Closed bryk closed 11 years ago
After deep considerations I've decided that "number of replicas per file (to simulate S3 like services)" point wasn't the best idea. Why? Because we want to add congestion to mostly simulate NFS-like storages. In those types of storage there's no per-file congestion.
Done.
I will implement congestion model in the Global Storage Manager. It will contain two parameters to simulate congestion: