Distributed File System | prof. Giovanni Chiola | ![]() |
Possible only if old file blocks are available locally and enough space is available in the local disk to store all new file blocks. | ||
Single point of failure (the client machine), so fault tolerance is achieved only after storage on servers of the new version. | ||
Risk of inconsistency in case of network partitioning. Semiautomatic sync probably required. | ||
| ||
previous | cache05.htm | next |