Browsing by Author "Wang, Jain J."

Now showing 1 - 7 of 7
  • Thumbnail Image
    Approximate Time-Parallel Simulation of Queueing Systems with Losses
    Wang, Jain J.; Abrams, Marc (Department of Computer Science, Virginia Polytechnic Institute & State University, 1992)
    This paper presents a guideline of a partial state matching approach for time-parallel simulation. Two algorithms using this approach to simulate FCFS G/G/1/K and G/D/1/K queues in which arriving customers that find the queue full are lost are proposed. Experiments with M/M/1/K and M/D/1/K models show that the performance of the algorithms in terms of convergence speed and accuracy is good in general cases. The worst performance of the algorithms occurs when traffic intensity approaches one. An argument is made to explain this phenomenon.
  • Thumbnail Image
    Determining Initial States for Time-Parallel Simulations
    Wang, Jain J.; Abrams, Marc (Department of Computer Science, Virginia Polytechnic Institute & State University, 1992)
    In this paper, we propose a time-parallel simulation method which uses a pre-simulation to identify recurrent states. Also, an approximation technique is suggested for approximate Markovian modeling for queueing networks to extend the class of simulation models which can be simulated efficiently using our time-parallel simulation. A central server system and a virtual circuit of a packet-switched data communication network modeled by closed queueing networks are experimented with the proposed time-parallel simulation. Experiment results suggest that the proposed approach can exploit massive parallelism while yielding accurate results.
  • The Impact of Lookahead on Conservative Simulation
    Wang, Jain J.; Abrams, Marc (Department of Computer Science, Virginia Polytechnic Institute & State University, 1995-02-01)
    This paper studies the impact on the performance of conservative simulation for both open and closed models. For open models, we derive an upper bound on the performance improvement due to the lookahead. We show that the benefit of lookahead diminishes as the simulation length increases. For closed models, on the other hand, the performance improvement converges to a constant as the simulation length increases.
  • Massively Parallel Simulations with Application to Queueing Networks
    Wang, Jain J.; Abrams, Marc (Department of Computer Science, Virginia Polytechnic Institute & State University, 1995-02-01)
    This paper investigates two parallel simulation methodologies, multiple replication and parallel regenerative simulation. Problems of applying these methodologies to massively parallel simulations are identified. Two approaches (MR-PI and PR) are proposed to overcome some of these problems. The MR-PI approach, based on multiple replication simulation, uses a pilot simulation to reduce the initial transient bias by starting the simulation from a state that is representative of the steady-state conditions. The PR approach allows approximate regeneration by using substate matching for models that do no regenerate frequently. Empirical results suggest that the proposed approaches can produce very accurate results.
  • Massively Time-Parallel, Approximate Simulation of Loss QueueingSystems
    Wang, Jain J.; Abrams, Marc (Department of Computer Science, Virginia Polytechnic Institute & State University, 1994-02-01)
    A time-parallel simulation obtains parallelism by partitioning the time domain of the simulation. An approximate time-parallel simulation algorithm named GG1K is developed for acyclic networks of loss FCFS G/G/1/K queues. In the first phase, a similar system (i.e., a G/G/1/infinity queue) is simulated using the GLM algorithm. Then the resultant trajectory is transformed into an approximate G/G/1/K trajectory in the second phase. The closeness of the approximation is investigated theoretically and experimentally. Our results show that the approximation is highly accurate except when K is very small (e.g., 5) in certain models. The algorithm exploits unbounded parallelism and can achieve near-linear speedup when the number of arrivals simulated is sufficiently large.
  • Thumbnail Image
    Time-Parallel Simulation Using Partial State Matching for Queueing Systems
    Wang, Jain J.; Abrams, Marc (Department of Computer Science, Virginia Polytechnic Institute & State University, 1992)
    This paper describes partial state matching for approximate time-parallel simulation. The notion of degree of freedom in time-parallel simulation is introduced. Two partial state matching algorithms are proposed to simulate acyclic networks of FCFS G/G/1/K queues in which arriving customers that find the queue full are lost. The algorithms are suitable for SIMD as well as MIMD architectures. The performance of the algorithms is studied. Experiment results with M/M/1/K and M/D/1/K queueing networks show that the potential speedup and simulation accuracy of the algorithms are good. The worst performance of both algorithms occurs when traffic intensity is one. Arguments are made to explain this phenomenon.
  • Thumbnail Image
    A Workload Emulator Architecture for Distributed Systems
    Abrams, Marc; Arnesen, John; Batongbacal, Alan; Chandraseka, Chokanath; Wang, Jain J. (Department of Computer Science, Virginia Polytechnic Institute & State University, 1991)
    This report presents the initial design of a general purpose workload emulator that emulates workstations, terminals, and communication equipment attached to host computers and servers in a distributed computing environment. The workload emulator can be programmed to model the time dependent behavior of a workload for performance evaluation of distributed systems.