Compare and contrast eager with lazy replication.

Question

What will be an ideal response?

Answer 1

In the Chapter 25 we examined protocols for updating data that worked on the basis that all
updates are carried out as part of the enclosing transaction. This was necessary because a
distributed transaction accesses different fragments on different sites; in other words, the updates
are immediately applied at every site. Atomicity is ensured by using the 2PC (two-phase commit)
protocol. The immediate propagation of updates in a replicated database is called eager or
synchronous update propagation. Eager update propagation ensures that all copies are updated
within the enclosing transaction and voting at the end ensures the atomicity.
An alternative mechanism to eager replication is called lazy or asynchronous update propagation.
With this mechanism, the target database is updated after the source database has been modified.
The delay in regaining consistency may range from a few seconds to several hours or even days.
However, the data eventually synchronizes to the same value at all sites (eventual consistency).
Although not all applications can cope with such a delay, it appears to be a practical compromise
between data integrity and availability that may be more appropriate for organizations that are able
to work with replicas that do not necessarily have to be always synchronized and current.