What Is the LANMAN Committee?

What Is the LAN/MAN Committee?

This chapter discusses the important standards as defined by the Local Area Network/Metropolitan Area Network (LAN/MAN) Standards Committee (LMSC). This committee consists of various working groups, each of which is devoted to a particular technology. For example, one committee was responsible for the basic Ethernet standard, and another was in charge of developing the standard for the first Token-Ring networks. As if this division were not enough, there are other working groups that further divide the responsibility for a particular technology. For example, there are several working groups actively working on wireless standards at this time.

The focus of the 802.* working groups is trained at the two lowest layers of the OSI Reference Modelthe physical and data link layersand some aspects of higher layers relating to network management. The LMSC also coordinates its activities with other standards bodies, such as the International Organization for Standardization (ISO), and other standards groups outside of the U.S.

The OSI Reference Model is discussed in Appendix A, "Overview of the OSI Seven-Layer Networking Reference Model."

The following sections cover important standards, some of which are mentioned for historical purposes, as well as newer standards that are still under development.

IEEE 802: Overview and Architecture

The IEEE 802 standards document sets forth the groundwork for the others that follow. It includes short descriptions of the types of networks that were to be considered by the committee, and a reference model for their development. The terms LAN and MAN are defined in this document:

• LAN A local area network is typically part of a single organization. The LAN is a peer-to-peer network that allows nodes on the network to communicate with one or more nodes on the same LAN. In the original specification, the bus topology was used, though the document has been updated to include switching and other modern LAN techniques.

• MAN A metropolitan area network is a collection of interconnected LANs that span a larger geographical area, such as a campus network or a city. MANs can also be owned and maintained by a single entity, but are more likely offered as services to multiple clients in order to provide a high-bandwidth connection to multiple client networks.

It is important to know that the OSI model is just that: a model. Very few protocols actually adhere strictly to this model. For example, TCP/IP was being developed before the OSI model was created, and thus the TCP/IP model differs from the OSI model.

Today, the OSI Data Link layer, for most network protocols, is usually divided into two sublayers. The Logical Link Control (LLC) and Media Access Control (MAC) sublayers are the result of this separation of the single OSI Data Link layer into two components (although in some standards the functionality provided by these sublayers becomes a little blurred).

The Media Access Control Sublayer

The Media Access Control sublayer provides a service to the LLC sublayers to get the data packets delivered to the destination node. At this sublayer the data to be transmitted are referred to as "frames." The MAC sublayer creates the frames to be transmitted on the physical network media, and includes some error checking to allow the receiving node to check the integrity of the data frame. The MAC address is an important concept when used in modern Ethernet networks. As you will learn in Chapter 24, "Overview of the TCP/IP Protocol Suite," the IP protocol enables the hierarchical address space that is used on WANs and the Internet. At the LAN level a flat address space is used for communications between nodes on the same LAN.

There is an important distinction to be made between MAC and IP addresses. MAC addresses are simply addresses that are burned into a network card or other hardware by the manufacturer. Part of the address represents the manufacturer, and the remaining portion of the address is assigned in a serial fashion to each network card that the manufacturer produces. Because there is no "organization" using MAC addresses, this addressing technique is known as a flat address space. IP addresses, however, are divided into two parts: a network address and a host address. Thus, IP addresses (which are described in greater detail in Chapter 24) allow for routing and other functions. MAC addresses, however, are generally used only on the local LAN, which consists of a much smaller number of network nodes.

On a LAN, which is used when only a handful of network devices are connected (up to a few hundred or even a few thousand), a MAC address is sufficient for getting the data delivered to the appropriate node. The data frame is broadcast to all nodesunless a switch is used. The switch learns the MAC addresses for attached devices and eliminates the "broadcast." When a node recognizes that a frame contains its MAC address, it responds by grabbing the frame, and from that point onward the MAC address is used to communicate between the two nodes. Many network devices have the capability of maintaining a table of MAC addresses to IP addresses. This information stays in the table for a short period in case additional communications take place. Eventually the entry is aged out of the table as new address pairs are added.

Tip

Distinguishing between MAC and IP addresses is simple. MAC addresses are used for any device on a LAN, whether it be a computer, printer, or router for exchanging data. IP addresses are used to exchange data between LANs. After the IP protocol has delivered the data to a router connected to a LAN, the MAC address is used to exchange data between the router and the destination device. And all communications on a LAN use the MAC address for communications.

