The basic idea is simple. Every time a file is written, the changed portion of the
file(not the entire file) are also saved immediately across the network to another
system in a remote location. At any given time, there are always two live, active
copies of the file in two separate locations.
With the NFS and CIFS protocol as the transport layer, there is no need to re-invent
the wheel with a proprietary protocol.
Only one copy of software is needed, on the active system. The NFS infrastructure
on the remote system takes care of all replication operations from the active system.
The two systems can be at geographically-dispersed locations, separated by hundreds
The MFS works with any remote system and NAS that supports NFS and CIFS.
The MFS works with any type of local and remote storage devices.
NFS and CIFS check the syntax and semantics of every replication operation (create,
write, delete) to maintain the integrity of the file system.
Files on the remote system are ready to be used immediately.
The file systems on both systems are active. When failover occurs, there is no need
for file system check and mount operations.
Using file system like NFS and CIFS to do the replication, only changed portion of
a file as unit of block need to be replicated to remote system. This saves a lot
of system resources and network bandwidth.