Mobile Data Networks
Stáhnout kompletní materiál zdarma (29,5 kB)
Níže je uveden pouze náhled materiálu. Kliknutím na tlačítko 'Stáhnout soubor' stáhnete kompletní formátovaný materiál ve formátu DOC.
11.Mobile Data Networks with Packet Switching Basic Methods of Packet Communication All messages get distributed already in terminal equipment or in a network-terminating device into smaller parts called packets.The packet is defined as a data unit that is transferred as a unit for purpose of routing in a data network.It usually involves a group of octets (octet is a group of eight bits) with the exactly specified structure that contains in general control information and user data. The packet length is not usually limited, but most often the user data field reach the length of several tens or hundreds octets. Therefore a message is initially converted on the transmitting side intoa packet form(packetising), in which it is transferred through the network so that on receiving side it can be converted again into the original form(depacketising). As soon as the packet arrives to the first switching node, it is stored in memory. Then it is sent to another node etc., up to the node the message receiver is connected to, and back to him. The packet movement through the network depends on one hand on so called packet service type, and on the other hand on methods of routing control used. First service type is based on the fact that every packet contains, except it’s own data and order number in the data message frame, also complete addresses of the message sender and the message receiver. It means, that every such packet can move over the network quite independently against other packets of the same message. The network node directs the packet onto the free route, going via other nodes up to the node the addressee is connected to. Individual packets of the same message can be in principle transferred to the receiver via various routes and they can arrive in random order. This procedure is called datagram service (also service without link) and appropriate packets we call datagrams. And advantage of the datagram service is a flexible response to changes of operational flows and the network topology, and also a relatively short delivery time of the packet. The main disadvantage is the transfer of relatively high number of service information in each packet, a necessity to provide putting packets in the right order only in the receiver’s place and difficult correction in case of an accidental packet loss. This type of packet communications is favourable namely for frequent exchange of relatively short messages in real time. Datagram principle is often used in the area of local data networks and it forms the base of the Internet protocol IP. More often used type of the packet data network service is the virtual circuit service (also the service with link). In this case the full addresses of both receiver and sender of a data message are contained only in service packets that serve to establish and abolish the virtual connection. The active subscriber transmits initially the call request packet (call packet) that penetrates through the network towards the called subscriber. A specific method of routing control in individual switchover nodes applies in the case of the virtual circuit packet service for these call packets only. Individual switchover nodes will then remember incoming and outgoing channels this call packet went through. The message receiver (addressee) responds to the call packet reception with transmitting the so-calledconfirmation packet that travels via the same nodes but in the opposite direction towards the sender. A circuit created this wayvirtually looks to an external observer like a real physical circuit even if it involves only the file of commands in „control computers“ of individual packet nodes. No other data or service packets contain full addresses and they get transferred via this so-called virtual circuit. After the successful transfer of all packets in a given message, suchcommutated virtual circuitgets automatically abolished. However this service enables to createfixed virtual circuits too, that keep the set-up and wait for another message transfer between these two subscribers (see the analogy with a fixed data circuit, e.g. in the telephone network. The virtual circuit service gives a smaller individual packet redundancy, the more efficient control during the data message transfer, but it is more sensitive for transfer parameter changes of transfer chain individual parts during the whole communication process. This service is more favourable principally at stable transmission parameters for longer data message transfers. One can say in general, that the majority of telecommunication systems, allowing packet communication, uses so far the service with link (fixed data networks packet switch-over, integrated services digital networks and others). Few basic methods are used for packet routing: Random routing methodis based upon the assumption, that the network node routes a received packet into quite randomly picked direction (excluding incoming direction, as long as into this node is not connected up already to a connection line to addressee). A shortcoming of this method is a fact that the optimum packet delivery time is not guaranteed. An advantage is the system simplicity and no chance of loosing the packet by sending it to a blocked direction. Fixed (static) routing method is based on establishing the shortest route algorithm, which is in every network node expressed by a fixed – routing table according to which the packets are routed. In case the specified direction is clogged, the packet gets lost. Adaptive routing methodsare based again on establishment of routing tables in individual nodes, whose contents is however changing flexibly upon other criteria, that most often is the shortest packet delivery time. Node adaptive method is based admittedly on routing tables, but gives a possibility to alter them on the basis of certain local information, most often when the nominal direction is clogged. Centralised adaptive method is the most efficient one from