The Physical Layer

At the bottom of the OSI model, you will find the physical components that perform the functions needed to transmit the data passed down from higher layers. These include the network adapter card and the network mediacopper wire or fiber-optic cables, for example.

It is easy to understand the packet or frame that a higher-level protocol constructs, which you will learn about in following chapters. The actual signaling mechanism at the physical layer, however, can be different depending on the network media, and this mechanism can be different depending on the transmission protocol used for the network media. For example, 10BASE-T and 100BASE-T and Gigabit Ethernet transmissions across a copper network media use different methods to send bits of data across the network. Similarly, the mechanisms used to send data across a fiber-optic cable will depend on the protocol used.

Consider the process of sending single bits of information across a network cable. You might expect that each bit is represented by some kind of state change (electrical or photo-optic) on the cable. This is not always the case. Instead, at the physical level, various techniques are used. It may be as simple as varying the voltage on a copper wire or as involved as using statistical methods to vary the voltage and frequency of the signal. A good example of the physical method for transmitting bytes of information across a network is Fibre Channel, in which 10 bits are used to send 8 bits of data. In this case, for each byte of information there can be either one or two possible bit combinations used to transmit the same byte! This is because Fibre Channel, usually implemented using fiber-optic cables, tries to maintain a "running disparity" on the network media. This is for several reasons, which are covered in detail in Chapter 11, "Network Attached Storage and Storage Area Networks." Another example of this, Manchester Encoding, is described in Chapter 13, "Ethernet: The Universal Standard."

Other Physical Layer Components

The physical layer standards established by the 802.* committee involve many other concepts, such as bridges and the protocols associated with these devices, like the Spanning Tree Algorithm. After you get past the LAN/MAN specifications, the routing terrain is the next step. Because routing implies connecting various physical LANs, routing is beyond the scope of the LAN/MAN committee.

The most widely used LAN technology in use today is Ethernet. And this protocol and its associated technologies have been extended over time to allow for end-to-end Ethernet connections across MANs. The various 802.* standards for Ethernet, Token-Ring, and other networking technologies are explored in the following sections.

IEEE 802.1: Bridging and Management

The 802.1 standards concern bridging. Bridging involves connecting two or more networks using an intermediary network device that serves one or two purposes. First, a bridge can be used to connect several LAN segments so that traffic between nodes on the network is confined to that LAN segment. Second, a bridge can be used to translate between different protocols.

Since the original publication of this standard, there have been several other standards that relate to the original 802.1. For example, IEEE 802.1Q discusses using bridges to create a virtual LAN. IEEE 802.1x provides for the use of a MAC bridge to create a virtual LAN. Both of these are discussed in further detail in Chapter 9, "Virtual LANs."

IEEE 802.2: Logical Link Control

As described earlier, the IEEE specifications divide the Data Link layer into two parts: the Logical Link Control (LLC) sublayer and the Media Access Control sublayer. The LLC sublayer provides services to the Network layer in the OSI model, independent of the underlying MAC sublayer.

This sublayer provides for three kinds of service. Type 1 defines an unacknowledged connectionlessmode link. Type 2 defines a connection-mode link. Type 3 defines an acknowledged connectionlessmode. It isn't important to understand exactly what these types of links really mean at this point. In Chapter 25, you will learn how TCP provides a connection-oriented link, whereas other protocols, such as UDP, provide a connectionless link. The important thing to remember here is that the IEEE 802 documentation defines the features and boundaries of these types of connections.

Type 1 services do not need any "setup" before communications can begin. This type of service provides no mechanisms for flow control or error detection.

Type 2 services dictate that a logical link must be established before data communications can begin. An example of this is the TCP protocol, which uses a "handshake" exchange of network packets to set up the link before the actual data exchange can begin. This type of service does provide for error detection and flow control.

Type 3 services provide for a connectionless link, in which no setup is required. However, acknowledgments are used to ensure that network packets are received intact and in the order in which they are sent.

IEEE 802.3: CSMA/CD Access Method

Until the development of full-duplex switches, the method used by nodes on an Ethernet network to gain access to the shared network media was called Carrier Sense Multiple Access/Collision Detect, or CSMA/CD. This simply means that before attempting to send data on a shared LAN segment, the computer (or other networked device) would first listen (carrier sense) to determine whether another device is already transmitting data (multiple access). If not, the node could begin to transmit data onto the network. If more than one node senses that the network media is not being used and both nodes begin to transmit data at about the same time, a "collision" occurs (collision detect). In that case, each node will stop transmitting for a semi-random interval before attempting to transmit again.

