coding and the convolutional coding but here in this lecture we're going to take a look at the block coding scheme which is easier than the convolutional coder so in the block coding it processes fixed block level

so do not use the memory during the transmission. On the other hand, convolutional coding will use the memory, which is applied for the continuous stream of processing. Okay, so let's take a look at block coding. Here, block coding will divide our message into blocks. So each block consists of k bits with the data ones.

k bits of the original data. Then it will add r redundant bits to each block to make the length n. So the length of n can be k plus r. So resulting n bit blocks are called as the code words. So the original data is data word and the coded data is code word.

And with k bits, we can create a combination of 2 to the k data words. Why is that? Because we consider the bit 0 or 1. Using the 0 or 1, we can generate 2 to the k combinations for the data words. And with n bits, we can create a combination of 2 to the n code words. That is the combination of the court words.

the n needs. And because n is greater than k, so n should be k+r, so the number of possible code words is definitely larger than the number of possible data words. So 2k is always less than 2^n. And then we can generate the redundant

redundant code word stat in addition to the original data words. The block coding process is the one-to-one process. Each data word is corresponding to each code word. But because the number of code words are greater than the number of data words, there are un-matched code words.

but the core of the error correction or detection arises here suppose that the receiver received some other code word, not the data and then the receiver can recognize that there is an error That is the main concept of

the block code here. Otherwise if the receiver received the such a kind of code word which is connected to the original data one and then the receiver will recognize that oh there is no error between the transmission because there exists the corresponding code word. But if the receiver received the other code word and

And then with 100%, there exists an error.

Okay, so before we move on to the error detection process, today's lecture will finish here because of the time limitation. So from the next lecture, we're gonna take a look at the details of how we can detect error or how we can correct error. Okay? Okay, thank you for today's lecture. - I'm gonna take a look at the picture. - Huh? I'm gonna take a look at the picture.

Thank you.

- 신주도 있대. 그 일호시하고. - 어? 일호시하고. - 그 예은님하고 같이 밥 먹고. - 있다가? - 아 예은이랑 밥 먹으니까. - 네. - 형도 그 무슨 일본 취업 단톡방에 들어갔어요. - 응? - 일본 취업은 오픈 택배나고 있다. - 오늘 처음 얘기한 데이터 레이트베이스먼트였었고요. 데이터 레이시 우리가. - 그 배수의 형태로 묶어가지고 멀티 플렉션 했잖아요. - 그 배수가 안 맞는 경우에 대해서 어떻게 해야 되냐. - 인구 없고. - 그래서 그런 방법이 세 가지가 있는데. 멀티레벨은 자기보다 더. 다른 애들보다 더 데이터 레이시이 낮은 애들 중에. - 최소한.

최대 공략수가 되는 애들을 멀티레벨로 쓰고 그리고 멀티플 슬라트 같은 경우는 다른 애들보다 더 큰 애들 중에 최대 공략수가 되는 애들을 멀티플러 슬라트를 쓰고 그렇지 않은 애들, GCD가 안되는 애들을 펄스 터핑을 쓴다 라고 하면 되고요 그래서 뭐 다른 것들은 보시면 그림보면 다 이해할 것 같고 그러면 펄스 터핑은 어떤 식으로 할 수 있느냐 라고 하면 그냥 의미 없는 예를 들면

0, 0, 0, 0, 0 이런 걸 붙여서 보내는 거예요. 그래서 리시버 측에서 헐 스톱핑을 어느 정도 했다고 하는 걸 알고 있으면 예를 그냥 빼고 나머지 값들만 고려하면 오리지널 46Kbps의 데이터를 받을 수가 있겠죠. 이렇게 하는 거고요. 그 다음에 이제 싱크로나이제이션 같은 경우도 타이밍을 맞추기 위해서 프레임으로 이렇게 사용을 하는 거죠. 그래서 싱크로나이제이션 패턴이 이것들이 샌드하고 리시버 둘 다 알고 있으면 그러면 리시버 측에서 1을 받았으니까 프레임 1이고 0을 받았으니까 프레임 2고 1을 받았으니까 프레임 3고 하는 걸 알겠죠.

이런 식으로 synchronization 하는 거였고요. 그 다음에 실제로 이제 통신사에서 전화망을 쓸 때 이런 식으로 구축을 한다고 하는 거고요. 이게 사실은 뭐 이상적으로는 전체 다 라인이 있으면 예를 들면 우리나라 몇 명이죠? 옛날 30년 전쯤이다 라고 하면 3천만 명이다라고 하면 집이 뭐 한 천만 가구가 있다고 하면은 천만 개나 되는 라인을 하나의 멀티플렉터가 다 처리를 해야 되잖아요. 이상적으로는.

