The Queen's University of Belfast

Parallel Computer Centre
[Next] [Previous] [Top]
ATM Technical Overview
Why ATM?
- ATM a merging of two well-known techniques
- Packet Switching which produces most flexible network in terms of
- bandwidth requirements
- resource allocation
- Time Division Multiplexing which permits Continuous Bit-stream Oriented (CBO) services such as voice
"The basic idea is that functions must not be repeated in the network several times if the required service can still be guaranteed when these functions are implemented only once at the network's boundary"
ATM Features
- Error & flow control is moved to the network boundary
- No error control on data field within the network
- No flow control on links within the network
- Connection oriented at the lowest level
- All information is transferred in a virtual circuit assigned for the duration of the connection
- Fixed cell (packet) size
- Permits high speed switching nodes
- No constraint on data services (segmentation)
- Efficient cell structure
- 48 bytes of data
- 5 byte header
ATM Cell Structure
- Cell

- Cell Header

ATM Cell Header Fields
- GFC (Generic Flow Control) permits transmissions of several terminals to be multiplexed on the same user interface - user to network traffic
- VPI/VCI (Virtual Path/Circuit Identifiers) field contains 24/28 bits UNI/NNI. VPI field of 8 to 12 bits allows 256 to 4096 virtual paths each of 64,000 VCIs
- PT (Payload Type) identifies payload type for OAM (Operations, Administration and Maintenance). Also reserved bits for explicit congestion control and an ATM Adaptation Layer 5 bit
- CLP (Cell Loss Priority) is used for buffer management
- If set (1) then the cell may be discarded according to network conditions - congestion control
- If not set (0) cell can not be discarded
- HEC (Header Error Control) is used to detect and correct errors in the header
- Single bit error - correction is attempted
- Multiple bit errors - cell discarded
- Calculated using the polynomial x8+x2+x+1
VPI/VCI
- ATM is connection oriented
- Connections are identified by the Virtual Channel Identifier
- A virtual channel is established at connection time and torn down at termination time
- Multiple VCIs may be used for a multi-component service e.g. sound and video over separate VCIs in video-telephony
- ATM switches perform switching on a per VCI basis

Virtual Paths
- Networks will have semi-permanent connections between endpoints
- Such connections may contain many VCs and are known as Virtual Paths
- Virtual Paths are identified by a Virtual Path Identifier

- VCs may not be switched within a Virtual Path
- Switching is normally performed on a per VPI basis

[Next] [Previous] [Top]
All documents are the responsibility of, and copyright, © their authors and do not represent the views of The Parallel Computer Centre, nor of The Queen's University of Belfast.
Maintained by Alan Rea, email A.Rea@qub.ac.uk
Generated with CERN WebMaker