Overview

Asynchronous Transfer Mode (ATM) is a packet‑switching technology that moves information in small, fixed‑size units called cells. Unlike variable‑length packets used in many data networks, every ATM cell is 53 bytes long, containing a 5‑byte header and a 48‑byte payload. ATM is connection‑oriented: before sending user data, a virtual connection is established through the network and cells are forwarded along that connection. The design emphasizes predictable latency and controlled bandwidth, making ATM well suited to carrying integrated voice, video, and data streams where Quality of Service (QoS) guarantees are important.

Structure and operation

ATM separates forwarding from payload content. Each cell header includes identifiers (commonly called virtual path and virtual channel identifiers) used to route cells through switches without examining their payload. Because the cells are small and of uniform size, hardware switching can be simpler and delay variation (jitter) is reduced for real‑time traffic. ATM operates over several layers, including a physical layer that defines electrical or optical transmission and an adaptation layer that fragments higher‑level packets into 48‑byte payloads (and reassembles them at the destination).

Service categories and QoS

One of ATM’s main advantages is explicit support for different service types with defined performance characteristics. Common categories include:

  • Constant Bit Rate (CBR): fixed bandwidth and timing for real‑time streams such as uncompressed voice or leased lines.
  • Variable Bit Rate (VBR): for bursty multimedia traffic that needs some timing guarantees.
  • Available Bit Rate (ABR): adapts to network congestion using feedback to allocate throughput fairly.
  • Unspecified Bit Rate (UBR): best‑effort delivery with no guaranteed bandwidth.

These categories allow service providers to offer contractual QoS levels specifying maximum delay, jitter, and minimum bandwidth, which was harder to achieve with early Ethernet and other simple LAN technologies. For further technical background and standards material see references on ATM.

History and deployment

ATM emerged during the late 1980s and early 1990s as part of efforts to create a broadband integrated services network that could carry voice, video and data on a common infrastructure. It was standardized and promoted by international bodies and adopted by many carriers as a backbone switching technology. ATM was also used within some LANs and became a component of DSL (ADSL) systems where ATM virtual circuits carried user traffic from customer premises to service provider equipment.

Uses, strengths and decline

ATM’s strengths were its support for strict QoS, predictable performance, and a unified model for different traffic types. These properties made it attractive to telecommunications carriers and for early multimedia services. However, ATM is complex and introduced significant protocol overhead. Over time, the rapid evolution of IP networks, improvements in Ethernet, and technologies such as MPLS and IP QoS mechanisms reduced the need for ATM. Today it remains in some legacy carrier networks and specialized contexts, while most new deployments favor IP/Ethernet‑based solutions. For a direct comparison with common LAN technology see Ethernet comparisons.

Notable distinctions

Key points that distinguish ATM from many packet networks are its fixed cell size, connection‑oriented switching using virtual circuits, and built‑in service classes for QoS. These design choices trade increased protocol complexity and header overhead for the ability to provide predictable latency and controlled bandwidth—features that influenced later networking approaches even as ATM itself became less dominant.