A DAQM-Based Load Balancing Scheme for High Performance Computing Platforms

This paper addresses the load balancing problem, which is one of the key issues in high-performance computing (HPC) platforms. A novel method, called decentralized active queue management (DAQM), is proposed to provide a fair task distribution in a heterogeneous computing environment for HPC platforms. An implementation of the DAQM is presented, which consists of an ON-OFF queue control and a utility maximization-based coordination scheme. The stability of the queue control scheme and the convergence of the algorithm for utility maximization have been assessed by rigorous analysis. To demonstrate the performance of the developed queueing control system, numerical simulations are carried out and the obtained results confirm the efficiency and viability of the developed scheme.

[1]  Norman T. J. Bailey,et al.  A Continuous Time Treatment of a Simple Queue Using Generating Functions , 1954 .

[2]  David M. C. Wishart,et al.  Queuing Systems in Which the Discipline is “Last-Come, First-Served” , 1960 .

[3]  James R. Jackson.  Queues with Dynamic Priority Discipline , 1961 .

[4]  Peeter Andrus Kivestu,et al.  Alternative methods of investigating the time dependent M/G/k queue , 1976 .

[5]  Andrzej Duda.  Transient diffusion approximation for some queuening systems. , 1983, SIGMETRICS '83.

[6]  D. Daley.  Certain optimality properties of the first-come first-served discipline for G/G/s queues , 1987 .

[7]  David Fernández-Baca,et al.  Allocating Modules to Processors in a Distributed System , 1989, IEEE Trans. Software Eng..

[8]  Offer Kella.  The threshold policy in the M/G/1 queue with server vacations , 1989 .

[9]  S. Sharma,et al.  Approximate models for the study of nonstationary queues and their applications to communication networks , 1993, Proceedings of ICC '93 - IEEE International Conference on Communications.

[10]  QUTdN QeO,et al.  Random early detection gateways for congestion avoidance , 1993, TNET.

[11]  Walter Willinger,et al.  Experimental queueing analysis with long-range dependent packet traffic , 1996, TNET.

[12]  L. Flatto.  The waiting time distribution for the random order service $M/M/1$ queue , 1997 .

[13]  Stephen Taylor,et al.  A Practical Approach to Dynamic Load Balancing , 1998, IEEE Trans. Parallel Distributed Syst..

[14]  Debra A. Hensgen,et al.  The relative performance of various mapping algorithms is independent of sizable variances in run-time predictions , 1998, Proceedings Seventh Heterogeneous Computing Workshop (HCW'98).

[15]  Philip S. Yu,et al.  Dynamic Load Balancing on Web-Server Systems , 1999, IEEE Internet Comput..

[16]  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).

[17]  Ladislau Bölöni,et al.  A comparison study of static mapping heuristics for a class of meta-tasks on heterogeneous computing systems , 1999, Proceedings. Eighth Heterogeneous Computing Workshop (HCW'99).

[18]  O. Brun,et al.  Analytical solution of finite capacity M/D/1 queues , 2000, Journal of Applied Probability.

[19]  Andreas Pitsillides,et al.  Congestion control for differentiated-services using non-linear control theory , 2001, Proceedings. Sixth IEEE Symposium on Computers and Communications.

[20]  Donald F. Towsley,et al.  On designing improved controllers for AQM routers supporting TCP flows , 2001, Proceedings IEEE INFOCOM 2001. Conference on Computer Communications. Twentieth Annual Joint Conference of the IEEE Computer and Communications Society (Cat. No.01CH37213).

[21]  FABRICE GUILLEMIN,et al.  Analysis of the M/M/1 Queue with Processor Sharing via Spectral Theory , 2001, Queueing Syst. Theory Appl..

[22]  Ladislau Bölöni,et al.  A Comparison of Eleven Static Heuristics for Mapping a Class of Independent Tasks onto Heterogeneous Distributed Computing Systems , 2001, J. Parallel Distributed Comput..

[23]  Albert Y. Zomaya,et al.  Observations on Using Genetic Algorithms for Dynamic Load-Balancing , 2001, IEEE Trans. Parallel Distributed Syst..

