Global dynamic load-balancing for decentralised distributed simulation

Distributed simulations require partitioning mechanisms to operate, and the best partitioning algorithms try to load-balance the partitions. Dynamic load-balancing, i.e. re-partitioning simulation environments at run-time, becomes essential when the load in the partitions change. In decentralised distributed simulation the information needed to dynamically load-balance seems difficult to collect and to our knowledge, all solutions apply a local dynamic load balancing: partitions exchange load only with their neighbours (more loaded partitions to less loaded ones). This limits the effect of the load-balancing. In this paper, we present a global dynamic load-balancing of decentralised distributed simulations. Our algorithm collects information in a decentralised fashion and makes re-balancing decisions based on the load processed by every logical processes. While our algorithm has similar results to others in most cases, we show an improvement of the load-balancing up to 30% in some challenging scenarios against only 12.5% for a local dynamic load-balancing.

[1]  Toyotaro Suzumura,et al.  X10-based massive parallel large-scale traffic flow simulation , 2012, X10 '12.

[2]  Jason Liu,et al.  Hierarchical Composite Synchronization , 2012, 2012 ACM/IEEE/SCS 26th Workshop on Principles of Advanced and Distributed Simulation.

[3]  Christopher D. Carothers,et al.  Scalable Time Warp on Blue Gene Supercomputers , 2009, 2009 ACM/IEEE/SCS 23rd Workshop on Principles of Advanced and Distributed Simulation.

[4]  J. D. Teresco,et al.  New challanges in dynamic load balancing , 2005 .

[5]  Anthony Ventresque,et al.  dSUMO: towards a distributed SUMO , 2013 .

[6]  Liam Murphy,et al.  Synchronisation for dynamic load balancing of decentralised conservative distributed simulation , 2014, SIGSIM PADS '14.

[7]  Katherine A. Yelick,et al.  Languages for High-Productivity Computing: the DARPA HPCS Language Project , 2007, Parallel Process. Lett..

[8]  Changjun Jiang,et al.  A fast adaptive load balancing method for parallel particle-based simulations , 2009, Simul. Model. Pract. Theory.

[9]  Xun Wang,et al.  Behavioral simulations in MapReduce , 2010, Proc. VLDB Endow..

[10]  W. Bossert,et al.  The Measurement of Diversity , 2001 .

[11]  Xueqi Cheng,et al.  Micro-Synchronization in Conservative Parallel Network Simulation , 2008, 2008 22nd Workshop on Principles of Advanced and Distributed Simulation.

[12]  J. Krebs,et al.  The Survival Value of Flocking in Birds: A Simulation Model , 1974 .

[13]  Toyotaro Suzumura,et al.  Highly Scalable X10-Based Agent Simulation Platform and Its Application to Large-Scale Traffic Simulation , 2012, 2012 IEEE/ACM 16th International Symposium on Distributed Simulation and Real Time Applications.

[14]  Liam Murphy,et al.  SParTSim: A Space Partitioning Guided by Road Network for Distributed Traffic Simulations , 2012, 2012 IEEE/ACM 16th International Symposium on Distributed Simulation and Real Time Applications.

[15]  Jack Dongarra,et al.  MPI: The Complete Reference , 1996 .

[16]  Bruce Hendrickson,et al.  Dynamic load balancing in computational mechanics , 2000 .

[17]  Karen D. Devinea,et al.  New Challenges in Dynamic Load Balancing , 2004 .

[18]  Michael Lees,et al.  Grid-based partitioning for large-scale distributed agent-based crowd simulation , 2012, Proceedings Title: Proceedings of the 2012 Winter Simulation Conference (WSC).

[19]  Tao Zou,et al.  Making time-stepped applications tick in the cloud , 2011, SoCC.