Networking II: LAN Protocols

By Ruffin Scott, ACI Sales Engineer

Technical Note 99-41

Technical Notes for Technical Notes for 99-09-September 1999

4th Dimension developers are often called upon to help manage or troubleshoot a client's network or internetwork. It is important to understand the basic concepts of networking so that installation and operation of 4th Dimension products goes smoothly. This technical note covers LAN Protocols including: media access, transmission method, topologies, and devices.

This is the second of a series of technical notes that cover networking. Before reading this technical note, please read the first of the series, Technical Note 99-28, "Networking I: Open System interconnect (OSI) Reference Model."

A LAN is a collection of computers, workstations, printers, and other devices that are connected over a relatively small geographic area. A good example of a LAN would be a small office with three or four computers, a shared printer, and a file server. The advantages of a LAN offer users increased efficiency through such things as email, shared printers and file exchange. The following figure shows three common LAN Protocols and their typologies.

As discussed in the Networking I technical note, the first two OSI Reference Model layers are the Physical and Data Link Layers. The Physical layer represents the medium, i.e. cabling/hardware, by which data is transmitted. The Data Link layer is split into two sublayers: Logical Link Control (LLC) and Media Access Control (MAC). Typically, LAN protocols operate at these two layers of the OSI Reference layers. The following figure maps several of the more popular LAN protocols to the OSI Reference Model layers.

A LAN protocol must have a standard for which it gains access to the network. If there were no rules or standards governing network access, collisions and failures would bring the network to a grinding halt. There are two basic media access methods used by LAN protocols: token passing and carrier sense multiple access collision detect, or CSMA/CD.

In CSMA/CD, the LAN protocol first checks for another node transmitting data. If there is, it waits for a specified time and then rechecks. If the network is free, it sends the data across the network physical medium. Occasionally there will be a collision of packets sent from two different nodes. If this happens, the LAN protocol senses the collision, initiates a backoff algorithm, and resends the data. Each node in the system will have a different backoff algorithm so that the packets will not immediately collide again.

The second major media access method is token passing. Possession of a "token" enables a node on the network to transmit data. Only one node at a time can have control of the token. After the transmission is completed, the sending node passes the token to the next node on the network. If a node that receives the token has data to transmit, it seizes the token and begins to transmit across the network. If the receiving node has nothing to transmit, it immediately passes the token to the next node.

Each LAN protocol has a particular method of transmitting data across the network. The three basic types of transmission methods are: unicast, multicast, and broadcast.

In unicast transmissions, a node sends a single packet to another specific node on the network. Multicast transmissions consist of a single packet that is copied and sent to specific nodes on the network. LAN protocols that use broadcast transmission send a single packet to all the nodes on a network. Nodes that are not intended to receive the packet ignore the transmission.

Nodes on a network are logically configured as LAN typologies. Bus, ring, and star are the three main LAN typologies in use today.

The bus typology is linear. Data is transmitted from on node to the next through the length of the LAN physical medium. This LAN architecture is often called "Daisy Chain".

In ring typologies, nodes are organized logically in the shape of a ring. Ring is similar to bus, except that the end nodes are attached to form a closed loop, and data transmission only flows in one direction.

The star typology is the most common. It involves connected nodes whose end points are attached to networking devices such as switches, hubs and routers. Star typologies are often made up of multiple bus and ring typologies.

There are many LAN protocols in use today. Some of the more common ones are Ethernet, Asynchronous Transfer Mode (ATM), Token Ring and Fiber Distributed Data Interface (FDDI). Ethernet is a common LAN protocol that can be found in most organizations. ATM is gaining popularity. The quality of service and class of service make ATM a very reliable LAN protocol that can handle voice and video more efficiently than other protocols. Token Ring is a well-established protocol originally developed by IBM in the 1970's. FDDI is a high bandwidth protocol that utilizes fiber optic cable as its physical medium of transmission. FDDI is less popular due to it's high cost of implementation.

The information required to cover each protocol is out of the scope of this technical note. Because it has been implemented to such a large degree, Ethernet and its specifications will be discussed.

The term "Ethernet" is like the term "4th Dimension." It can refer to a specific protocol or can be used as a catch-all for a variety of Ethernet implementations. Ethernet has been widely implemented due to its versatility and simplicity. It can be implemented quickly with little knowledge of the protocol. For this reason, most small to medium sized businesses use Ethernet as their LAN protocol. However, the simplicity does not limit it to smaller implementations. Ethernet is also widely implemented in larger organizations and campuses. There are three basic variants: 10 Mbps, 100 Mbps and 1000 Mbps.

The Ethernet specification was developed by Xerox Corporation in the 1970's. Operating at 10 Mbps, it was designed to handle sporadic and occasionally heavy network traffic. Data is transmitted over coaxial cable while media access is achieved through CSMA/CD.

In 1980, a new Ethernet specification, IEEE 802.3, was developed from the original design. Thus began the two separate specifications for Ethernet. Like other LAN protocols, both IEEE 802.3 and Ethernet operate at the lower layers of the OSI Reference model. They both refer to a 10 Mbps implementation. Most often they are implemented in the node's network interface card (NIC). Both specifications use broadcast-based transmission, meaning that frames are sent to every node on the network and each node must decipher if it is the recipient. Each physical layer Ethernet and IEEE 802.3 specification can be described by a three-part naming scheme--LAN speed, signaling method, and physical media type are described by the specification name. The following figure shows an example of the naming scheme.

Although very similar, IEEE 802.3 and Ethernet have some basic differences. Ethernet operates at both layer 1 and layer 2, while IEEE 802.3 services only layer 1 and the channel access portion of layer 2. The following table compares Ethernet and IEEE 802.3.

As seen in the following figure, the two specifications are very similar. Both data frames are the same size, but the IEEE 802.3 frame has a slightly different format.

These are the fields of the data frames:

Preamble - A series of bits that indicate to the receiving node that a frame is coming.

Start-Of-Frame - Synchronizes the frame reception for all nodes on the network. This is accomplished in the Preamble for the Ethernet specification.

Destination and Source Addresses - The source address defines the node that is transmitting the data. The source address can only be unicast, meaning one address. The destination address defines the nodes that should receive the data. The destination address can be either unicast or multicast.

Type (Ethernet) - Defines the upper layer protocol that should handle the data transmission.

Length (IEEE 802.3) - Defines the length, in bytes, of the data to follow.

Data (Ethernet) - The data that is being transmitted. The data is sent to the upper layer protocol defined in the Type field.

Data (IEEE 802.3) - The data that is being transmitted. The upper layer protocol that should handle the data is specified in this field.

Frame Check Sequence (FCS) - Four bytes in length, the FCS contains a Cyclic Redundancy Check, or CRC, of the frame. The sending node calculates the value and places it into the FCS field. The receiving node calculates its own value and compares it with the FCS. If the values are different, an error results.

This technical note presents the basic concepts of LAN protocols and describes the Ethernet LAN protocol. This is the second technical note in a series that address networking issues.

Technical Note #99-28, Networking I: Open System Interconnects (OSI) Reference Model, by Ruffin Scott.