Introduction to ATM
Asynchronous Transfer Mode: goals, reference models, and comparison with other technologies.
What is Asynchronous Transfer Mode (ATM)?
Asynchronous Transfer Mode (ATM) is a high-speed, cell-based switching and multiplexing technology designed to be a single, unified network for a wide variety of traffic types, including voice, data, and video. Unlike the variable-length packets used in the internet (IP), ATM uses small, fixed-size packets called cells.
The "asynchronous" in its name refers to the fact that cells are allocated on demand, as needed by the traffic source, rather than being placed in a rigid, pre-assigned time slot like in synchronous systems (such as ). This on-demand allocation makes it efficient for the "bursty" nature of computer data.
Analogy: The Postal System
Imagine a postal system. Sending data using IP is like sending parcels of various sizes. Sending data using ATM is like sending everything in standard-sized, small postcards. The fixed size of the postcards makes sorting and handling at the post office (the network switch) extremely fast and predictable, as all processing can be done in hardware. This was the key idea behind ATM: achieve very high speeds by simplifying the switching process.
A Historical Perspective: The Vision of B-ISDN
ATM was not developed in a vacuum. It was the chosen technology for an ambitious vision from the 1980s and 90s called B-ISDN (Broadband Integrated Services Digital Network). The goal was to create a single, global network that could seamlessly handle every type of communication service.
- Standardization: The technology was initially defined by the CCITT (now ) and ANSI.
- ATM Forum: In 1991, a consortium of equipment manufacturers and service providers formed the ATM Forum to accelerate the development and deployment of ATM standards, ensuring interoperability.
- Scope: ATM was designed to be scalable, operating across Local Area Networks (LANs), Metropolitan Area Networks (MANs), and Wide Area Networks (WANs).
Although B-ISDN never fully materialized as envisioned, ATM technology found significant use in the core of telephone and internet provider networks for many years, especially in mobile network backhauls.
The Heart of ATM: The 53-Byte Cell
The single most defining feature of ATM is its fixed-size cell. Every piece of information, regardless of its source, is broken down into these small units.
Cell Structure: A Tale of Compromise
An ATM cell consists of:
- 5-Byte Header: Contains addressing and control information.
- 48-Byte Payload: Carries the actual user data.
Why 48+5?
The choice of a 48-byte payload was a famous engineering and political compromise. Voice carriers in the US preferred a small payload (like 32 bytes) to minimize the delay when filling a cell with voice samples, which is critical for call quality. Data carriers in Europe preferred a larger payload (like 64 bytes) for better efficiency (less overhead). The 48-byte payload was the middle ground they agreed upon. The 5-byte header was deemed sufficient for addressing and control.
Key Header Fields
- VPI/VCI (Virtual Path/Channel Identifier): This is the "address" of the cell. ATM is connection-oriented, and these numbers identify the the cell belongs to.
- PT (Payload Type): Indicates what type of data is in the payload (e.g., user data or management information for the network itself).
- CLP (Cell Loss Priority): A single bit that indicates whether the cell can be discarded if the network encounters congestion. A value of '1' means the cell is lower priority and can be dropped first.
- HEC (Header Error Control): A checksum calculated only on the 5-byte header. This allows switches to quickly detect and often correct header errors, ensuring the cell is not misrouted.