Read an Excerpt
Chapter 1: Introduction to Emerging Switching Technologies and the Campus NetworkIn order to design and build multilayer switched networks, it is important to first understand the demand for switches and what problems they solve. This chapter explains the problems the networking industry faces with existing technology, and the solution that switches provides.
The first sections of this chapter cover the origins of switching technology and explain in detail the problems facing businesses. The next sections introduce Virtual Local Area Networks (VI.ANs), and discuss where switching resides in the OSI model. The final section is an introduction to the switch block in Chapter 2, "Switch Block Concepts."
Origins and Directions of Switches/LAN Segmentation-Broadcast Issues About 15 years ago, LANs (local area networks) had little need for high-speed network connectivity. Networks usually consisted of a mainframe with "dumb" terminals or perhaps a few personal computers (PCs) connected to a central server for data storage and printing. Most business applications were installed on local desktop machines with very little data transmitted between the server and client. These "networks" were primarily used for storing and sharing data among small groups of users (workgroups).
This soon changed, and LANs quickly became the central point for all enterprise applications, data sharing, and storage. More and more applications were being staged from centralized servers. Companies soon were connecting all of their workgroups together into larger and larger organizations providing centralized data storage, workgroup collaboration (e-mail and scheduling software), accounting packages,and other enterprise applications. As the number of users located on these networks increased, so did the need to physically segment the users while logically connecting them to the core resources. Early networking devices were developed that allowed network administrators to geographically group centralized resources and link them with geographically disparate users.
These early networking devices were hubs, bridges, and routers. Each was responsible for a separate layer of communication. These devices worked very well during the first years of explosive growth. Technology soon caught up with these first- and second-generation networking devices. New applications were developed that pushed current technology beyond its capacity. Multimedia applications, networked voice applications, powerful engineering tools, and Enterprise Resource Planning packages took their toll on networks. It was time for data communication technology to catch up, and the answer was the switch. This new device combined the functionality of multiport bridges and sometimes routers into a single layer 2/3 device. The device allowed for reliable high-speed transmissions of data, while expanding port density and eliminating unnecessary networking devices.
New switches contain application-specific integrated chips (ASICs) that allow data to be switched several thousand times faster than predecessor switches, while allowing administrators more granularity and features.
The primary issue fueling the push for high-speed switching was the inherent nature of the primary transmission protocol used in the industry, Ethernet. While not the most efficient transport protocol, it's the most widely used. Due to inherent limitations in the protocol, steps had to be taken to design equipment that allowed Ethernet to work as efficiently as possible while scaling to thousands of users.
Ethernet Broadcasts Look for a moment at Ethernet. Ethernet is a broadcast (contention) based media protocol. Ethernet technology states that, "each station must listen to the physical line for the absence of other traffic before transmitting." If no traffic is sensed, the transmitting station will broadcast its data over the wire. Each station on the Ethernet segment receives the packet and copies it to its buffer. The stations then check the packet's destination address against their own Media Access Control (MAC) address. If the destination address matches the receiving station's MAC address, the packet is processed; no match, the packet is dropped...