ATM Asynchronous Transfer Mode

ATM is a protocol that transmits data as fixed sized packets . This is the culmination of all the developments in switching and transmission of data in the last twenty years . It was designed to make Broadband-ISDN (B-ISDN) a reality. B-ISDN was created conceptually as just an extension of ISDN so it functions as a communication network that can provide integrated broadband services such as high-speed-data service, video phone, video conferencing, CATV services along with traditional ISDN services such as phone and telex. To have the B-ISDN services mentioned above an interface between the ATM layer and higher layers was necessary . The ATM adaption layer provides this service . Its main purpose is to resolve any disparity between a service required by the user and services available at the ATM layer . It lies between the ATM layer and the higher layers of the B-ISDN protocol reference model (see diagram).

The ATM cell is the basic unit of information transfer in the B-ISDN ATM protocol . The cell is comprised of 53 bytes. Five of the bytes make up the header field and the remaining 48 bytes form the user information field .

ATM cells are transported via virtual channels and indrectly in virtual paths . A virtual channel is a unidirectional pipe . A virtual path is made from of a set of these channels . The ATM cells traveling through the virtual pipes and channels mentioned above will need a physical medium (e.g. fibre optic cable ) to go anywhere . So another layer in the B-ISDN protocol reference model was created , below the ATM layer (see the diagram above) .
The physical layer , its essential purpose is to collect and organize ATM cells sent down from the ATM layer, transport them to the physical medium and also perform the reverse of the process . An ATM network needs certain traffic control capabilities to deal with applications such as video conferencing . Which needs a guarnteed amount of bandwidth is available for the communications link while still using the network as efficently as possible and to cope with potential errors within the network at any time (eg. a problem with the physical layer) .