Consideration is given to the implementation of distributed discrete-event simulation (DDES) using what has been commonly called the Misra approach, after one of its inventors. A major problem with DDES is that deadlock can occur. Therefore, DDES algorithms must either avoid deadlock in the first place, or detect the existence of deadlock when it does occur and eliminate it. J. Misra (1986) proposes the use of null messages as one way to circumvent the deadlock problem. However the number of null messages can become quite large. Methods are presented for reducing the number of null messages through the prediction of channel times. A framework is presented on the basis of which distributed discrete-event simulation can be built for applications that can be decomposed into feedforward and feedback networks. >
[1]
Eric G. Manning,et al.
Distributed Simulation Using a Network of Processors
,
1979,
Comput. Networks.
[2]
Jayadev Misra,et al.
Distributed discrete-event simulation
,
1986,
CSUR.
[3]
David R. Jefferson,et al.
Virtual time
,
1985,
ICPP.
[4]
Devendra Kumar,et al.
Simulating feedforward systems using a network of processors
,
1986,
ANSS '86.
[5]
K. Mani Chandy,et al.
Distributed Simulation: A Case Study in Design and Verification of Distributed Programs
,
1979,
IEEE Transactions on Software Engineering.