[24]  Donald F. Towsley,et al.  Fixed point approximations for TCP behavior in an AQM network , 2001, SIGMETRICS '01.

[25]  E. Joseph Metzger,et al.  Real-time ocean modeling systems , 2002, Comput. Sci. Eng..

[26]  Y. Tipsuwan,et al.  Control methodologies in networked control systems , 2003 .

[27]  Selim G. Akl,et al.  Scheduling Algorithms for Grid Computing: State of the Art and Open Problems , 2006 .

[28]  U. Narayan Bhat,et al.  An Introduction to Queueing Theory: Modeling and Analysis in Applications , 2006 .

[29]  Bharadwaj Veeravalli,et al.  Design and Performance Evaluation of Queue-and-Rate-Adjustment Dynamic Load Balancing Policies for Distributed Networks , 2006, IEEE Transactions on Computers.

[30]  Zhong-Ping Jiang,et al.  Network flow control under capacity constraints: A case study , 2006, Syst. Control. Lett..

[31]  Daniel Pérez Palomar,et al.  Alternative Distributed Algorithms for Network Utility Maximization: Framework and Applications , 2007, IEEE Transactions on Automatic Control.

[32]  Stephen P. Boyd,et al.  Subgradient Methods , 2007 .

[33]  Gustavo de Veciana,et al.  Dynamic association for load balancing and interference avoidance in multi-cell networks , 2007, IEEE Transactions on Wireless Communications.

[34]  Archana Ganapathi,et al.  Analysis and Lessons from a Publicly Available Google Cluster Trace , 2010 .

[35]  Xiandong Meng,et al.  A High-Performance Heterogeneous Computing Platform for Biological Sequence Analysis , 2010, IEEE Transactions on Parallel and Distributed Systems.

[36]  Hong He,et al.  Task assignment in heterogeneous computing systems using an effective iterated greedy algorithm , 2011, J. Syst. Softw..

[37]  Robert A. van de Geijn,et al.  Unleashing the high-performance and low-power of multi-core DSPs for general-purpose HPC , 2012, 2012 International Conference for High Performance Computing, Networking, Storage and Analysis.

[38]  Sven Leyffer,et al.  Heuristic static load-balancing algorithm applied to the fragment molecular orbital method , 2012, 2012 International Conference for High Performance Computing, Networking, Storage and Analysis.

[39]  T. K. Ghosh,et al.  Load balanced static grid scheduling using Max-Min heuristic , 2012, 2012 2nd IEEE International Conference on Parallel, Distributed and Grid Computing.

[40]  Randy H. Katz,et al.  Heterogeneity and dynamicity of clouds at scale: Google trace analysis , 2012, SoCC '12.

[41]  Richelle V. Adams,et al.  Active Queue Management: A Survey , 2013, IEEE Communications Surveys & Tutorials.

[42]  Stefano Grivet-Talocia,et al.  High-Performance Passive Macromodeling Algorithms for Parallel Computing Platforms , 2013, IEEE Transactions on Components, Packaging and Manufacturing Technology.

[43]  Kenli Li,et al.  Energy-Efficient Stochastic Task Scheduling on Heterogeneous Computing Systems , 2014, IEEE Transactions on Parallel and Distributed Systems.

[44]  Mohamed Faten Zhani,et al.  Dynamic Heterogeneity-Aware Resource Provisioning in the Cloud , 2013, IEEE Transactions on Cloud Computing.

[45]  Giovanni Giambene.  Queuing theory and telecommunications , 2014 .

[46]  Deokho Kim,et al.  Dynamic Load Balancing of Parallel SURF with Vertical Partitioning , 2015, IEEE Transactions on Parallel and Distributed Systems.

[47]  Anthony A. Maciejewski,et al.  Scalable linear programming based resource allocation for makespan minimization in heterogeneous computing systems , 2015, J. Parallel Distributed Comput..

[48]  Yichuan Jiang,et al.  A Survey of Task Allocation and Load Balancing in Distributed Systems , 2016, IEEE Transactions on Parallel and Distributed Systems.

[49]  Xin Chang,et al.  Suitability Analysis of FPGAs for Heterogeneous Platforms in HPC , 2016, IEEE Transactions on Parallel and Distributed Systems.