About IBM’s Cabling System
IBM’s Cabling System is designed to reduce or eliminate the expense of rewiring a building or pulling cable when DTEs such as computers and terminals are installed or moved.
It is meant to be an alternative to dedicated coaxial cable, which is used conventionally to attach remote terminals to mainframe computers.
IBM’s approach in a LAN configuration is to: minimize the length of the cable pulled through a building, and provide for concentration points in the cabling system or link. The first objective reduces costs and the transmission distance between the DTEs. The second objective facilitates installation, configuration, and reconfiguration of the topology. Most importantly, it facilitates network maintenance and simplifies troubleshooting.
Given these two goals, IBM states one approach is to provide a serial bus connection, as indicated in Figure. While this is quite good for meeting the first goal (very little cable is pulled through a building for a serial bus), the serial interconnection is very poor in meeting the second goal. It is difficult to troubleshoot a serial interconnection, and it is very difficult to reconfigure a serial bus. Option TWO provides for concentration points. In contrast to Option One, it is quite poor in meeting the first goal but quite good at meeting the second.
IBM’s Cabling System, IBM’s approach is to combine Options One and Two. The combination provides for concentration points which bridge linear buses (which are actually rings).
The cabling system consists of wiring closets (concentrators) placed in various locations in the building. Each closet has a distribution panel. This panel is used to connect the wiring which runs through the various walls of the offices. Each panel can accept up to 64 cable pairs or, stated another way, 64 individual devices. The wiring closets and their attached cables fan out to special outlets in walls. The wiring permanently connected to the wiring closet permits connections to be made into the outlets in the office walls with a special plug. This concept is quite similar to the electrical outlets we have in our homes and offices, where if we wish to move something (for example, a clock), we simply unplug it and move it to another outlet in the wall.
Figure 6-ll(c) depicts the token-ring bridge. This component serves as an off-ring communications link to other ring networks. Ring-to-ring communications are provided by a backbone of 4 Mbit/s twisted-pair coaxial cable or 16 Mbit/s optical fiber. The individual nodes can be outfitted with PC adapter boards, allowing an IBM PC network interface into the token ring.
As stated earlier, IBM’s Cabling System, IBM supports the conventional telephone-type unshielded twisted-pair cable. The token ring permits each node to utilize this technology. This can be quite attractive for small business or users who wish to use the wiring already installed in a building. However, unshielded telephone twisted-pair cable is subject. to more problems than other kinds of media. Consequently, IBM also offers higher quality cable connections, called data-grade cable. If the user chooses the conventional telephone wiring, the maximum number of stations that can be attached to the ring is 72. In addition, the mixing of wires is not permitted on the individual ring.
All stations must use the same type of media. However, this does not preclude two rings using twisted-pair cable connecting to each other through the wiring closet from which the wiring concentrators could use the higher grade cable. The use of twisted-pair cable limits the distance of the devices to no more than 45 meters to the wiring closet. Wisted-pair cable is more subject to noise, jitter, decay, and clocking problems, but the reduced cost may justify its use.