A good customer-owned OSP design depends on your knowledge of cable types, placement methods, and network topologies.
Installers wiring buildings for voice, data, and video (V/D/V) can usually expect to “home-run” their cable runs, because a star-shaped wiring topology is the preferred infrastructure design for inside wiring. However, the requirements for customer-owned outside-plant (CO-OSP) network designs are much more varied than those for premises networks, so many more options than the home-run are available.
The Second Edition of BICSI’s Customer-Owned Outside Plant Design Manual identifies the following concerns for CO-OSP network designers:
The characteristics of the cabling media to be pulled.
The methods by which the cabling media will be installed.
Restrictions resulting from the geography of the campus, along with the transmission requirements of the campus network.
OSP cabling media. Although overall cable construction for outdoor installations is likely to be much more varied than for indoor installations, the actual cabling media used—singlemode optical fiber, 62.5/125- and 50/125-micron multimode optical fiber, unshielded twisted-pair (UTP) cable, and 75-ohm coaxial cable—are very similar to those employed for premises backbone cabling.
Optical fiber, which carries signals in the form of light pulses, can be used for extended distances and has much greater bandwidth, or carrying capacity, than copper-wire media. Optical fiber is also lighter and more compact than copper wire, is immune to electromagnetic interference (EMI), and offers greater security because it is difficult to tap. As a result, fiber is well suited to heavy-industrial applications, where a great deal of electrical interference is common. It is also widely used in military installations for security reasons.
Copper cabling media transmit electrical signals rather than light pulses, and twisted-pair cable—often in high pair counts—is the mainstay of many regional and local telephone companies. Although twisted-pair cable is popular among installers, it’s subject to electrical interference and has distance limitations when it comes to high-bandwidth applications.
Coaxial cable, or coax, is also a copper-based transmission medium, but it operates on a different principle from twisted-pair. In addition to being the choice of cable-television providers, coax has found a place in private broadband video networks like those providing television service to college dorm rooms. Coax offers higher bandwidth than twisted-pair, and it’s also less susceptible to interference. However, it’s more expensive than other media, and it presents installation complications because its shielding must be grounded.
Installation methods. Last month’s installment of this series discussed the three methods of installing outside-plant cable: aerial, direct-buried, and underground.
Aerial installations, consisting of cables and other apparatus mounted on utility poles, are the least expensive of the three types and are readily accessible for maintenance. However, they also pose several problems, from aesthetic concerns and susceptibility to environmental damage to tricky planning necessitated by considerations of tension, sag, clearance, and wind- and ice-loading.
Direct-buried OSP installations are usually installed by means of trenching, plowing, or directional boring. They are less expensive than underground installations and don’t clutter the appearance of the property because they’re out-of-sight. Yet although they can easily bypass obstacles, they are less flexible than conduit once installed and cannot be upgraded or expanded. They may be difficult to relocate for repair, they provide less physical protection for transmission media than conduit, and they may provide avenues for water or gases to enter buildings they serve.
Underground installation, in which cable is pulled through conduit, offers the aesthetic appeal of the direct-buried method, as well as providing greater cable protection and offering more potential for future upgrades. However, this method is more costly than direct burial and requires more careful route planning.
Network topologies vary. The topology of a network is its layout scheme. It can become difficult, however, when the physical layout of the network, or its physical topology, differs from the way in which the network exchanges information, or its logical topology. CO-OSP networks are similar in structure to premises networks. The three main topologies—star, bus, and ring—can be combined in a number of ways to produce hybrid networks offering workable solutions to many different geographical constraints and administrative concerns (Fig. 1).
All other factors being equal, the star is the preferred configuration for outside-plant installations. The star arrangement provides a central hub from which the network administrator can manage the physical network. If possible, this main cross-connect (MC) or campus distributor (CD) should be co-located with the enterprise’s data center.
If the distances spanned by a campus network exceed the transmission distances allowed by a star topology, the hierarchical star is a hybrid arrangement that may solve the problem. In this arrangement, an intermediate cross-connect (IC) or building distributor (BD) at one “point” of the star links buildings still farther away from the hub. Ideally, such IC/BDs should be linked to outlying buildings via a star topology, and no more than five such “hierarchical stars” should be designed into any one-campus network (Fig. 2).
Hierarchical stars are best used when geographical constraints make a single star configuration impossible, or when user requirements make it advisable to segment the network. In large networks, active electronics such as bridges and switches can often be used more effectively to balance bandwidth requirements and distance limitations with such a topology.
Ring-shaped networks come in two varieties: virtual rings and physical rings. Virtual rings have the physical topology of a star-shaped network but operate logically as a ring. Physical rings are widely used in CO-OSP networks because the networking technologies employed on a campus often call for a ring-shaped infrastructure; examples of such LAN and WAN technologies include Token Ring, SONET (synchronous optical networking), ATM (asynchronous transfer mode) and FDDI (fiber distributed data interface) (Fig. 3 above).
A virtual ring is used when obstacles on the campus prevent a physical star from being implemented, and when existing star-topology cabling facilitates such a design. Its main advantage is that it concentrates monitoring, maintenance, and backup operations at a central hub. However, this hub also is the scheme’s main disadvantage, because it’s a central point of failure in the event of a disaster.
The physical ring topology is used when there is a conduit system in place to support it, redundant cable paths are provided, and the predominant networking technology is ring-shaped, such as SONET, FDDI, or Token Ring.
A bus topology (Fig. 1) can be used when a fault-tolerant design isn’t required. The linear bus design is of limited use because it will fail if a cable break occurs anywhere in the network. It doesn’t provide backup or redundant transmission paths, so it can’t be used for mission-critical applications.
Other hybrid designs include the clustered star, which combines star and ring networks, and the tree and branch network, which is used for telephone and video distribution over copper twisted-pair and coaxial cables. Next month we’ll cover pathways and spaces from Chapter 3.