Research on the Fast Parallel Recomputing for Parallel Digital Terrain Analysis

With the rapid increasing of spatial data resolution, the high volume datasets make the geocomputation become more time-consuming especially operating some complex algorithms, i.e. viewshed analysis and drainage network extraction. Parallel computing is regarded as an efficient solution by utilizing more computing resource, which has also been proved its availability in digital terrain analysis according to many published literatures. Among them, the stable and credible services play an irreplaceable role in the high performance computing, especially when a failure occurs in the large-scale processing. However, litter research focuses on this issue. In this paper, a new approach for the parallel digital terrain analysis considering the performance of fault tolerance, named the Fast Parallel Recomputing (FPR) was proposed. FPR owns fast self-recovery ability. Once some failures are detected, all the surviving processes repartition the data block and recompute the sub-blocks in parallel to improve the efficiency of failure recovery. The experiments show that the proposed the FPR method achieves better performance than the traditional checkpointing method.