Network Layer

The major function of the network layer in the OSI model is to enable messages to pass between connected networks. Devices, protocols, and programs that run at the network layer are responsible for identifying, selecting, and directing traffics through the internetwork.

Network describes any collection of devices connected together to share information and resources and communicate with each other. Physically, networks are identified by transmission media segments and network addresses.

Network Subnetting

A network is defined by its network address. Network address can have both an internal and an external significance. Externally, a network address can identify a network under one administration. Internally, the network itself can be divided into multiple networks, each with its own network address. This is what we call it subnetting.

Externally this LAN appears as one network managed by the one organization. But internally, the network has multiple subnets. Each subnet cannot communicate each other but with routers all devices allow to communicate each other between networks. Routers connect separate networks, network segments with different addresses.

Network layer addresses

At data link layer, addresses identifying individual physical devices. If you are familiar with a NIC card of a computer, it has a unique MAC address. However, identifying physical devices alone is insufficient to enable internetwork routing. The ability to route between networks depends on identifying the networks. This is done through network addresses, also called logical addresses to differentiate them from the physical addresses used at the data link layer. Logical addresses identify both network the network segment and the device, even though the device has a physical address as well.

Network layer addresses typically contain two components: a network segment address, and a logical device address – both are used to route messages.

The following is an illustration how computer A on network AA wants to send a message to computer D on network CC.

1. First packet arrives at the network layer on the computer A. The device appends the destination addresses (Device D and network CC). The device also appends the source addresses (device A and network AA) so the receiving device can replay it as necessary. The packet now contains D|CC|A|AA (the destination and source addresses)

2. At the data link layer the device adds its own device address (30) and the next hop device (router B) address on the path (40). The packet now contains 30|40|D|CC|A|AA

3. The packet arrives at router B and the router B examines the data link destination address of the packet and recognizes that it addresses to it, the router drops off the data link addresses and examines the destination address at the network layer. It recognizes that the destination device is on the network address that is not attach to it.

4. The router repacked the destination data link address and adds its own device address (40) and the next hop device address router C (50). The packet now contains 40|50|D|CC|A|AA. And then the packet travels to router C.

5. Router C receives the packet and drop the device addresses (40|50), and examine the destination network address and recognizes the packet is to device D and repacked the packet in the data link layer, adds its own device address (50) and the destination device address D (60). The packet now contains 50|60|D|CC|A|AA and then the packet travels to computer D.

6. Computer D receives the packet and removes the data link layer addresses, the original resource address (A|AA) and the destination address (D|CC) are contained in the network layer.

The following points should be understood regarding how the packet is addressed as it travels through an internetwork.

• Both data link physical addresses and network logical addresses are used
• The network address contains both a logical network address and a logical device address.
• Both the source and destination network and data link addresses are typically contained in the packet
• The data link destination address indicates the physical address of the next hop on the route
• Data link addresses in the packet change as the packet is delivered from hop to hop
• The network destination addresses indicate the address of the final destination device
• Network addresses remain constant as the packet is delivered from hop to hop

The network address indicates both the logical network and logical device address.

Routing protocols

Network layer protocols are software processes that perform internetwork routing functions. A Cisco router can run several network layer protocols at once where each protocol runs independently of the others. A routing protocol is the actual network layer protocol that performs internetwork routing functions. Routing protocols learn and share information about internetworks, and make decisions about which parts should be used. Routing protocols include the following:

1. Routing information protocols (RIP)
2. Interior Gateway Routing Protocols (IGRP)
3. Open shortest path first (OSPF)
4. Netware link service protocol (NLSP)

Routed protocol

A routed protocol is an upper-layer protocol that can be routed through an internetwork. For a protocol to be routable, it must contain network layer addressing information. Routed protocols are routed through an internetwork by routing protocols including: IP; IPX; AppleTalk; DECNet.

Nonroutable protocol

Not every protocol can be routed, it’s a nonroutable protocol which:

1. Does not support network layer data; it does not contain logical addresses.
2. Use static – predefined routes that cannot be updated.

For examples:

1. NetBIOS (network basic input/output system)
2. NetBEUI (NetBIOS Extended user interface
3. LAT (Local Area Transport)

Switching

Besides routing, another function of the Network layer is switching.

1. The ability of a router to receive data on one port from one network and send it out another port to different network.
2. Moving data between interconnected networks to reach the final destination

There are two methods how the packets travel through a complex network; circuit switching and packet switching.

Circuit switching has the following characteristics:

1. The path is determined from start to finish
2. The path must be established before communication begins
3. It is like setting-up a call, and uses the same technology used as phone network
4. All packets take the same route
5. The route is dedicated to conversation, and must be opened and close each time
6. It uses a switched virtual circuit (SVC) between devices

WAN connection that use this type of circuit switched is ISDN circuit-switched network.

Packet switching has the following characteristic

1. The path is determined as communication takes place
2. Establishing a path or connection is not necessary before beginning to send data
3. Packet switching is always on, and does not have to be reestablished for each session
4. Each packet can take different routes
5. Each route can be used by other devices at the same time
6. It uses a permanent virtual circuit (PVC) between devices

Example of this packet switching is frame relay technology in WAN connection.

see also:

• WAN services technologies
• Local area network topology
• routing protocols guide