RPC error and Checkpoint granularity?

Of late we have been facing issues with RPC timeouts across zones.

We have tweaked the keep alive time interval and have limited success for the issue to crop up intermittently.

Would enabling checkpoint and setting checkpoint granularity to File level help us in sustaining the connection in such cases where clients are in other zones and have large save sets...?

The logic I am applying here is as the checkpoints aren't enabled currently and in case of large save sets there would be a prolonged silence between the client and the backup server.

This may be the reason why the firewall may be breaking the connection. If there is a continuous or atleast periodic traffic like the checkpoints' communication the connection may be sensed as active rather than idle for a long time.

Please let me know your views....