The asynchronous transfer mode makes use of fixed-size cells, consisting of a 5-octet header and a 48-octet information field. There are several advantages to the use of small, fixed-size cells. First, the use of small cells may reduce queuing delay for a high-priority cell, since it waits less if it arrives slightly behind a lower-priority cell that has gained access to a resource (such as the transmitter). Second, it appears that fixed-size cells can be switched more efficiently, which is important for the very high data rates of ATM.
Figure 3a shows the header format at the user/network interface. Figure 3b shows the cell header format internal to the network. Internal to the network, the Generic Flow Control (GFC) field, which performs end-to-end functions, is not retained. Instead, the virtual path identifier field is expanded from 8 to 12 bits. This allows support for an expanded number of VPCs internal to the network, to include those supporting subscribers and those required for network management.
Figure 3 ATM cell format.
The GFC field can be used for control of cell flow at the local user/network interface. The details of its application are for further study. The field could be used to assist the customer in controlling the flow of traffic for different qualities of service. One candidate for the use of this field is a multiple-priority-level indicator to control the flow of information in a service-dependent manner. In any case, the GFC mechanism is used to alleviate short-term overload conditions in the network.
The virtual path identifier and virtual channel identifier fields constitute a routing field for the network. The virtual path identifier indicates a user-to-user or user-to-network virtual path. The virtual channel identifier indicates a user-to-user or user-to-network virtual channel. These identifiers have local significance (as with X.25 and frame relay) and may change as the cell traverses the network.
The Payload Type (PT) field indicates the type of information in the information field. A value of 0 in the first bit indicates user information (that is, information from the next-higher layer). In this case, the second bit indicates whether congestion has been experienced; the third bit, known as the service data unit (SDU) type bit, is a one-bit field that can be used to discriminate two types of ATM SDUs associated with a connection. The term SDU refers to the 48-octet payload of the cell. A value of 1 in the first bit of the PT field indicates that this cell carries network management or maintenance information. This indication allows the insertion of network-management cells onto a user's VCC without impacting the user's data. Thus, the PT field can provide inband control information.
The Cell Loss Priority (CLP) field is used to provide guidance to the network in the event of congestion. A value of 0 indicates a cell of relatively higher priority, which should not be discarded unless no other alternative is available. A value of 1 indicates that this cell is subject to discard within the network. The user might employ this field so that extra information can be inserted into the network, with a CLP of 1, and delivered to the destination if the network is not congested. The network can set this field to 1 for any data cell that is in violation of an agreement between the user and the network concerning traffic parameters. In this case, the switch that does the setting realizes that the cell exceeds the agreed traffic parameters but that the switch is capable of handling the cell. At a later point in the network, if congestion is encountered, this cell has been marked for discard in preference to cells that fall within agreed traffic limits.
The Header Error Control (HEC) field is an 8-bit error code that can be used to correct single-bit errors in the header and to detect double-bit errors. In the case of most existing protocols, such as LAPD and LAPB, the data that serve as input to the error code calculation is in general much longer than the size of the resulting error code. This allows for error detection. In the case of ATM, the input to the calculation is only 32 bits, compared to 8 bits for the code. The fact that the input is relatively short allows the code to be used not only for error detection, but, in some cases, for actual error correction. This is because there is sufficient redundancy in the code to recover from certain error patterns.
The error-protection function provides both recovery from single-bit header errors and a low probability of the delivery of cells with errored headers under bursty error conditions. The error characteristics of fiber-based transmission systems appear to be a mix of single-bit errors and relatively large burst errors. For some transmission systems, the error-correction capability, which is more time-consuming, might not be invoked.