Which statement describes Hybrid Cache behavior?

Which statement describes Hybrid Cache behavior?

• A. Cache data is replicated on each node
• B. Cache data is evenly distributed across all Confluence nodes in a cluster
• C. Cache data is replicated on each node and invalidated remotely by other nodes
• D. Cache data is invalidated locally only

Answer: (C)

In a Hybrid Cache in a Confluence cluster, data is kept on every node so reads can be served quickly from the local cache, while updates from one node trigger invalidation messages to the other nodes to ensure their caches don’t serve stale data. This combination—replication on each node plus remote invalidation—provides fast local access without sacrificing consistency across the cluster. That’s why the best description is: cache data is replicated on each node and invalidated remotely by other nodes. It captures both the local availability of cached data and the mechanism that keeps all nodes in sync when changes occur. The other ideas don’t fit as well because: replicating data on each node without remote invalidation can lead to stale results after updates; simply distributing cache data evenly describes distribution rather than maintaining coherence across the cluster; invalidating only locally would leave other nodes with potentially stale caches since they wouldn’t receive invalidation signals.

In a Hybrid Cache in a Confluence cluster, data is kept on every node so reads can be served quickly from the local cache, while updates from one node trigger invalidation messages to the other nodes to ensure their caches don’t serve stale data. This combination—replication on each node plus remote invalidation—provides fast local access without sacrificing consistency across the cluster.

That’s why the best description is: cache data is replicated on each node and invalidated remotely by other nodes. It captures both the local availability of cached data and the mechanism that keeps all nodes in sync when changes occur.

The other ideas don’t fit as well because: replicating data on each node without remote invalidation can lead to stale results after updates; simply distributing cache data evenly describes distribution rather than maintaining coherence across the cluster; invalidating only locally would leave other nodes with potentially stale caches since they wouldn’t receive invalidation signals.