Distributed File System prof. Giovanni Chiola

Availability


Files should be CONTINUOSLY accessible all the time (even in case of transient failures).
Client failures can be overcome by fast substitution.
Complete network failures can only be overcome by local caching on the client machine (detached operation).
Partial network or server failures can be overcome by redundancy and redirection of requests to currently available servers.

previous intro04.htm next