Chapter Eight Network Control and Monitoring: Circuit Switching

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Chapter Eight Network Control and Monitoring

his chapter covers the third layer of ISO/OSI reference model. Network layer is

primarily responsible for routing the packets amongst the network nodes and avoiding network congestion. Upon completion of this module, you should understand: Different switching techniques The functions of network layer Different routing techniques Different congestion controls

he purpose of Network Layer is to provide data transfer from the source to the

destination across various communications media/sub-networks. It has to deal with issues relating to routing, addressing, congestion and flow control.

Switching Techniques
There are FOUR types of switching techniques available: Circuit Switching (Telephone network) Message Switching (Telex network) Packet Switching (X.25 network) Cell switching (ATM network) Figure is the diagram showing different line speed for various equipment to communicate through a packet switching network. In this diagram, a terminal operating at different line speed can access the mainframe through the data network.

Circuit Switching

A dedicated path between the source node and the destination node is set up for the duration of communication session to transfer data. A typical example is the use of FAX machine in telephone network as shown in figure . The characteristics are: Dedicated hardware connection between sender and receiver Real-time network response (limited delays due to equipment and signal propagation) Good For Burst traffic (such as telephone conversation) Relatively large set-up time Charging is usually based on distance/connect time Can you figure out how long it takes to set up a FAX connection?

Longer than setting up a telephone call.

Message Switching
The user message is forwarded across the network one hop at a time. The entire message is transmitted and stored as a whole at each node awaiting for the routing decision to be made. This switching is sometimes called store-and-forward. A similar example is to program the electronic mail to send to different location at different time as shown in figure . The Characteristics are: Greater line efficiency as each inter-nodal like can be shared by many messages from different users. Non-blocking, unless circuit switching a connection must be established prior to delivering data. Caters for different line speed. Delay is relatively high. What is the full name of IMP? Why message switching is less efficient than packet switching?

Packet switching
The user messages are split up into packets of a fixed maximum size to be sent across the network as shown in figure . The whole user messages are reassembled at the destination node. Packet switching is efficient than message switching as pipelining effect can be achieved. This can significantly reduce the transmission time between the sender and receiver. The characteristics are:

Share communications channels within the network Flexible routing (can use the same physical channel to deliver messages to two different nodes within the network.) Limited block size (usually 512 bytes or 1 kbytes) Can be used for interactive access (Response time dependent on network usage) Charging is related to data volume DATAPAK owned by Hong Kong Telecom is a packet switching network using SL-10 manufactured by Northern Telecom.

Cell Switching
Figure is the diagram showing the use of ATM switches to support multi-media traffic. These include voice, video, data etc. ATM network uses a cell of 53 bytes divided into 5byte header and 48-byte payload field. The header contains the information regarding the destination node and error checking and handling while the payload field contains the data. The cell size of header is shown in Figure . Here, it has VPI and VCI both of which refers to the virtual channel identifier and virtual path identifier. Note that ATM network is ideal for optical fibre operating speed up to 8G bps, it is more efficient to divide into a number of physical paths and each path supports different channels. ATM Layer The ATM layer provides for the transparent transfer of fixed size ATM layer Service Data Units (53 bytes) between the physical layer and the adaptation layer. Currently, different headers are added to the data cell between the user and the network and between nodes. The header field for the network and network interface (NNI) and user network interface (UNI) comprises the fields of Generic Flow Control, Virtual Path Identifier, Virtual Channel Identifier, Header Error Control and Cell Loss Priority (CLP). Figure shows the position of ATM cell layer. The cell header error control (HEC) is designed using polynomial convolution to check against multiple transmission errors and will correct a single bit error in the header field. The header error protection scheme in this layer was accepted by the IEEE 802.6 subcommittee and T1S1 in 1989. There is, however, no error checking on cell contents at this layer and no error protection or flow control from an ATM terminal onto the network. As the headers virtual path identifier and virtual channel identifier are modified by the network at each hop, the header must be recomputed and checked by each ATM node to ensure no transmission error.

Virtual Circuits and Datagrams

The switching techniques mentioned previously are related to the properties of subcommunication network. From the network layer points of view, it has to make sure the packets received are in correct order. There are a lot of models existed to help address this problem, among them, two conceptual models namely virtual circuit and datagram dominate.

