COMMUNICATION FACILITIES
DATA LINK LAYER PROTOCOLS
As stated in the definition above, the data link layer of protocol encompasses two functions: 1) controlling access to the shared communication medium and 2)
structuring the format of messages in the network. This section describes the two functions and gives examples of their implementation.
Since a communication network (or subnetwork) consists of a single communication medium with many potential users, the data link protocol must provide the rules for arbitrating how the common hardware is used. There is wide variety of approaches to implementing this function, some of which are used only with certain network topologies. They are defined as follows:
1. Time division multiplexing access (TDMA)- This approach is used in bus type network topologies . A bus master transmits a time clock signal to each of the nodes in the network, each of which has a preassigned time slot during which it is allowed to transmit messages. In some implementations, the assignment of time slot is dynamic instead of static. While this is a simple approach, it does not allow nodes to get rapid access to the network, nor does it handle bursty message traffic( where the message comes in spurts, or bursts) very efficiently. Also, it requires the use of a bus master, which can be a single point failure unless it is made redundant (which increases cost and complexity).
2. Polling- This approach can be used in either bus or ring networks. Like TDMA, it requires that a network “master” be used to implement the protocol. In this approach, the master polls each of the nodes in the network in sequence and asks it whether it has any messages to transmit. If the reply is affirmative, the node is granted access to the network for a fixed length of time. If not, the master moves on to next node.
Since time is reserved for nodes that do not need to transmit, this protocol is more efficient than TDMA. However it suffers from the same disadvantages as TDMA:
slow access to the network and need for a redundant master for reliabilty. The polling approach has been used extensively in computer based digital control systems and in certain proprietary distributed control systems.
3. Token passing-This method can be used in either bus or ring networks. In this protocol, a token is a special message that circulates from each node in the network to the next in a prescribed sequence. A node transmits a message containing
information only when it has the token. An advantage of this approach over the previous protocols is that it requires no network master. The access allocation method is predictable and deterministic, and it can be used in both large and small distributed networks. The main disadvantage of this approach is the potential danger that a token may get “dropped” (lost) or that two nodes may think they have the token at the same time. Reliable recovery strategies must be implemented to minimize the chance of these errors causing a problem in the network communication function.
Token passing is one of the access protocols defined by the IEEE 802 local area network (LAN) standard.
4. Carrier sense/Multiple access with collision detection (CSMA/CD) –This approach is used in bus networks. It is analogous to a party line telephone network, in which a caller (in this case a node or device in the network) listens on the line until it is no longer busy. Then the device initiates the call (i.e., the message transmission), while listening at all times for any other device trying to use the line . If another device starts to send a message at the same time, both devices detect this and back off a random length of time before trying again. This approach has a number of advantages. It is simple and inexpensive to implement, it doesn’t require a network master, and it provides nodes or devices with fast access to the network. Its efficiency decreases in geographically large networks, since the larger signal propagation times require a longer wait before the device is sure that no other device is trying to use the network. Also, it is not possible to define an absolute maximum time it can take to gain access to the network, since the access process is not as predictable as in other access protocols (the token-passing approach, for example).
However, queuing analyses and simulations can provide excellent information on the behaviour of CSMA/CD network, so predicting its performance generally is not a problem. The CSMA/CD protocol is used in the Ethernet proprietry communication system, and is specified in the IEEE 802 local area network standard.
5. Ring expansion – This approach is applicable only to ring networks. In this technique, a node wishing to transmit a message monitors the message stream passing through it . When it detects the lull in the message traffic, it inserts its own message, while at the same time buffering (and later retransmitting) any incoming
message. In effect, this method “expands” the ring by one message until the original message or an acknowledgement returns back to the original sender. This protocol is very useful in ring networks, since it does not require a network master;it also permits multiple nodes to transmit messages simultaneously( thereby increasing the effective bandwidth of the communication system) This approach is used in the serial version of the CAMAC (computer-automated measurement and control) system and in certain proprietry communication networks.
Once the control of the communication medium has been established by one of the mechanisms just described, data can be sent from one node to another in the form of a sequence of bits. It is the data link layer of protocol that defines the format in which the bits are arranged to form an intelligible message. It also defines the details of the message transmission and reception operations(including error detection and correction) . Most commercial communication systems used in distributed control implement this level using one of a number of protocols that have become standards in the communication industry. Some of the more popular ones are:
1. BISYNC (binary synchronous communications protocol) – Character oriented protocol developed by International Business Machines (IBM).
2. DDCMP (digital data communications message protocol) – Character oriented protocol developed by the Digital Equipment Corporation (DEC).
3. SDLC (synchronous data link control ) – Bit oriented protocol developed by IBM.
4. HDLC (high level data link control) – Bit oriented protocol standard defined by the Consultative Committee for International Telephony and Telegraphy (CCITT).
5. ADCCP (advanced data communication control procedures) – Bit oriented protocols standard defined by the American National Standard Institute(ANSI).
The first two protocols are similar in that they define node-to-node messages on the basis of characters or bytes (eight bit units of data). In contrast, the last three protocols are bit – oriented; that is the messages are broken into frames in which the individual messages bits have significance. The second group of protocols has largely supplanted the first in current communication systems because of their superior performance and efficient use of the communication medium. The protocols defined
in the second group also have been implemented in off-the shelf chips available from the semiconductor manufacturers, thus simplifying their usage in commercial control systems.