∙Port switching—Takes place in the backplane of a shared hub. For instance, ports 1, 2, and 3 could be connected to backplane 1, whereas ports 4, 5, and 6 could be connected to backplane 2. This method is typically used to form a collapsed backbone and to provide some improvements in the network.

∙Cell switching—Uses Asynchronous Transfer Mode (ATM) as the underlying technology. Switch paths can be either permanent virtual circuits (PVCs) that never go away, or switched virtual circuits (SVCs) that are built up, used, and torn down when you’re finished.

A Bit of History

The first local area networks (LANs) began as a result of the introduction of personal computers into the workplace environment. As computers became more common, the need arose to share resources, such as printers or files. These early networks were pretty simple, with a handful of computers sharing a few printers and not much more. As more items such as servers, applications, and peripherals came along, the increasing numbers of interfaces—along with application designs that could take advantage of the network—created a weakness in the current network design.

The limitations of traditional Ethernet technology brought forth a number of innovations that soon became standard in the Ethernet protocol. Innovations such as full duplexing, Fast Ethernet, and Gigabit Ethernet began to appear—innovations that have also made possible a transition to switches from shared hubs.

Other limitations to the way networks operated in a shared environment created a need for alternative methods to permit the use of bandwidth−intensive applications such as video and voice. Switches are one of these alternative methods. In many respects, switches are relatively simple devices. A switch’s design and self−learning features require very little manual configuration to get it up and running. To properly use these devices in your network, you must have an in−depth knowledge of the issues involved in implementing switching.

Knowing the basics of Ethernet technology can help you effectively troubleshoot and install switches in the network. You also need a good grasp of the different technologies and how switches work, as well as the constraints of each type of device you may use in the network. As you read the following sections, make sure you get a clear understanding of the fundamentals and basics of Ethernet technology.

The types of devices you use in the network have important implications for network performance. For example, bridges and routers are both devices that network administrators use to extend the capabilities of their networks. Both of them have advantages and disadvantages.

Bridges, for example, can easily solve distance limitations and increase the number of stations you can have on a network, but they can have real problems with broadcast traffic. Routers can be used to prevent this problem, but they increase the time it takes to forward the traffic.

This has been the pattern throughout the history of networking. When a new product is introduced, problems or bottlenecks are soon found that limit the product’s usefulness. Then, innovations are invented or implemented to aid the product and allow it to perform better. To see this occurrence in action, let’s take a look at some of the traditional network architectures. As you will see in upcoming sections, the pattern of new innovation after new innovation started in the earliest days of networking and continues in today’s networks.

Networking Architectures

Network designers from the beginnings of networking were faced with the limitations of the LAN topologies. In modern corporate networks, LAN topologies such as Ethernet, Token Ring, and FDDI are used to provide network connectivity. Network designers often try to deploy a design that uses the fastest functionality that can be applied to the physical cabling.

7

Many different types of physical cable media have been introduced over the years, such as Token Ring, FDDI, and Ethernet. At one time, Token Ring was seen as a technically superior product and a viable alternative to Ethernet. Many networks still contain Token Ring, but very few new Token Ring installations are being implemented. One reason is that Token Ring is an IBM product with very little support from other vendors. Also, the prices of Token Ring networks are substantially higher than those of Ethernet networks.

FDDI networks share some of the limitations of Token Ring. Like Token Ring, FDDI offers excellent benefits in the area of high−speed performance and redundancy. Unfortunately, however, it has the same high equipment and installation costs. More vendors are beginning to recognize FDDI and are offering support, services, and installation for it—especially for network backbones.

Network backbones are generally high−speed links running between segments of the network. Normally, backbone cable links run between two routers; but they can also be found between two switches or a switch and a router.

Ethernet has by far overwhelmed the market and obtained the highest market share. Ethernet networks are open−standards based, more cost−effective than other types of physical media, and have a large base of vendors that supply the different Ethernet products. The biggest benefit that makes Ethernet so popular is the large number of technical professionals who understand how to implement and support it.

Early networks were modeled on the peer−to−peer networking model. These worked well for the small number of nodes, but as networks grew they evolved into the client/server network model of today. Let’s take a look at these two models in more depth.

Peer−to−Peer Networking Model

A small, flat network or LAN often contains multiple segments connected with hubs, bridges, and repeaters. This is an Open Systems Interconnection (OSI) Reference Model Layer 2 network that can actually be connected to a router for access to a WAN connection. In this topology, every network node sees the conversations of every other network node.

In terms of scalability, the peer−to−peer networking model has some major limitations—especially with the technologies that companies must utilize to stay ahead in their particular fields. No quality of service, prioritizing of data, redundant links, or data security can be implemented here, other than encryption. Every node sees every packet on the network. The hub merely forwards the data it receives out of every port, as shown in Figure 1.1.

Figure 1.1: A flat network topology.

Early networks consisted of a single LAN with a number of workstations running peer−to−peer networks and sharing files, printers, and other resources. Peer−to−peer networks share data with one another in a non−centralized fashion and can span only a very limited area, such as a room or building.

8