Medium access control

From Citizendium
Jump to navigation Jump to search
This article is developing and not approved.
Main Article
Related Articles  [?]
Bibliography  [?]
External Links  [?]
Citable Version  [?]
This editable Main Article is under development and subject to a disclaimer.

All practical communications devices share a medium, even a point-to-point connection, and medium access control (MAC) techniques are the means by which sharing is possible. Applications include local area networks both wired and wireless, longer-ranged wireless networks such as cellular telephony and military tactical radio communications, and long-haul resources such as satellites and optical networks. It is not limited to data networks.

In the Open Systems Interconnection Reference Model, medium access control is at the data link layer, but there are many communications systems in which the medium access has to be arbitrated at the physical layer.

Medium access techniques for local area networks have found it useful to have a standard way to define hardware-unique addresses; see IEEE 802 address‎. Some wireless networks also use hardware-unique addresses, but other networks use software-negotiated time slots or codes. Yet other free-space networks use unique frequencies; the term wavelength, or lambda, λ, is preferred for optical networks.

Methods include reservation, contention, and token-based. Unique identifiers simplify the assignment of computers to resources (e.g., time slots). A reservation method reserves some resource, such as a time slot relative to the start of a cycle of communications, a frequency within a shared part of the electromagnetic spectrum, or a physical connection within an interconnection fabric.

Contention methods have the potential of multiple devices trying to use the medium at once, but have mechanisms of avoiding or resolving such collisions. This is the fundamental Ethernet or IEEE 802.3 technique, with a variant called carrier sense multiple access with collision detection (CSMA/CD).

Token-based methods use the shared channel, or possibly a separate control channel, to send a message called the token; only the station with the token can have access to the medium. Control of the token may be from a central station, or it may be distributed. Central control is also called polling. Distributed token control is used in the obsolescent IEEE 802.5 (Token Ring) and Fiber Distributed Data Interface (FDDI) methods. A somewhat more recent niche system, still due to be replaced, is the military JTIDS Link 11, typically with a mission controller acting as central controller to a group of aircraft.

Reservation methods

Reservation methods are more common in communications over a wider area, be it the area of coverage of a cell in cellular telephony or a shared electromagnetic spectrum for tactical or air traffic control communications. The resource reserved may be a frequency, a time slot within a cycle made up from a sequence of time slots, or a particular signal pattern that is recognized within a mixture of signals. Reservation may be preassigned (i.e., static) or dynamic.

The Synchronous Optical Network (SONET) and Synchronous Digital Hierarchy (SDH) use statically reserved slots, either manually defined or defined with the assistance of a non-real-time provisioning tool. Self-organizing networks, such as the Automated Identification System (AIS) used in maritime safety, negotiate for time slots.

Contention methods

The most common contention method is carrier sense multiple access with collision detection, which is the basic method of the original Ethernet, and the subsequent IEEE 802.3 standardized version of CSMA/CD.