Caching in large-scale distributed file systems
Princeton University, 1993 - Computers - 178 pages
Abstract: "This thesis examines the problem of cache organization for very large-scale distributed file systems (DFSs). Conventional DFSs, based on the client-server model, suffer from bottlenecks when the total client load exceeds the server's capacity. Previous work has suggested that hierarchical client organizations can ameliorate the problem somewhat, but at the expense of a substantial increase in client latency. An analysis of existing DFS workloads reveals that there is considerable regularity in client file access patterns and that widely shared files lend themselves especially well to caching techniques. In particular, a large proportion of 'cache miss' traffic is for files that are already copied in another client's cache.
What people are saying - Write a review
We haven't found any reviews in the usual places.
File Access Patterns and Caching
4 other sections not shown
actual algorithms bottleneck bytes cache hit cache miss cached copies cached files caching schemes client and server client cache client performance client response client-driven invalidation client-server command computing cost DEC-SRC Trace disk distributed file systems dynamic hierarchies Ethernet Figure file access patterns file cache File Entropy file open file server file system activity file system tracing file transfer getattr hierarchical file system hit rate interface intermediate cache invalidation messages logged machines maintain megabytes miss rate miss traffic monitoring name cache nf strace NFS server number of clients number of file opens for read operating system Overall Client overhead messages overwritten packets Princeton and DEC-SRC Princeton Trace problem Prototype File System prototype implementation read and write reduce server load replacement rule requests rpcspy server daemon server traffic server-driven invalidation simple simulation results stateless stateless server static system call level thesis Total Server trace data trace-driven Unix workstation