| Abstract | In the standard token ring network, the performance will be
deteriorated when the number of stations in the ring is increased
and/or the arrival rate of message is increased. To improve the
characteristic of the basic token ring, an architecture providing
multiple physical channels was found to be the alternative. A
bridged token ring architecture is one way to improve the
performance of the basic token ring . However, since the
allocation of stations to ring and traffic dictated the
performance of the bridged token ring architecture, therefore a
multiple channel token ring architecture was found to be the
suitable solution.
A multiple channel tokening architecture with separate
buffer, simultaneous transmission has a number of concentric
rings with each station attached to all the rings. Each station
is capable of transmitting and receiving messages to and from any
free rings whenever receiving a free token. The message from the
station are not equally distributed among the different rings.
There is one transmit buffer in which the message are stored and
from which the transmit logic extracts the messages for
transmission when any ring becomes available.
All of them, the basic token ·ring, the bridged token ring
and the multiple channel token ring, were assumed to conform with
the IEEE 802.5 standards. They were modeled to work under
symmetric traffic, non-exhaustive service discipline, fixed
message length and single message token generation policy with
source removal. They were then simulated to evaluate the
performance. The multiple channel token ring architecture was
found to give a better performance than the single ring and the
bridged token ring by offering lower delays and yielding the
higher throughput under the same load. In addition, it also kept
the traffic and allocation of station dependent characteristic as
in the basic token ring.
In Ethernet, it has been clearly noted from the performance
point of view that CSMA/CD is wasteful in that, at any instant,
only one successful transmission can be accommodated on a common
channel at a time since a successful transmission even between
two adjacent stations will render the entire channel inaccessible
to all other stations. In addition, in real situation the
stations could be located at any point of the network and mostly
formed a cluster of nodes .
To improve the performance
successful transmission protocol is
of CSMA/CD, is able to dynamically
independent segments, each capable
transmission when ever possible.
- ii -
of Ethernet, simultaneous
proposed. Segnet, a variation
partition a bus network into
of accommodating a successful
Therefore it utilizes the
channel more efficient than normal CSMA/CD. However, in Segnet
the message transmission may not be terminated only because of
message collision but also because of blocking by the node at the
segmentation point . Thus, buffered Segnet capable of storing the
message in the queue of the station at the point of segmentation
was also investigated.
The simulation models of buffered and non-buffered Segnet
were simulated under the clustered condition . Only the best case
and the worst case of clustered setup were chosen. It was found
that the performance of both buffered and non-buffered Segnet
were outperformed over the normal Ethernet due to its ability in
simultaneous successful transmission and buffered Segnet was
superior to the non-buffered Segnet resulting from its absence of
transmission failures due to segmentation which implies fewer
retransmissions and lesser mean message transfer delay . In
addition, Segnet also improves the fairness of each station for
the same reason. |