A Fault-Tolerant File Management Algorithm in Distributed Computer System “THUDS”

A concurrent control with independent processes from simultaneous access to a critical section is discussed for the case where there are two distinct classes of processes known as readers and writers. The readers can share the file with one another, but the interleaved execution with readers and writers may produce undesirable conflicts. The file management algorithm proposed in this paper is the activity of avoiding these results. This algorithm not only guarantees the consistency and integrity of the shared file, but also supports optimal parallelism. The concept of dynamic virtual queue is introduced and serves the foundation for this algorithm. Our algorithm with its implicit redundancy allows software fault-tolerant technique.