Efficient Load Balancing on Irregular Network Topologies Using B+tree Structures

In this study, a heterogeneous distributed computing environment (Grid) is employed as a computing platform to perform some computationally intensive tasks. In order to increase the efficiency of the system (utilization and average response time), a dynamic task scheduling algorithm is proposed to balance the load among the nodes of the system. The technique is composed of two main phases. Firstly, the heterogeneous distributed system is embedded onto a structure similar to B+tree and then the algorithm of load-balancing is executed on the virtual structure. The optimum degree of the tree is identified in order to achieve optimal performance. The experimental and simulation results show the benefit of using the technique and additionally they show that it is relatively easy to identify the crossover points (thresholds) for which the system should be load balanced.

[1]  Jaafar Gaber,et al.  Dynamic and Randomized Load Distribution in Arbitrary Networks , 1998, Euro-Par.

[2]  Imtiaz Ahmad,et al.  An Integrated Technique for Task Matching and Scheduling onto Distributed Heterogeneous Computing Systems , 2002, J. Parallel Distributed Comput..

[3]  Keqin Li.  On the Performance of Randomized Embedding of Reproduction Trees in Static Networks , 2004, International Journal of Parallel Programming.

[4]  Vittorio Scarano,et al.  Embedding Graphs onto the Supercube , 1995, IEEE Trans. Computers.

[5]  Mohand Tahar Kechadi,et al.  Topology Preserving Dynamic Load Balancing for Parallel Molecular Simulations , 1997, ACM/IEEE SC 1997 Conference (SC'97).

[6]  Wentong Cai,et al.  How network topology affects dynamic loading balancing , 1996, IEEE Parallel Distributed Technol. Syst. Appl..

[7]  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..

[8]  Andrew S. Tanenbaum,et al.  Distributed systems: Principles and Paradigms , 2001 .

[9]  David S. Johnson,et al.  Computers and Intractability: A Guide to the Theory of NP-Completeness , 1978 .

[10]  Mark Baker,et al.  Cluster Computing White Paper , 2000, ArXiv.

[11]  M. N. Shanmukha Swamy,et al.  Graphs: Theory and Algorithms , 1992 .

[12]  Rudolf Bayer,et al.  Organization and maintenance of large ordered indexes , 1972, Acta Informatica.

[13]  Eunseuk Oh,et al.  Hyper-Star Graph: A New Interconnection Network Improving the Network Cost of the Hypercube , 2002, EurAsia-ICT.

[14]  Debasish Ghose,et al.  Scheduling Divisible Loads in Parallel and Distributed Systems , 1996 .

[15]  Douglas Comer,et al.  Ubiquitous B-Tree , 1979, CSUR.

[16]  Thomas L. Casavant,et al.  A Taxonomy of Scheduling in General-Purpose Distributed Computing Systems , 1988, IEEE Trans. Software Eng..

[17]  M. N. S. Swamy,et al.  Graphs: Theory and Algorithms: Thulasiraman/Graphs , 1992 .

[18]  Guy E. Blelloch,et al.  Prefix sums and their applications , 1990 .

[19]  Keqin Li.  A Comparative Performance Evaluation of Randomized Tree Embedding in Hypercubic Networks , 1998 .

[20]  Bruce S. Davie,et al.  Computer Networks: A Systems Approach , 1996 .

[21]  M. Tahar Kechadi,et al.  Dynamic task scheduling for irregular network topologies , 2005, Parallel Comput..

[22]  R. Bayer,et al.  Organization and maintenance of large ordered indices , 1970, SIGFIDET '70.

[23]  Rudolf Bayer,et al.  Binary B-trees for virtual memory , 1971, SIGFIDET '71.