Chapter 22: Phase III: Multicast-enabled Network

Overview

In chapter 20, the Fabrikam Media and Training departments produced a series of training videos to kick off the company-wide deployment of Windows Media. The company can now deliver on-demand streaming media to every desktop and computer-equipped meeting room. It can also deliver broadcast content on a limited basis using unicast delivery. After performing the steps outlined in this chapter, the company will have full capability to deliver broadcast content to every computer on the corporate network.

The key to delivering broadcast content to every desktop is a multicast-enabled network. As mentioned previously, unicast delivery is not the best solution for delivering hundreds or thousands of concurrent streams. For example, 5,000 clients receiving a 100-Kbps unicast stream would require a network infrastructure capable of handling a 500-Mbps load. The same number of clients receiving a multicast stream would require a maximum bandwidth of only 100 Kbps. Edge servers help the unicast bandwidth problem by distributing content closer to end users. However, the current number of cache/proxy servers deployed by the company would not be able to handle the expected broadcast load if unicast was the only method of delivery.

Multicast is by far the most bandwidth-efficient method currently available for delivering a broadcast to large numbers of end users in an enterprise. But one impediment to multicast delivery is the ability of the network infrastructure to handle multicast traffic. It is not a matter of bandwidth or computing horsepower because one multicast stream can easily feed every client in the company with no more bandwidth or server usage than one unicast stream. In order to use multicast on a network, the routers, switches, other network devices, and subnetworks must be configured to handle multicast traffic. Once the network is multicast-enabled and solutions have been implemented on network segments that cannot be enabled, multicast delivery can begin. If the system is configured properly, clients everywhere in the company can either join or subscribe to the multicast group and receive the broadcast.

In this chapter, we describe how Fabrikam enables multicasting on their enterprise network. It would be impossible to describe how to enable every network because each enterprise network is unique. The Fabrikam network is fairly typical for a mid-sized company, and includes modern networking equipment. Many networks have both new and older equipment, and contain a collage of operating systems and hardware.

There are certain minimum requirements for a network to handle multicast traffic. If parts of your network do not meet these requirements, you can either upgrade those parts or work around those areas. Chapter 20 described some common solutions for existing networks. You can, for example, add a cache/proxy server and use unicast stream splitting or multicast distribution to leap over troublesome areas.

To understand the concepts in this chapter, it helps to be familiar with the Open System Interconnection (OSI) Reference Model that defines a networking framework for implementing IP protocols. For more information, see the sidebar, The OSI Model. You should also be familiar with how the protocol layers work together to convey data. Before jumping right into enabling the network, you need to understand the basics of a multicast system.

The OSI Model

The Open System Interconnection (OSI) Reference Model defines seven layers that describe how applications running on network-aware devices communicate with each other. The model is generic and applies to all network types, not just TCP/IP; and to all media types, not just Ethernet.

• Layer 1 is the Physical Layer. It defines the physical and electrical characteristics of the network. The NIC in a computer and the interfaces on routers all run at this level because they send strings of ones and zeros through the cables.

• Layer 2 is known as the Data Link Layer. It defines the access strategy for sharing the physical medium, including data link and media access issues. Protocols such as PPP, SLIP, and HDLC operate at this layer.

• Layer 3 is the Network Layer. It provides a means for communicating open systems to establish, maintain, and terminate network connections. The IP protocol operate at this layer, and so do some routing protocols. All the routers in a network operate at this layer.

• Layer 4 is the Transport Layer, and is where TCP operates. The standard says that “The Transport Layer relieves the Session Layer (Layer 5) of the burden of ensuring data reliability and integrity.” It is at this layer that, should a packet fail to arrive (perhaps due to misrouting, or because it was dropped by a busy router), it will be retransmitted when the sending party fails to receive an acknowledgement from the device with which it is communicating. The more powerful routing protocols also operate here. OSPF and BGP, for example, are implemented as protocols directly over IP.

• Layer 5 is the Session Layer. It provides for two communicating presentation entities to exchange data with each other. E-commerce uses this layer for shopping baskets, which are placed on a Web server and are not load balanced in order to preserve the content of the session.

• Layer 6 is the Presentation Layer. This is where application data is either packed or unpacked, ready for use by the running application. Protocol conversions, encryption/decryption, and graphics expansion all take place here.

• Layer 7 is the Application Layer. This is where end user and application protocols operate, such as telnet, FTP, and mail (POP3 and SMTP).

Devices that communicate across networks use the OSI seven layer model as follows: An application forms a packet of data to be sent; this takes place at layer 7. As the packet descends the layers, also known as the stack, it is wrapped in headers and trailers, as required by the various protocols. Having reached layer 1, it is transmitted as a frame across the medium in use. When the packet reaches device B, it moves up the stack. The networkdevice strips off the appropriate headers and trailers, delivering just the data to the application.