Logical bus topology

From Hill2dot0
Jump to: navigation, search
The Logical Bus Topology
Although point-to-point physical topologies are rapidly eliminating the need to discuss the difference between physical and logical topologies, or even to discuss topologies at all, there are still sufficient legacy systems in place to warrant a discussion of the predominant logical topologies. It is also germane to understanding why these environments may not be well suited to modern applications like IP telephony or videoconferencing.

In a LAN that implements a logical bus topology, the LAN delivers the signal generated by a transmitter to all attached devices. This is a true broadcast environment; no attached end device is required to regenerate a signal in order for other attached devices to receive it.

Obviously, a physical bus topology is intrinsically a logical bus topology as well. In fact, in the early days of Ethernet and LocalTalk, it would have seemed somewhat odd to separate the concepts at all. However, throughout that 1990s, logical bus topologies were more commonly associated with physical star topologies. That is to say, a logical bus was implemented over a physical star (i.e., a hub-based network). The hub is essentially a multiport repeater that is responsible for regenerating the signal to all attached systems.

Although the cabling plant differed, the concept of the logical bus topology remains essentially the same: No attached device (other than the repeaters themselves) regenerates another device’s signal within the LAN (i.e., passive attachment).

Interestingly, the wireless environment operates both physically and logically as a bus. Unless wireless antennae are directed, the wireless signal radiates outward in all directions from the transmitter, creating a small “cell” of signal. Any device in that region can read the signal and use the same environment for a return transmission. Therefore the physical and logical bus environment, while on the decrease in a wired LAN world, is alive and well in the wireless world.

Logical Bus Over Physical Star

Logical Bus Over Physical Star

The visual depicts a physical star topology that implements a logical bus topology. These environments are commonly implemented using twisted pair or optical fiber. The logical bus is implemented as a physical star by using a multiport repeater (also known as a hub), and a cable plant that radiates outward to all attached devices. While some environments permit daisy chaining, it is far more common to limit each length of twisted pair or optical fiber to two attachments. At one end of each cable segment is the hub. At the other end is either a LAN adapter or another multiport repeater. In a working LAN environment, the multiport repeater would be placed in a wiring closet. From the repeater would stretch two (or more) twisted pair or two optical fibers to each desktop.

If the port capacity of a multiport repeater has been exhausted, it is possible to connect two multiport repeaters to one another, effectively increasing the capacity of the LAN. The result is a tree-like configuration, as depicted on the visual. Note that, to the attached devices, such an environment operates almost exactly as it would if this were one large multiport repeater. Each repeater regenerates any incoming signal on all ports except the one on which it arrived. This is true whether the signal is arriving from a directly attached end station or via another repeater.

When two multiport repeaters are connected in this manner, a transmit/receive problem is likely to surface. Most physical star topologies use two twisted pair or two optical fibers to connect the end station to the repeater. The end station transmits on one pair (or fiber) and receives on the other. For this to work correctly, the repeater must implement ports that reverse this relationship (i.e., the end station's transmit pair must be the repeater's receive pair, and the end station's receive pair must be the repeater's transmit pair). Most multiport receivers are designed in exactly this way. In other words, they are designed to expect that the port is being used to attach an end station.

When a port is actually being used to attach another repeater, a Physical Layer communication problem arises. The multiport receivers are both receiving on the same pair and transmitting on the same pair. This would be like two people trying to talk on the telephone when one of them is holding the handset backwards! There is one general approach to solving the problem: The link between the two devices must be rolled, which effectively exchanges the two pairs. Some multiport repeaters provide a special port that has an associated switch. Set in one position, the port is configured to attach an end station. Set in the other position, the transmit/receive links of the port are exchanged and the port is configured to attach to another repeater. If the repeater does not implement this capability, then a special crossover cable must be wired to make the exchange.

Most twisted pair implementations support cable lengths up to one hundred meters. Although greater distances could be achieved using shielded twisted pair (STP), most implementations use data-grade (i.e., Category 3 or Category 5) unshielded twisted pair (UTP). When optical fiber is used, distances of two kilometers can be achieved with multimode fiber (MMF), and up to 60 kilometers with single-mode fiber (SMF).