The availability of resources in a distributed environment is an important factor to be considered when designing a distributed system. Replication is a well-known technique to achieve fault tolerance in distributed systems, thereby enhancing availability. But some efficient techniques are required to keep all the replicas consistent i.e every replica should have the same copy of the data. When a user changes or does any modification, the similar modification has to be done on every other replica for the purpose of achieving consistency. Proposed method uses write through mechanism i.e. whenever the client make any modification on the file, that modification will be immediately transmitted to the server, so the server can also perform similar modification on the server copy of the file and then asynchronous replication is followed i.e whenever any modification is done, first the primary copy is updated and new version number is added with the dirty bit marked as 0. This shows that the data is a valid data. A threshold is set between subsequent modification and once the threshold is reached, then the updates will be propagated to other replicas also. It ensures data consistency from the viewpoint of client.
[1]
Irene Zhang.
Reducing the Frequency of Data Loss in Cloud Storage
,
2013
.
[2]
Li Li,et al.
Adaptive Replica Synchronization for Distributed File Systems
,
2015,
IEEE Systems Journal.
[3]
GhemawatSanjay,et al.
The Google file system
,
2003
.
[4]
Yutaka Ishikawa,et al.
Partial Replication of Metadata to Achieve High Metadata Availability in Parallel File Systems
,
2012,
2012 41st International Conference on Parallel Processing.
[5]
Sachin Katti,et al.
Copysets: Reducing the Frequency of Data Loss in Cloud Storage
,
2013,
USENIX Annual Technical Conference.
[6]
Sushil Jajodia,et al.
An adaptive data replication algorithm
,
1997,
TODS.