Layering similarities between computer and postal networks

Computer networks theory of operation has lot of similarities with postal networks. This post briefly describes the similarities between a postal network and a computer network with respect to layering. The Hybrid 5-layer model is considered for comparison purposes.

Computer networks theory of operation has lot of similarities with postal networks. This post briefly describes the similarities between a postal network and a computer network with respect to layering. The Hybrid 5-layer model is considered for comparison purposes.

The diagrams below illustrate the similarities between a postal network and computer network with respect to layering

postalcomputer1

postalcomputer2

Layering Similarities between postal and computer network
Layering Similarities between postal and computer network

As seen in the diagram above, postal and computer networks have similar operations in each layer.

  • For instance, at the Network layer, each home has a unique postal address and each computer has a unique IP address.
  • Also, in postal networks, to uniquely identify an area a ZIP or PIN code is used, whereas in computer networks, to uniquely identify a subnet, the subnetwork identification is used.
  • At the transport layer, TCP is similar to registered post (with guaranteed delivery of information without loss) whereas UDP is similar to ordinary post (with best effort delivery with no guarantee on lossless information delivery) etc.
  • At the datalink and physical layers,  in postal networks letter is carried hop by hop via. vehicles through different paths like road(land), sea, air etc., whereas in computer networks, digital data units are carried hop by hop via. datalink layer protocols like Ethernet, ATM etc. through wired and wireless telecommunication links.

Layer processing at End and Intermediate nodes

This post briefly discusses about the layers that are processed at the end nodes (sending and receiving computers) and at the intermediate nodes (Layer1 Hubs/repeaters, Layer2 switches, Layer3 Routers) when a data unit (packet) is transferred between the end nodes.

This post briefly discusses about the layers that are processed at the end nodes (sending and receiving computers) and at the intermediate nodes (Layer1 Hubs/repeaters, Layer2 switches, Layer3 Routers) when a data unit (packet) is transferred between the end nodes.

The Hybrid layering model is used as the reference to illustrate the processing. To summarize, the layering process happens as follows:

At the end nodes,  processing happens at all the layers. In a hybrid model, this means the application, transport, network, datalink and physical layers are all involved.

At the intermediate nodes,

  •    If the intermediate node is a physical layer device like repeater/hub, then the device   confines its processing to the physical layer (bits). It does not try to interpret the rest of the higher layers information like datalink, network, transport and application headers.
  •     If the intermediate node is a Layer2 switch, then the L2 switch examines and processes  only the datalink and physical layer headers/trailers of a data unit. It does not process the rest of the higher layers information like network, transport and application headers. 
  •     If the intermediate node is a Layer3 router, then the L3 router examines and processes only the Network, datalink and physical layer headers/trailers of a data unit. It does not process the rest of the higher layers information like transport and application headers.

The diagram given below illustrates the layering processes at the end nodes and at the intermediate nodes:

Layer Processing at End and Intermediate Nodes
Layer Processing at End and Intermediate Nodes

As seen from the diagram above,

  • at the sending node, the actual application data is processed and encapsulation happens first at the application layer (application protocol header is added to the actual data that is to be transmitted), then at the transport layer (transport header with L4 protocol, source port, destination port etc. are added), then at the network layer (IP header with source IP address, destination IP address etc. are added), then at the datalink layer (packet is converted into frames by adding a frame header with source and destination MAC address and a frame trailer with checksum) and then at the physical layer (where the frame is treated as a sequence of bits and transmitted through the wireless/wired media to an adjacent neighbor).
  • at the L2 switches, the incoming data unit is processed by the physical and datalink layers. The data unit’s frame header would typically be changed based on the media type of the outgoing physical layer interface through which the data unit has to leave the L2 switch.
  • at the L3 Router, the incoming data unit is processed by the bottom three layers, namely physical, datalink and network layers, so as to send out the packet via. an appropriate outgoing interface. The processing in a router would typically result in updating one or two fields in the network header and also in updating the frame header to match that of the outgoing physical interface.
  • at the receiving end, all the layers are involved, with the decapsulation process happenning in the reverse order to that at the sending end. The physical layer assembles the bits of a data unit and hands it as a frame to the datalink layer. The datalink layer removes the frame header and trailer and hands the packet to the network layer. The network layer processes and removes the packet header and hands the data unit (segment) to the appropriate transport layer protocol (TCP/UDP). The transport layer reads the transport header and gives the segment to the appropriate application layer protocol after removing the tranport header. The application layer finally process the application data after examining and decapsulating the application header.