Distributed storage and parallel processing for large-size optical images
Nikolay Lvovich Kazanskiy, Sergey B. Popov · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2012
We present a distributed image approach to image data management in distributed storage and parallel processing systems. This approach allows to avoid parallel processing bottlenecks and overlap concerned with image data read/write. The distributed image is defined as a set of the overlap fragments. Their overlap size is determined by the requirements of fault tolerance and dynamic load balancing. Within this concept, we propose a decentralized method for dynamic load balancing of parallel computing and visualization solutions for large-size distributed images.