Information Technology Reference
In-Depth Information
mation without allowing others a turn for their work. To resolve
this problem, Ethernet rules specify a maximum length for any mes
sage. If a computer wishes to communicate more information than
is possible within this maximum, then the data are divided into mul
tiple messages, and each is sent separately. With this rule in place,
other computers will have the chance to send their materials during
the gaps between the parts of any long message.
Altogether, an Ethernet network allows reliable communica
tions at reasonably high speeds (now up to 1 and sometimes 10 gi
gabytes per second) at a reasonable cost. In order to function suc
cessfully, however, it does require that the communications of one
computer do not interfere unduly with those of another. In practice,
this means that individual messages must be relatively short, and
that the network cannot become too busy. For example, if the net
work is busy more than 50% or 60% of the time, then interference
among computers becomes problematic and communication can be
come bogged down. In addition, although it may support many ma
chines, a simple Ethernet network maintains a single connection
that runs from one computer or hub to the next. Thus, a single
break in a cable or malfunction at one connection point will divide
the network into two parts, isolating one piece from the other.
TokenRing Network: Rather than connect all computers to a sin
gle cable or bus, a
token-ring network
organizes computers logically
into a circle or ring; that is, each computer logically maintains two
connections, one to the computer before and one to the computer
after it in the circle. Communication also progresses in a single di
rection around the circle, so each computer receives messages from
one side and sends messages to the other side.
For example, if we were to label computers consecutively along
the circle as A, B, C, and so on, then computer A would send mes
sages to B, B would send to C, C would send to D, and so on.
However, B would never send directly to A, and C would never
send to B.
Following this system, if A wanted to send a message to com
puter D, then A would send the data to B, B would forward it to C,
and C would finally deliver it to D. To coordinate this transmission,
and to know where messages were intended to go, each message has
both a sender and an addressee designator. Thus, in sending a mes
sage from A to D, A would deliver the material to B—indicating it
