Data replication is the process of storing data at multiple sites to improve availability. The data contained at one server is replicated at multiple nodes to make it consistently available in case a node is compromised. This results in a distributed database.
It is essential that the replica is in a consistently updated state and that it is fully synchronized with the source, especially when frequent transactions are involved.
There are two kinds of replication techniques:
As the name suggests, full replication involves replicating the entire database at all sites of the distributed system. This drastically improves data availability, but all the sites need to be constantly updated so that only updated data is available to every user.
In this replication technique, only a portion of the database is replicated while the other portion is left out. The size of the portion to be replicated at different sites depends on the size of the distributed system. This does reduce the amount of data to be replicated and updated, but it also means that there are fewer sites (than full replication) that hold some particular data (data availability is compromised).
View all Courses