| Summary: | The increasing traffic flow of packets on the network computers that are not
offset by an increase in service will cause a queue of data packets until the
availability of facilities resource. Queues and packet transmission process is very
complicated on a computer network that can not be solved analytically, and
therefore used modeling and simulation to analyze the queuing system.
Queuing model discussed in this study is an open network of queues where the
packets of data arriving from the outside and from within the system which passes
through several nodes to complete the job. The arrival of data packets follow a
Poisson distribution with Exponential service time and Erlang distribution are used
as a comparison. Analysis of the data packets are using a single server model
where, if the service time of data packets using the Exponential distribution, then
the model become M/M/1, whereas the service time using the Erlang distribution
with parameters m and μ, then the model becomes M/E[m]/1. The process of queuing
networks in general will be served by many servers.
This research use an open Jackson method to perform queueing network
simulations to calculate the characteristics of the network with parameters as
follow: the average waiting time of data packets in queue on a network system, the
average waiting time of data packets on a node in a network system, the average
number of packets in a network system , the average service time of data packets in
a network system, and the probability of a busy network node system.
Examination of simulation system use data packets on a computer network
server of Department Computer Science University of Udayana to determine
system performance using with different service time distribution. The result of
examination indicate that service time of Exponential distribution has better
characteristic then Erlang distribution at parameter m-Erlang � 2.
|
|---|