Multiplexing is the process of data transmission, where data character by character from different devices is combined into single data stream, which is transmitted over a single communication channels. This process aims at enhancing the data transmission capacity of the communication channels. For this purpose, special devices know as multiplexers are used, which produce the multiplexing. The same devices known as multiplexers are used, which produce the multiplexing. The same device is used at both the transmitting and the receiving and, as it is used to combine the characters from different devices at transmitting end and to separate then at receiving end. Using multiplexers, one can transmit more data, using a Single Communication Channel. The multiplexers are quite less expensive and are very efficient.
Multiplexers are capable of multiplexing and de-multiplexing is also commonly known as MUX.
Multiplexers use any of the following schemes for communicating concurrently:
In this scheme of multiplexing, the frequency spectrum is divided among various logical channels, with each user getting an exclusive possession of a frequency band. For example, the radio broadcasting channel AM uses a spectrum of about 1 MHZ frequency, different frequencies are allocated to different logical channels, so that the broadcasting takes place with each frequency operating in a portion of a spectrum, with the inter channel separation, sufficient for preventing interference of frequencies.
This multiplexing scheme usually uses a broad band, enabling the frequency distribution/division.
Time Division Multiplexing
In this scheme of multiplexing, different devices get the entire bandwidth of the channel, for a small time slice and during that time, the data transmission takes place for that device. Usually Round Robin Algorithm is implemented for determining the device, to which the time slice is to be allocated. The time slots (slices) are fixed and each device gets its time slice, within which the data transmission takes place, else it has to wait for next time slot. The order of allocation of time slice is always. For example, let us consider, the same radio broadcasting example.
Here, we will assure that the individual station (logical channel), has two logical sub-channels, for two different programmers, for example songs and story. For broadcasting them, on the same frequency, they are released at alternate times, which are constant, first may be burst of songs, then story and so on.Thus, two different data pieces are transmitted using same frequency, same channel with time gaps.
This multiplexing scheme usually uses a base band for transmission.
If the Conventional Time Division Multiplexing Scheme is used, there are quiet many chances, of the bandwidth getting wasted, as there may be some slots of time, which may not be used by the allocated device. To remove this short coming, another scheme of multiplexing is used, which is known as Statistical Time Division Multiplexing (stat TDM). In this scheme, the time slots are allocated dynamically to the active devices using either “First Come First Serve” basis or “priority” basis. The data transmitted in each slot is tracked, so as to keep track of the device which has transmitted it. This is quite an effective scheme of the data multiplexing.
For networking, which use telephone cables, which were set up to be used for human-human communication, not for computer to computer communication, these multiplexing schemes are in appropriate for data traffic. We need to have a different kind of switching altogether. The following section discusses, the two basic techniques, which are quite often used in Computer-Computer Data Transmission.
Circuit Switching: Whenever one uses a telephone, and makes a call (Either human being or a computer can make a call), the switching element, within the telephone system seeks out a physical “copper” route all the way from the calling telephone to the called telephone.This technique is known as circuit switching. The vital property of this technique as that we need to set up an end to end path before transmitting any data.
Let us for example take an office, where they have four incoming lines and four outgoing lines. These lines are connected through a switching office as illustrated in following figure:
In above figure all the six rectangular boxes represent switching office has four incoming lines and four outgoing lines. Whenever a call passes through a switching office, a physical connection is established between the lines on which the call came in to one of the output lines, so that the call can be received.
In past times, this connection was made by the operator manually, now it is automated. The logic here is that, once the path is set up, it continues till the call is finished. This way enables that there is no congestion, as once the call is put through; it will be concerned till it ends.
Here, instead of using physical copper path, between sender and receiver, whenever the sender sends data, it is stored in first switching office, and then it is forwarded later, a big block at a time. Every block is received in totality, inspected for errors and then re transmitted, this type of switching is implemented by a network it is called “Store and Forward Network”. The problem here is that, there is no limit for the block size, i.e. the switching offices should have disk to buffer very long blocks, which may make it that one line is totally blocked for many minutes which makes it useless for interactive traffic. Thus message switching may not be advisable alternative.
This technique of switching overcomes, the short coming of message switching which it does by fixing the size of the block. This allows the data packets to be buffering in switching offices using main memory, instead of the disk. By ensuring, that no user can block any transmission line for more than a few milliseconds, these packet switching networks are suited for handling interactive traffic. This method of switching reducing delay and improves through put. For these reasons, this switching mechanism is most preferred in computer networks.