Distributed File System prof. Giovanni Chiola

Detached Modify


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