To know when and how to use the different kinds of cable media, you need to understand what users do on the corporate network. The way to find this information is to ask questions. After that, you can use monitoring equipment to really see what is going on inside the network cabling. Before you deploy an application on a corporate network, carefully consider bandwidth requirements as well as latency issues. More and more users need to compete for bandwidth on the network because of bandwidth-consuming applications. Although layer 2 switches break up collision domains and certainly help a congested network if correctly designed and installed, you must also understand the different cable media types available and where to use each type for maximum efficiency. That’s where this chapter comes in.
In 1980, the Digital Equipment Corporation, Intel, and Xerox (DIX) consortium created the original Ethernet. Predictably, Ethernet_II followed and was released in 1984. The standards- setting organization, the Institute of Electrical and Electronics Engineers (IEEE), termed this the 802.x project. The 802.x project was initially divided into three groups:
The High Level Interface (HILI) group became the 802.1 committee, and was responsible for high-level internetworking protocols and management.
The Logical Link Control (LLC) group became the 802.2 committee, and focused on end- to-end link connectivity and the interface between the higher layers and the medium-access- dependent layers.
The Data Link and Medium Access Control (DLMAC) group became responsible for the medium-access protocols. The DLMAC ended up splitting into three committees:
802.3 for Ethernet
802.4 for Token Bus
802.5 for Token Ring
DEC, Intel, and Xerox pushed Ethernet, while Burroughs, Concord Data Systems, Honeywell, Western Digital—and, later, General Motors and Boeing—pushed 802.4. IBM took on 802.5.
The IEEE then created the 802.3 subcommittee, which came up with an Ethernet standard that happens to be almost identical to the earlier Ethernet_II version of the protocol. The two differ only in their descriptions of the Data Link layer. Ethernet_II has a Type field, whereas 802.3 has a Length field. Even so, they’re both common in their Physical layer specifications, MAC addressing, and understanding of the LLC layer’s responsibilities.
See CCNA: Cisco Certified Network Associate Study Guide, 4th edition, by Todd Lammle (Sybex, 2003) for a detailed explanation of Ethernet frame types.
Ethernet_II and 802.3 both define a bus-topology LAN at 10Mbps, and the cabling defined in these standards is identical:
10Base2/Thinnet Segments up to 185 meters using RG58 coax at 50 ohms.
10Base5/Thicknet Segments up to 500 meters using RG8 or RG11 at 50 ohms.
10BaseT/UTP All hosts connect by using unshielded twisted-pair (UTP) cable with a central device (a hub or switch). Category 3 UTP is specified to support up to 10Mbps, category 5 to 100Mbps, category 6 to 155Mbps, and category 7 to 1Gbps.
Ethernet is the most popular type of network in the world and will continue to be so. It is important to understand how hubs and switches work within an Ethernet internetwork.
By using switched Ethernet in layer 2 of your network, you no longer have to share bandwidth with the different departments in the corporation. With hubs, all devices have to share the same bandwidth (collision domain), which can cause havoc in today’s networks. This makes a switched Ethernet LAN much more scalable than one based on shared Ethernet.
Hubs are layer 1 devices. The best way to think of them is as multi-port repeaters, repeating everything that comes their way, including runt frames, giant frames, and frames failing the frame check sequence at the end.
Even though layer 2 switches break the network into smaller collision domains, the network is still one large broadcast domain. Nowadays, switched Ethernet has largely replaced shared hubs in the networking world because each connection from a host to the switch is in its own collision domain. This is often referred to as micro-segmentation (as opposed to segmentation, where a legacy bridge may have created only two LAN segments). Remember that, with shared hubs, the network was one large collision domain and one large broadcast domain, whereas layer 2 switches break up collision domains on each port, but all ports are still considered, by default, to be in one large broadcast domain. Only virtual LANs, covered in Chapter 3, “VLANs, Trunks, and VTP,” break up broadcast domains in a layer 2 switched network.
Switched Ethernet is a good way to dynamically allocate dedicated 10Mbps, 100Mbps, and 1000Mbps connections to each user. By also running full-duplex Ethernet, you can theoretically double the throughput on each link. In the next sections, we’ll discuss how Ethernet is used in your internetwork, the differences between the Ethernet types, and half- and full-duplex.