High performance pattern matching with dynamic load balancing on heterogeneous systems
Jin Hwan Park, B.A. Demirdag · 2006
This paper presents an efficient parallel-and-distributed string matching scheme in a heterogeneous environment. The contribution of the proposed approach is in two-fold; exploiting a non-trivial parallelism in each node that possibly is SMP, and distributed processing with dynamic load balancing on a network-connected system. For the parallel processing in each node, a multithreaded computation model is designed and implemented with Java Threads. For the distributed processing, a dynamic load balancing method is developed and implemented with Java RMI. The proposed approach manipulates both exact string matching and approximate string matching with k-mismatches. Experimental results demonstrate that the approach is an efficient parallel-and-distributed solution to the problem in a clustered system that is either homogeneous or heterogeneous.