Fine-tuning the Replication Intervals may save some bandwidth on WAN links. Figure 8.10 illustrates a network configuration based on the network configuration first illustrated in Figure 8.6. The network configuration in Figure 8.6 was changed to the configuration shown in Figure 8.10 to illustrate one possible method for reducing convergence time.
Figure 8.10 Reducing WINS Server Replication Convergence Time
By keeping the pull Replication Intervals between WINS-C and WINS-B short (15 minutes), WINS servers A, B, and C can be reasonably synchronized. Replicas are never pulled in twice; only replicas with a higher version ID are copied. When WINS-B has an entry directly from WINS-C, it does not pull that entry from WINS-A.
However, under the example configuration there is the chance that both WINS-D and WINS-B can pull replicas from WINS-C by using the link between WINS-B and WINS-C. The resultant load on the link between WINS-B and WINS-C would increase.
In the example, this problem can be avoided if WINS-D is configured to pull replicas from WINS-B first and then check WINS-C. The pull Replication Interval between WINS-D and WINS-C would typically be the same 12 hours. Remember to configure push Update Counts (on WINS-D and WINS-C) to correspond to the 12 hours pull Replication Interval; otherwise unexpected replication is triggered by the Update Count threshold and not by the pull Replication Interval.