사실 하나의 멀티플렉서가 천만개를 다 처리할 수 없기 때문에 지금 PC라면 가능하려나? 아무튼 그 당시는 PC 한 대당 처리할 수 있는 캠페시티의 제한이 있기 때문에 여러 개를 묶어서 하나로 보내고, 묶어서 하나로 보내고 그런 식으로 했다고 보시면 되겠습니다. 그게 이제 하이라기커란 계층적인 구조를 가지고 있고요. 그래서 우리가 배웠던 PCM이나 이런 것들을 사용해서 디지털 변조를 시킨 다음에 그 다음에 TDM으로, 멀티플렉서 전체 묶어서 하나로 보냈다.

라고 하는 거고요. 여기서도 이제 프레임을 써가지고 하는 거죠. TDM이니까 Time Division Multi-Plex이니까 프레임을 써가지고 하는 거고 각 프레임마다 아까 Synchronization Step이 있었다 라고 하는 거고요. 그리고 디지털이니까 2세대 이동통신, 셀룰로통신부터 디지털로 사용하게 되었고 2세대 이동통신은 1세대 이동통신보다 좀 더 캐페시티가 높고 그리고 할 수 있는 게 좀 더 많았다. 문자메시지 같은 그런 것들이 가능했다.

그래서 여기서 멀티플렉싱에 대해서 배웠었고, 그 다음에 렉쳐 8로 넘어가면서 우리가 하이어 피지컬 레이어에서 데이터 리클레이어로 넘어가는 파트였죠. 그래서 데이터 리클레이어에서 처음에 여러 기능들이 있는데 물리계층하고 데이터 리클 계층과의 차이점을 명확하게 알아야 되고 거기에 대해서 이야기를 했었고, 데이터 리클 계층에서는 가장 중요한 게 여러 유저들이 어떻게 하나의 리클을 공유할 거냐, 멀티플 억세스가 제일 중요한 컨셉이라고 이야기를 했었고,

이것들을 하기 위해서 여러 가지 서포팅을 하는 기능들이 여러 가지 있다는 거였는데 이것들이... 어? 죄송합니다. 여기 왔네. 그냥 말로만 설명할게요. 여러 가지가 있었는데 그것 중에 첫 번째가 프레임. 여러 개 흩어진 비트를 어떻게 하나도 못 봐가지고 보낼 거냐에 대한 프레임이라는 거였고 데이터 링크레이어는 프레임마다 하나씩 이렇게 관리를 하는 거니까 이 프레임이 닫아서 그 다음 프레임으로 넘어갈 때 이게 어디에서 시작을 하고 어디에서 끝나고 하는 것들을 알아야 했다라고 하는 거였고요.

두 번째로 중요한 게 에러 디텍션, 에러 프렉션 넘어가기 전에 플로우 컨트롤도 했죠. 플로우 컨트롤은 sender와 리시버가 있으면 리시버가 지금 버퍼가 있는데 그 버퍼에 이미 데이터가 많이 쌓여있어요. sender가 엄청 많이 보내서 근데 이게 리시버 버퍼가 오버플로우 하는 거예요. 그러면 sender한테 "그만 보내라"라고 말하거나 '좀 천천히 보내라'라고 이야기를 해야, 그래야 오버플로우가 덜 발생하겠죠.

그렇게 얘기를 하는게 Flow Control인거구요. Flow Control은 데이터 링크레이어에도 있지만 나중에 보면 트랜스포트레이어에도 Flow Control이 있거든요. 그 둘의 차이는 제가 항상 강조를 했죠. 데이터 링크레이어와 트랜스포트레이어의 차이는 엄청 큰 차이가 있는거죠. 데이터 링크레이어는 직접적으로 연결된 두 디바이스 간의 트랜스미션이고, 트랜스포트레이어는 논리적으로 연결된 로지컬리 커넥티드 두 디바이스들 간의 커넥션이죠. 감사합니다.

It's like a similar role, flow control, and the flow control is different. Then we talked about error correction and detection. The important thing is that there is a code word and data word. Data word is original data, and the code word is to do error detection and correction. All-Nator's code is made to make other data.

여기서 말하는 코딩은 인코딩이라고 생각하시면 됩니다. 그래서 코딩을 하려면 좀 더 많은 데이터 빛이 필요했다고 더 많은 데이터 빛이 왜 필요하냐 라고 하면 우리가 그걸 가지고 조합을 만들었을 때 1:1로 원래 오리지널 데이터하고 매칭을 하는 것 외에 다른 애들도 같이 고려해가지고 만약에 리시버가 다른 애들을 받았으면 에러가 100% 나은 거잖아요. 그러니까 그걸로부터 에러를 디텍션할 수 있다. 라는 것 때문에 이제 추가적인 키트가 필요한 거였죠.

여기까지 했었고 다음 시간부터는 에러 디텍션과 프로렉션을 어떻게 디테일하게 하는지에 대해서 배울 예정입니다. 여기까지 고맙습니다.