Virtual Circuits
In this model in figure , the network layer provides the transport layer with a perfect channel and all packets delivered in order. A virtual path or circuit is set up so that packets can pass through over this connection. This connection can be a permanent virtual circuit or switched virtual circuit analog to leased or switched line. The characteristics are: Easier for the user host to use as the data is already in correct sequence. Circuit setup and disconnection is required each time. Sophisticated user may want to do their own error and flow control schemes.

Datagram
In this type of service as shown in figure , each message in the network is not related to any other messages. There is no connection between the sender and receiver. Also the transport layer of the receiver must handle error and flow control on its own. As there is no dedicated path between the sender and receiver, the subnetwork accepts packets (often called datagrams) which contain sufficient addressing information so that the packets can be individually routed within he network. Basically, the user supplies the packets and the subnet transports them to the destination. The packets are routed individually and there are usually no delivery assurance between the sender and receiver. The characteristics of datagram service are given below: Datagrams are individually routed within the subnet No delivery assurance relating to the packets as the packets can be lost, out-ofsequence, contaminated, duplicated etc. Transaction (Sending a short message), connectionless oriented (No need to establish call prior to sending data.)

X.25 uses virtual circuit approach while TCP/IP Internet Protocol uses datagram.

Comparison between Virtual circuit and Datagram

Below is a summary about the Pros and Cons for the Virtual circuit and Datagram Virtual Circuit Datagram Host to host address is always needed in Host to host address is needed in link sending the datagram (Embedded in the setup only datagram itself) Error checking is required by host to Errors is handled by subnetwork. Host will resemble the packet and find out the receive the packets in correct sequence. missing packets. messages may be out of order in the Messages passed in order to the network. communication sub-network Connection setup is initially required prior Connection setup is not required to sending data Is a flexible foundation to support a range Network component failure in path may of higher level protocols which can affect the result provide for additional network services Less overhead in addressing embedded in Overhead in addressing the packet Example is X.25 Level 3 Example is Internet Protocol of (TCP/IP) Can you list a few more discrepancies in terms of packet processing, error handling, endto-end flow control or efficiency?

Example of IP Protocol
TCP/IP is the protocol used in the Internet. It consists of five layers and the network layer consists of four protocols. The commonly used protocol is Internet protocol (IP) which uses Datagram to transmit data. Apart from this, it has three more protocols for signalling and control purpose as follows: Internet Control Message protocol (ICMP) Routing Information Protocol (RIP) Address Resolution Protocol (ARP) Internet Control Message Protocol(ICMP) It is designed to pass control messages between gateways, routers and destination machines. This machine can be sent out in any of the following situations: Use command such as PING to check if another host (machine) is available

When a packet cannot reach its destination When a router can direct a host to send traffic on a shorter route When a host request a time stamp to check the machine time In case the router does not have the buffering capacity to forward a packet Routing Information Protocol (RIP) It is used by gateways to exchange routing information between any two nodes. Address Resolution Protocol (ARP) It is used to collect and distribute the information for mapping tables and is used to find out the corresponding physical address of destination node for an IP address. Internet protocol It is the basic transport mechanism for routing IP packets through gateways, router etc. It uses connectionless approach and is responsible for sending the blocks of data from source machine to destination machine. It does not provide a reliable transmission as it does not require acknowledgement from the receiving machine. It provides checksum on the IP header not the data.

The decision to have the network layer provide an unreliable connectionless service evolved gradually from an earlier reliable connection-oriented service. By putting all the reliability mechanisms into the transport layer, it was possible to have reliable end-to-end (transport-to-transport) connections even when some of the underlying networks were not very dependable. The IP protocol works as follows: The transport layer takes messages and breaks them up into datagrams of up to 64K bytes each. Each datagram is transmitted through the Internet, possibly being fragmented into smaller units depending on the network frame size as it goes. When all the pieces finally get to the destination machine, they are reassembled by the gateway or transport layer to reform the original message based on the fields of IP identification and offset in the IP header as shown in Figure . The IP header has a 20byte fixed part and a variable length as shown below:

32 bits

Version IHL Type of Total length service


Identification

DF MF

Fragment offset

Time to live protocol

Header Checksum

Source address Destination address Options

You might also like