Linear Programming-Based Affinity Scheduling of Independent Tasks on Heterogeneous Computing Systems

Resource management systems (RMS) are an important component in heterogeneous computing (HC) systems. One of the jobs of an RMS is the mapping of arriving tasks onto the machines of the HC system. Many different mapping heuristics have been proposed in recent years. However, most of these heuristics suffer from several limitations. One of these limitations is the performance degradation that results from using outdated global information about the status of all machines in the HC system. This paper proposes several heuristics which address this limitation by only requiring partial information in making the mapping decisions. These heuristics utilize the solution to a linear programming (LP) problem which maximizes the system capacity. Simulation results show that our heuristics perform very competitively while requiring dramatically less information.

[1]  Ali Sharifnia,et al.  Instability of the Join-the-Shortest-Queue and FCFS Policies in Queueing Systems and Their Stabilization , 1997, Oper. Res..

[2]  J. Dai.  On Positive Harris Recurrence of Multiclass Queueing Networks: A Unified Approach Via Fluid Limit Models , 1995 .

[3]  Jack Dongarra,et al.  Users' Guide to NetSolve v1.4.1 , 2002 .

[4]  Michael Mitzenmacher,et al.  The Power of Two Choices in Randomized Load Balancing , 2001, IEEE Trans. Parallel Distributed Syst..

[5]  Anthony A. Maciejewski,et al.  Heterogeneous Computing: Goals, Methods, and Open Problems , 2001, HiPC.

[6]  Michael Mitzenmacher,et al.  How Useful Is Old Information? , 2000, IEEE Trans. Parallel Distributed Syst..

[7]  Anthony A. Maciejewski,et al.  A Stochastic Approach to Measuring the Robustness of Resource Allocations in Distributed Systems , 2006, 2006 International Conference on Parallel Processing (ICPP'06).

[8]  Kevin D. Glazebrook,et al.  Dynamic routing to heterogeneous collections of unreliable servers , 2007, Queueing Syst. Theory Appl..

[9]  R. F. Freund,et al.  Scheduling resources in multi-user, heterogeneous, computing environments with SmartNet , 1998, Proceedings Seventh Heterogeneous Computing Workshop (HCW'98).

[10]  R. F. Freund,et al.  Dynamic matching and scheduling of a class of independent tasks onto heterogeneous computing systems , 1999, Proceedings. Eighth Heterogeneous Computing Workshop (HCW'99).

[11]  D. Yao,et al.  Fundamentals of Queueing Networks: Performance, Asymptotics, and Optimization , 2001, IEEE Transactions on Automatic Control.

[12]  Ian T. Foster,et al.  The Anatomy of the Grid: Enabling Scalable Virtual Organizations , 2001, Int. J. High Perform. Comput. Appl..

[13]  Sean P. Meyn,et al.  Stability and convergence of moments for multiclass queueing networks via fluid limit models , 1995, IEEE Trans. Autom. Control..

[14]  Nicholas Bambos,et al.  Optimal processor allocation to differentiated job flows , 2006, Perform. Evaluation.

[15]  Hong Chen.  Fluid Approximations and Stability of Multiclass Queueing Networks: Work-Conserving Disciplines , 1995 .

[16]  William Gropp,et al.  Beowulf Cluster Computing with Linux , 2003 .

[17]  Douglas G. Down,et al.  Linear Programming Based Affinity Scheduling for Heterogeneous Computing Systems , 2007, PDPTA.

[18]  Hui Li,et al.  Workload Characteristics of a Multi-cluster Supercomputer , 2004, JSSPP.

[19]  Anthony A. Maciejewski,et al.  Dynamic resource allocation heuristics that manage tradeoff between makespan and robustness , 2007, The Journal of Supercomputing.

[20]  Anthony A. Maciejewski,et al.  Measuring the Robustness of Resource Allocations in a Stochastic Dynamic Environment , 2007, 2007 IEEE International Parallel and Distributed Processing Symposium.

[21]  Douglas G. Down,et al.  MARO - MinDrift affinity routing for resource management in heterogeneous computing systems , 2007, CASCON.

[22]  Leonidas I. Kontothanassis,et al.  Profile driven scheduling for a heterogeneous server cluster , 2005, 2005 International Conference on Parallel Processing Workshops (ICPPW'05).

[23]  Robert Kyle Armstrong,et al.  Investigation of effect of different run-time distributions on SmartNet performance , 1997 .

[24]  David D. Yao,et al.  Fundamentals of Queueing Networks , 2001 .

[25]  R. F. Freund,et al.  SmartNet: a scheduling framework for heterogeneous computing , 1996, Proceedings Second International Symposium on Parallel Architectures, Algorithms, and Networks (I-SPAN'96).

[26]  J. Moreira,et al.  An Evaluation of Parallel Job Scheduling for ASCI Blue-Pacific , 1999, ACM/IEEE SC 1999 Conference (SC'99).

[27]  Kevin D. Glazebrook,et al.  Generalised ‘join the shortest queue’ policies for the dynamic routing of jobs to multi-class queues , 2003, J. Oper. Res. Soc..

[28]  Alexander L. Stolyar,et al.  OPTIMAL ROUTING IN OUTPUT-QUEUED FLEXIBLE SERVER SYSTEMS , 2005, Probability in the Engineering and Informational Sciences.

[29]  Sigrún Andradóttir,et al.  Dynamic Server Allocation for Queueing Networks with Flexible Servers , 2003, Oper. Res..

[30]  Anthony A. Maciejewski,et al.  Dynamically mapping tasks with priorities and multiple deadlines in a heterogeneous environment , 2007, J. Parallel Distributed Comput..

[31]  Yu-Tong He.  Exploiting Limited Customer Choice and Server Flexibility , 2007 .