Previous Table of Contents Next


The payload data is encapsulated by a beginning flag (8 bits) and an ending flag (8 bits) that serve to the distinguish each packet from other packets traveling the same path. The beginning flag also serves as synchronizing bits in order that the packet nodes (intelligent switches) and the receiving terminal equipment synchronize on the rate of transmission. As discussed in Chapter 7, this approach reduces overhead in data transmission and improves the efficiency of transmission. A packet address field of 8 bits (4 bits for the calling DTE and 4 bits for the called DTE) is prepended to the data in order that the various packet nodes might route each packet to the target device. Control data (8-16 bits) includes the packet sequence number so that the target node and terminal equipment are able to identify errored, corrupted, or lost packets, or to resequence the packets should they arrive out of order. Further, the packet sequence number allows the identification of lost or corrupted (errored) data packets in a stream of data. Additionally, the control data includes the number of the virtual circuit (4 bits) and virtual channel (8 bits) over which the data will travel, if a path has been preordained. Finally, error control data is included in the form of a CRC check (16 bits), As discussed in Chapter 7, this level of error control offers a high degree of reliability [8-17].

Packet Switching and Transmission

In a typical scenario, the transmitting terminal, equipped with a modem, dials a telephone number to gain access to a local packet node on a circuit-switched basis through the LEC central office exchange. Alternatively, a short-haul dedicated circuit might connect the user location directly to the packet node. Once the connection to the packet node is established, the transmitting device sends a control packet across the network to establish a data session with the target host computer. The originating node receives that packet, checks for transmission errors, reads the address, and forwards the packet towards the destination, across the most direct and available link. The process is repeated at the next node and so on, until the data reaches the packet node serving the target host. That node sends the packet to the target, which acknowledges its receipt and establishes a session by responding with a control packet to the originating device.

At that point, the originating device begins a stream of data, segmented into packets, with each packet numbered sequentially. Each packet is routed through the network independently, from node to node, in the direction of the target device, based on the most direct and available path at that instant. Should a reasonable route not be available immediately, the packet will be held in queue in buffer storage for a reasonable length of time, until a link becomes available. Once the communication session is complete, a control packet is sent across the network to terminate the data call.

In this scenario in Figure 8.7, each packet may take a different route from transmitter to receiver, in what is known as a datagram mode of transmission. This mode, therefore, requires that the packets be resequenced before being transmitted from the final node to the receiving device, ARPANET and its X.25 successor, pioneered the concepts of locally adaptive routing, network message segmentation, and datagram transmission mode.


Figure 8.7  Packet switched network, supporting transmission in datagram mode.

The internodal links originally were dedicated analog trunks, which later were replaced by digital circuits—usually 56 Kbps DDS circuits. Over time, many of those circuits were replaced with T-carrier facilities. Currently, the facilities generally are high-speed fiber optic in nature, although all variety of analog and digital media are employed in consideration of specific network economics.

Error Control

X.25 also provides for error control through the use of a Cyclic Redundancy Check (CRC). Should an error occur in transmission, the node receiving that packet will recognize the error and correct for it through requesting a retransmission of the corrupted packet. Each node acting on the packet repeats that process. Additionally, lost packets are discovered at the destination node through examination of packet sequence numbers. Retransmissions of lost packets are then requested.

Error control was extremely important in early packet networks, as the facilities consisted of analog twisted-pair. Through a cascading error control process, the integrity of the individual packets and of the packet stream could be improved. However, this process is CPU-intensive, adding to the cost of the packet nodes. Additionally, the process is time-consuming, as each packet must be checked for errors prior to being forwarded to the next node.

Connection-Oriented

It should be noted that X.25 packet switching is a connection-oriented service. A call is set up over a shared path (virtual circuit), over which all packets may travel in support of a logical connection. On the other hand, each packet may travel over a different path, depending on the availability of the various network links at any given moment in time. Each packet of data is separately addressed and, therefore, is capable of working its way through the network independently of the other packets in a stream of packetized data. This characteristic of packet networks is a critical advantage, as the network and all of its elements are shared among large number of users—the cost of transmission across such a network is very low in the context of an appropriate application. Remember X.25 is an interface standard and does not address the internal operation of the network.

Permanent Virtual Circuits (PVCs) and Switched Virtual Circuits (SVCs)

Packet switching supports a large number of conversations over virtual circuits using the same, previously designated circuit or path. While the individual packets of the typical user may travel different paths, large users may be supported by PVCs. In this scenario, all packets will travel the same path between two computers, which path is established by routing instructions programmed in the involved nodes. Alternatively, the network may select the most available and appropriate path on a call-by-call basis using Switched Virtual Circuits (SVCs). Again, all packets in a given session travel the same path [8-19].


Previous Table of Contents Next