9.1 Local Area Network (LAN) Terminology

Ethernet — A network communications technology using coaxial or twisted-pair cable, originally developed by Intel, Xerox, and Digital. It has a data transmission rate of 10 megabits/second. It is characterized by the use of the CSMA/CD network access method. It is described by the IEEE 802.3 standard. Ethernet is also used as an adjective to describe Ethernet characteristics, such as an Ethernet address, or an Ethernet application.

Fast Ethernet — Ethernet operating at 100 megabits/second over twisted-pair cable or multimode fiber. Fast Ethernet devices support 10 and 100 megabits/second operation over twisted-pair media or 100 megabits/second over multimode fiber.

Gigabit Ethernet — Ethernet operating at 1000 megabits/second over twisted-pair cable or multimode fiber. Gigabit Ethernet devices support 10, 100, and 1000 megabits/second operation over twisted-pair media or 1000 megabits/second over multimode fiber.

FDDI — Fiber Distributed Data Interface, a token-passing network communications technology characterized by use of a dual ring configuration to improve availability upon failure of a node or connection. It has a data transmission rate of 100 megabits/second. It operates over multimode fiber or twisted-pair cable. It is described by the American National Standards Institute (ANSI) standard X3T9.5.

Token Ring — A token-passing network communications technology characterized by a star topology in most implementations. It has a data transmission rate of 4 or 16 megabits/second. It operates over twisted-pair cable. It is described by the IEEE 802.5 standard.

ATM — Asynchronous Transfer Mode, a cell-based network communications technology, where network data is divided into 48-byte chunks and transferred across the network with a 5-byte header that contains addressing and control information. The ATM Forum describes the communications protocol, and specifies how it is to be used to interoperate with Ethernet networks, in the LAN Emulation (LANE) standard. To interoperate with Ethernet, the ATM device hardware transparently breaks transmit packets into 48-byte chunks and adds a 5-byte header and transmits the cells onto the ATM network. On receive, it transparently re-assembles the 48-byte chunks to construct each receive packet.

IEEE — Institute of Electrical and Electronics Engineers, an organization that, among other activities, develops and maintains standards for the computer and electronics industries, including the 802 standards that define local area networking.

ANSI — American National Standards Institute, an organization that develops and maintains standards for the computer and communications industries

802.3 — The IEEE standard for Ethernet network technology, including 802.3u for Fast Ethernet, and 802.3z for Gigabit Ethernet.

802.5 — The IEEE standard for Token Ring network technology.

CSMA/CD — Carrier Sense Multiple Access with Collision Detection, the network access protocol used on half-duplex Ethernet networks to resolve contention between nodes competing for access to the network medium.

NIC — Network Interface Card. Other terms that may be used interchangeably include Adapter, Controller, Device, Card, Port. LAN On Motherboard (LOM) is a variant where the NIC hardware is included on a system board. A multiport adapter consists of multiple adapters on one card, so, for example, a quad Ethernet NIC may be referred to as a 4-port card. A combo adapter consists of multiple adapters, some Ethernet and some storage, SCSI or Fibre Channel.

Bus — Data and control paths that connect the functional units of a computer. In relation to LAN devices, it refers to the hardware interface between the CPU and the I/O devices. Each LAN device connects to a particular type of bus, such as PCI, PCI-X, PCI-Express, EISA, ISA, XMI, TurboChannel, each of which typically has multiple slots to accommodate several I/O devices.

Duplex — A characteristic of a 2-way communication channel that indicates whether the channel can allow transmission in both directions at the same time (full-duplex) or not (half-duplex).

Flow Control — A technique where the flow of data along a communications channel is adjusted to ensure that the receiving side can handle incoming data without loss. Many network applications implement flow control techniques in software. Here, this term refers to the implementation of flow control in hardware independent of the network application or protocol, as specified by the IEEE 802.3x standard. The receiver side hardware sends special packets, called pause frames, that asks the transmitting side to stop transmitting for a certain amount of time. When the receiver has caught up, it sends a pause frame with a zero time to re-enable the transmitter.

Packet — A unit of data transmission on the network, also called frame. It consists of a header, body of data, and a Cyclic Redundancy Check (CRC). The frame may be encapsulated by additional data needed for the particular network technology. Note that LAN Emulation over ATM imposes packet concepts over the underlying cell-based network technology.

Jumbo Frames — Oversize Ethernet packets, where the range of sizes on Ethernet is from 64-1518 bytes, jumbo frames are packets ranging in size from 1519 to 9216 bytes depending on the hardware and software implementation.

Link Up/Down — Network connection state, for Ethernet devices. Most Ethernet devices that connect to twisted-pair cables have the ability to detect if an active link connection exists. When both ends of the network connection can detect a valid connection, the link is considered to be 'up' and the Ethernet device is capable of using the network channel to transmit and receive packets. When the Ethernet device cannot detect a valid connection, the link is considered 'down' and the device will not transmit or receive over the network communications path.

Ring Available/Unavailable — Network connection state, for FDDI, Token Ring, or ATM devices.

Open Systems Interconnect (OSI) Model — Defines the following seven layers in a networking framework:

Port — One end of a communications channel, or the channel itself. When correlated to the OSI Model, port may refer to a communications channel at various layers. At the physical layer, a port is a LAN device, so a quad Ethernet device is said to be a 4-port card. At the data link layer, the LAN drivers allow multiple applications to run on one LAN device. Each application will have opened a port to the LAN driver. At the application layer, an application may allow multiple ports to be opened to it, with the application itself doing the multiplexing of the ports through itself to the underlying network. An example of this would be a network application written to send and receive data over a TCP/IP port.

In this chapter, applications open a port to the LAN driver to communicate over a particular LAN device. In OpenVMS terms, opening a port is done by assigning a channel.

User — Refers to the application that has opened a port to the LAN driver. A LAN device may be described as having a number of different users. Each user would have opened a port to the LAN device. Examples of users are LAT, TCP/IP, DECnet, Clusters (NISCA).

In this chapter, the terms application and user may be used interchangeably.