Getting information to its destination is a big part of network administration. With users demanding Web-based solutions, such as streaming video for training, the role of the administrator is looking more and more like that of a traffic cop. Dictating which applications will or will not be supported on the network is essential to walking the fine line between network efficiency and user happiness. That is why it is necessary to know how the data gets put out on the wire so your network can deliver the goods. An overview of transmission types is a good place to start. There are three basic methods of transmitting data: unicast, broadcast, and multicast.
When a transmission occurs using the unicast method, a copy of each packet (or cell) must be sent to every client who wants to receive the data. For example, if 10 users want to communicate with a video-conferencing server, and the video-conference stream consists of 2 Mbps, the amount of bandwidth using a unicast transmission method would equal 20 Mbps. If these users were on a 10BaseT network, the video-conference unicast traffic would completely saturate the network. Now imagine that there are 100 users involved in a four-hour video conference, and you can see where using unicast transmissions could bring your network to a standstill.
Using the broadcast transmission method to transmit voice and video is an improvement over the unicast method, but it does have limitations. When a transmission occurs using the broadcast method, only one copy of a packet (or cell) is sent to a broadcast address. For example, if 100 users on the 10.1.1.0 network wanted to connect to a video-conferencing server sending a data stream of 2 Mbps, the total amount of bandwidth required would be 2 Mbps. Broadcast transmissions use bandwidth more efficiently because only one packet (or cell) is sent to a single broadcast address. In our example, the broadcast address would be 10.1.1.255.
There are two major problems with broadcast transmissions. First, every machine on a segment must look at every broadcast packet that crosses its network interface. This robs systems of valuable processing time and can cause a dramatic drop in performance on older PCs. In our example, there are 100 members of the network participating in the conference call. Unfortunately, if there are 200 members on the 10.1.1.0 network segment, even devices that are not participating in the conference call will receive the broadcast packets.
The second problem with using broadcast transmissions for voice and video is the way most networks are designed to handle broadcast packets. In routed networks, by default, broadcasts are kept on the segment from which they originate. This is because most Layer 3 devices, such as routers, drop broadcast packets. Because many protocols make extensive use of broadcasts, routed networks drop broadcast packets to prevent them from saturating the network. Therefore, in our example, a user in the 18.104.22.168 network would not be able to join the conference call with the 10.1.1.0 network.
Multicasting was once used almost exclusively for home users connected to the Internet. However, over the past few years, multicasting has evolved into a valuable tool that allows companies to transmit voice, video, interactive training, and software updates. As more companies become aware of the benefits of multicasting, the amount of multicast traffic on corporate networks is bound to increase.
Essentially, multicasting is the transmission of data from a single source to multiple destinations using a single data stream. Transmitting voice and video over a network can require an enormous amount of bandwidth. The use of multicasting for those transmissions has many benefits over traditional transmission methods.
Using multicast transmissions for voice and video is an efficient means of transport that allows users of discontinuous subnets to receive a single data stream without flooding the network with traffic. A multicast transmission sends a single copy of a packet (or cell) to multiple multicast addresses. Unlike broadcasts packets (or cells), only those stations that are participating in the multicast group receive the data stream. This is accomplished using special multicast addresses.
Many network devices are configured by default to handle multicast traffic in the same manner as broadcast traffic—that is, to drop the packet. Therefore, to forward multicast packets, many devices require special configuration. In addition, to support a multicast network, the network devices must use protocols such as Internetwork Group Management Protocol (IGMP) and Cisco Group Management Protocol (CGMP) to manage the multicast traffic.
Multicast address structure
Multicast addresses use 32-bit Class D IP addresses. Multicast applications use the multicast address to join multicast groups. Each station running a multicast application can choose to participate in a multicast group and see the multicast packets destined for the group. Conversely, each station can choose not to participate in a multicast group and ignore multicast packets. Table A lists some well-known Class D addresses.
Coming soon to a network near you
As multicast applications continue to grow in popularity and as businesses begin using their data networks to replace telephones and video-conferencing equipment, the amount of multicast traffic on the network will be an issue for network administrators. Managing a multicast network requires a fundamental understanding of the basic multicasting principles and terminology. In future articles, we’ll build upon these basics and discuss how to configure Cisco devices to support multicast transmissions. For more information on configuring multicasting in a Cisco environment, check out CCIE Professional Development: Cisco LAN Switching from Cisco Press.
Warren Heaton Jr., MCSE+I, CCNP, CCDP is the Cisco program manager for A Technological Advantage in Louisville, KY.If you'd like to share your opinion, please post a comment below or send the editor an e-mail.