Backbone The portion of a network that ties different departmental networks into a single whole. Backbones are primarily used in medium to large networks, such as those occupying a building or a group of buildings on a campus. The backbone carries the bulk of the network traffic and must be designed accordingly. Backbone designs generally fall into two basic categories: Distributed backbone: Refers to using cabling to join different departmental networks in a bus topology or mesh topology. This cabling is referred to as backbone cabling, and it connects the hubs, switches, or routers of each network into a single whole. In a typical scenario, each floor or building might have a local area network (LAN) and wiring closet containing, among other things, a main hub or router. Backbone cabling is then run between floors or buildings, connecting the main hub or router for each department into a bus-style network (see illustration). Collapsed backbone: Refers to using cabling to directly join each departmental network’s main hub or router using backbone cabling to a central hub, switch, or router in a star topology (see illustration). The central unit is often referred to as the collapsed backbone, although this term properly describes the entire configuration. The central unit can be located in the building’s main equipment room or, in a campus scenario, in the IS (Information Services) department’s building. Distributed backbones generally have a greater degree of fault tolerance than collapsed ones, because the collapsed backbone unit forms a single point of failure. However, collapsed backbones usually have better traffic flow than distributed backbones because of the underlying star topology. Collapsed backbones generally offer better performance because of the reduced number of hops that traffic must make when passing between departmental LANs. Collapsed backbones are also easier to manage because they bring all the backbone switching and routing equipment into a single room or building. Collapsed backbones are used frequently for connecting departmental LANs within a single building, but less often for connecting building LANs across a campus network because of the increased distances and cabling costs. Backbone cabling should have the highest bandwidth of any cabling in your network, since backbones are used to join together hubs, switches, and routers, linking departmental LANs or subnetworks into building-wide or campus-wide internetworks. In buildings, backbone cabling often refers to the vertical cabling running through the risers or elevator shafts that connects the hubs and switches in each floor’s wiring closet. Depending on performance requirements, anticipated growth, and cost, any of the following might be suitable for backbone cabling: ? Category 5 (Cat5) unshielded twisted pair (UTP) cable ? Type 1A shielded twisted pair (STP) cable ? Thinnet coaxial cabling ? Multimode fiber-optic cabling ? Single-mode fiber-optic cabling Notes The term backbone is also used to refer to the collection of networking components (cabling, hubs, switches, and routers) that form the supporting network into which workgroup and departmental LANs are connected. A mesh topology is often used for network backbones to provide fault tolerance for critical high-speed data paths. You should put considerable thought and planning into the design and implementation of your network’s backbone, because the overall performance of networking services is largely dependent on the backbone’s bandwidth and reliability. Design your backbone with network expansion in mind. Planning for growth is especially important if the cost of cable reinstallation is high. Fiber-optic cabling is preferred for most network backbones because of its high bandwidth, security, and resistance to electromagnetic interference. Collapsed backbone An enterprise networking methodology in which the network backbone consists of a single device. In a traditional network, local area networks (LANs) are multipoint connections connected using a backbone cable. For example, in a building, a fiber-optic backbone might run from floor to floor and connect with a hub in a wiring closet on each floor. In contrast, collapsed backbones make use of centralized switches, which provide virtual point-to-point connections for LAN connections. These switches are located in the same place as the network servers—in fact it was the move toward centralized location of network servers that helped drive the development of collapsed backbones. In a typical collapsed backbone scenario, instead of having a hub for each floor located in that floor’s wiring closet, a set of stackable Ethernet switches would be located in the equipment room in the basement, with individual cables running from this closet through vertical rises to wiring closets on each floor where hubs distribute connections to stations in work areas. Advantages and Disadvantages The advantages of using a collapsed backbone are that they eliminate the costs of backbone cabling installation, they require fewer devices, their equipment administration is more centralized, and they offer higher available bandwidth for each station. The disadvantages are that collapsed backbones generally are not feasible for use in more than one building, they require more cabling, they use more expensive devices, and they have a more limited distance capability.