For small LANs this technique provides an inexpensive method to allow computers to use a shared network media. As networks have grown in size, switches have replaced hubs in most networks. Switches remove the "collision domain" so that communications take place between just the switch port and the computer connected to that port. If the switch operates in half-duplex mode, a collision can occur if the switch and the attached computer both try to transmit data at the same time. In full-duplex mode, which is the most widely used mode today, the switch port and the attached computer do not share the same wires, but instead each has a dedicated set of wires so that the switch can send data while the attached computer is sending data to the switch.

It is important to understand the CSMA/CD technique, however, so that you can see how Ethernet has evolved from a shared media networking technology to the switched environment used today. For more information about CSMA/CD, see Chapter 13.

IEEE 802.4: Token-Passing Bus Access Method and IEEE 802.5: Token-Ring Access Method

Token-Ring and Token-Bus technologies have a lot in common. They both assume a ring topology, and a token frame is passed from one node to another. When a node on the network needs to transmit data, it waits until it receives the token frame and then transmits a data frame. The data frame travels around the ring until the destination node receives it. Upon successfully receiving a data frame, the destination node sets a few bits in the frame to indicate that it was successfully received and retransmits the modified frame on the network. When the sending node receives the frame it originally sent, it can check the bits to see that the data was received, and it then transmits a token frame so that another node can use the network.

The major difference between the Token-Bus and the Token-Ring network is that the Token-Ring network is physically wired in a ring topology. That is, the transmitter of a node is connected to the receiver of the next node in the ring, until the last node in the ring connects back to the receiver of the first node in the ring. When a Token-Bus is used, a single network media is used that connects all nodes, similar to a bus topology in early Ethernet networks. However, the ring topology is maintained as a logical ring. Instead of using the CSMA/CD method to access the shared bus media, each node on a network passes the token frame from one node to another, in a predetermined order, and thus a ring formation is still used.

Token-Ring networks are still in use today, though their numbers are far outweighed by the installed base of Ethernet networks. Token-Bus networks were generally used in industrial situations, such as factory floors, where a guaranteed minimum access time was crucial. Development of Token-Bus topologies has now been discontinued. To quote the IEEE, "These standards were administratively withdrawn by the IEEE Standards Board." Although Token-Ring networks are still marketed today, the speeds at which Ethernet now operates has all but rendered Token-Ring to history.

IEEE 802.7: Recommended Practices for Broadband Local Area Networks

This standard, first published in 1989, described various items that were pertinent to offering broadband communications at that time. These included a bus topology and amplifiers using coaxial cabling and frequency division multiplexing (FDM) that allows for communications in two directions, by using different frequencies for each direction of the link. This standard can be considered the grandfather of the standards used for cable and the xDSL modems that are so popular today. It is mentioned here so that you'll understand that broadband communications based on cable networks were envisioned many years ago, and are commonplace today.

IEEE 802.10: Security

First published in 1998, this standards document defines "IEEE Standards for Local and Metropolitan Area Networks: Standard for Interoperable LAN/MAN Security (SILS)." As the title implies, this standard discusses many aspects of security for both local area networks and metropolitan area networks. This document was updated in 1999, and it defines many concepts that have been adopted, or modified for use in networks today. For more information about current security standards, read Part VIII of this book, "System and Network Security," which contains several chapters devoted to this important topic.

IEEE 802.11: Wireless

Perhaps the most fascinating new development in the past few years is the widespread adoption of wireless networking technologies. From short-range technologies such as Bluetooth (see Chapter 22, "Bluetooth Wireless Technology") to the now popular Wi-Fi (802.11a, 802.11b, and 802.11g) networks, wireless LANs have become commodity items as the price of hardware has dropped dramatically. The 802.11 standards are covered in several documents. For example, 802.11g is now available in your local discount electronics store, and the Wi-Fi branding enables you to choose equipment from different manufacturers with the guarantee that the equipment has been tested for interoperability.

Today, the 802.11g standard (a faster version of the pioneering 802.11b standard) has become the de facto standard for wireless networking, with increasing numbers of dual-band 802.11g/802.11a adapters, routers, and switching being used in both SOHO and corporate networks. Other 802.11 standards have been developed to encompass security measures, and other such items as virtual LANs, as discussed in Chapter 9.

Part V, "Wireless Networking Protocols," covers the implementation of wireless networking standards in greater detail.