Chapter 4. Building the Network Infrastructure

Understanding Network Architectures

In Chapter 2, "Different Needs, Different Networks," we took a look at the different physical topologies used to describe how devices are physically laid out on a LAN. Topology is a fairly broad way to discuss how a particular LAN is organized. A network architecture , on the other hand, provides more specific information on not only physical layout but also the cabling specifications that can be used and the actual method that the computers and other devices use to access the network media. Network architectures are defined by strict specifications provided by the Institute of Electrical and Electronics Engineers (IEEE), an international organization whose mandate is to develop and share electrical and information technology specifications worldwide.

In Chapter 5, "Network Protocols: Real and Imagined," we will take a look at the OSI model, which provides a theoretical look at how network communication takes place between a sending and a receiving device. Although the IEEE standards relate to real-world network functionality, these specifications have actually been grouped at the Data Link layer of the OSI theoretical model. Now, having said this, we will save the theory until we discuss OSI and LAN protocol stacks in Chapter 6. Just be aware that for the purpose of discussion and development, the IEEE folks have placed their actual network architecture specifications into a model that explains network interaction in a purely theoretical manner.

As with any technology, network architectures have come and gone. For example, ARCnet (Attached Resource Computer Network), the oldest network topology ( old is a relative term because this architecture was created in 1977), used a ring topology in which an electronic token was passed from computer to computer. The computer with the token could access the network and send data.

You would have to look far and wide to find an ARCnet network (although, I'm sure they are out there); other network architectures such as Ethernet (the most popular in the world) and IBM Token-Ring have replaced this early network architecture. What's more, new architectures, such as Fiber Distributed Data Interface (FDDI), and Asynchronous Transfer Mode (ATM) have been developed to provide faster throughput on network backbones (ATM is discussed in Chapter 13, "Expanding a LAN with WAN Technology").

Before we take a look at the commonly used network architectures (as of today) and how even these architectures are continuing to evolve , we need to discuss the terms data transmission speed and bandwidth . Data transmission speed is measured in bits per second and noted as bps . A bit is one binary digit, either a 1 or a 0 (it is the smallest unit of data; 8 bits actually make up a byte). All the architectures that we will look at provide data transmission speeds in excess of a million bits per second, which is noted as Mbps .

In terms of networking, bandwidth is considered the number of bits that can be sent across the network medium at a given time. So, the terms data transmission speed and bandwidth are often used interchangeably when discussing an architecture's data throughput.

Another issue related to these different architectures is the actual distance that data can be transmitted along a particular medium type. For example, Ethernet has a limitation of 100 meters over copper twisted-pair wire (without using any type of amplification device such as a repeater, which is discussed later in the chapter). If fiber- optic cable is used, an Ethernet run can be up to 2,000 meters (that's 2 kilometersa meter is slightly longer than a yard, so 2,000 meters would be more than 20 football fields long).

We will be discussing bandwidth and maximum cable lengths as they relate to each of the network architectures we discuss. Now, let's take a look at the LAN architectures themselves .