Network Layer

In the preceding tutorials we have studies how the DLL is responsible for moving the frame one end to the other end of the media, the network layer is responsible for the moving the packets from the source to the destination. This mean that the network layer is the lowest layer in the hierarchy to deal with the end-to-end transmission.

For this purpose the network layer should:

  1. Know the topology of the network
  2. Deal with taking an optimum route to destination such that no path is overloaded or idle.
  3. To deal with problems that may arise if the source and destination host are on different networks

Network Layer Design Issues

  1. Store-and-forward packet switching
  2. Services provided to transport layer
  3. Implementation of connectionless service
  4. Implementation of connection oriented services
  1. Store-and-forward Packet switching

As seen in the figure, the major components of the network are the Hosts H1, H2 etc which may be the terminals, PCs etc at home or in an office environment, network of routers i.e. a network of the ISP. Here Home PC directly connect to the ISP equipment which is in the form of a modem connected to the telephone network which contains all the infrastructure of ISP (since telephone company is the ISP). The host at office may have its own router which then connect to the router of ISP. In any case it is the same router as in the ISP having same algorithm running in it, and, hence can be considered as part of ISP. The working procedure is as follows:

  1. The host transmits a packet to the nearest router, either on its LAN or over a point-to-point link to ISP equipment such as modem
  2. The packet is stored in the router until the whole packet has arrived and then compute the checksum to find if there any transmission error.
  3. If the packet is intact and error free, then it is transmitted to the next router and reaching the function defined in above point until the packet reaches the destination.
  1. Services Provided to the Transport Layer

Since routers are the major components of the network layer and it the network/transport layer interface to which the routers have to communicate. So, the network layer should provide the following services to the transport layer:

  1. The services should be independent of the network layer components i.e. routers.
  2. Transport layer should be shielded from the topology of the network layer (routers)
  3. The network addresses made available to the transport layer shoulduse a uniform numbering plan, even across LANs and WANs.

The network engineer have enough freedom in designing the network layer specifications of the services to be provided to the transport layer. There are differing view of the of the internet community and the telephone company community. These viewpoints are:

Internet community believe, even if the network of router is designed properly, there are inherent error and therefore the services like flow and error control be responsibility of the hosts, if the responsibility is given to routers, this would lead to doing the same thing twice. Without error and flow control, the service is like connection-less service. So this is an example of end-to-end example.

The telephone company engineers argue that the network has been for over a century old. So this network should be connection-oriented and reliable. Without this quality of service cannot be achieved especially for audio and video services.

This controversy still exists. Data networks like X.25 in 1970s and its successor frame-relay were connection oriented. However, since inception of ARPANET internet, connection-less network layer have grown. In the 1980s it was thought that ATM which is connection-oriented technology will take overtake the IP technology, but the reverse seems to be happening.Under the cover however the internet is evolving connection-oriented service as QoS become more important. Two example of connection oriented technology are:

  • MPLS - Multi-protocol labeling Switching
  • VLAN - Virtual LAN
  1. Implementation of Connectionless Service

Depending on whether connectionless or connection-oriented service is used two types of organizations are possible:

  1. For connectionless service, datagram networks are possible in which the datagram (same as packets) are injected into the network individually and routed independently OF EACH OTHER. No prior route is established.
  2. For connection-oriented service, Virtual LAN technology is used. A path from thesource router all the way to the destination router must be established before any data packets can be sent. This connection is called a VC (virtual circuit), in analogy with the physical circuits set up by the telephone system, and the network is called a virtual-circuit network

Suppose process-P1 running on host H1 generates some message for the process-P2 running on another host H2host. Host H1 network layer prepares the packets and passes to ISP router A. Router A stores the packets and carry out checksum the releases as per the routing information in its routing table for next router B or router C. Figure above sows four routing table two for router A (initial and new table) and other two in the router C and router E. The routing table (as shown in figure) contain two fileds : destination and the outgoing line. Table shows all the destination from this node and the outgoing line for next hop router. For example A’s initial table shows path to destination F and the outgoing line is to router C from A. Routing table of C shows the destination F through E as outgoing line. So the path for packet 1,2 and 3 is A -> C -> E -> F where the destination host H2 is connected by a Point-to-point link.

How these routing tables are generated and maintained will be dealt with in the next section. The IP which is a connectionless protocol is the basis of internet and the addresses containe in the packet are 32 bit IPv-4 addresses or 128-bit addresses for IPv-6 packets.

  1. Connection-oriented Service:

The connection-oriented service make use of Virtual circuit technology. In this approach a virtual path between the source host Ha and H2 is established and maintained till the communication has finished. If another host H3 also wish to transmit to same destination, then how the circuite is maintained is described below:

The routing entry at each router containsfour field information, these are: incoming host/router, host identifier: outgoing Host, identifier. A’s table thus has H1 host as incoming with identifier-1 and the outgoing to router C and identifier -1. Now, if H3 also wish to transmit to H2, so, A’s table will have H3 as incoming host and chooses connection identifier-1 to establish a virtual route, hence the second entry in A’ table.

Now we have a conflict, though A can distinguish between H1 and H3 identifier but C cannot differentiate. Therefore A assigns a different identifier to its outgoing line, this is shown in outgoing connection detail of A’s table.Similarly, routing table at C and E updates their information for the incoming and outgoing lines. The points to note here are:

  • This process is called label switching.
  • MPLS (Multi-Protocol Label Switching) is an example of a connection-oriented network service used within ISP networks in the Internet.
  • MPLS is often hidden from customers.
  • ISP establishes a long-term connections for large amounts of traffic.
  • It is increasingly being used to help when quality of service is important.
  • Also used with other ISP traffic management tasks.

Questions:

  1. Explain the function of the network layer. Also explain the network layer design issues.
  2. Write short notes on the followings:
  1. Store-and-forward packet switching
  2. Services provided to transport layer
  3. Implementation of connectionless service
  4. Implementation of connection oriented services