What are different ways you can configure WINS replication?

Answer 1

WINS replication can be configured in several ways.

Each server has the following replication options:
push : send records to another server
pull : get records from another server
push/pull : send records to AND get records from another server

If you manage all WINS servers and clients then you should use push/pull replication exclusively.
If there are other networks outside of your control then you may wish to configure your server to push records to their servers but not pull any records. This should ensure that your records are not updated or added to by any of their details. Note that the replication partners must be configured correctly on both servers.

With regard to replication topology, you really have two options:
chain or hub and spoke

Consider four servers: A, B, C managed by you and server D managed by Bob who is a bit of a cowboy. You do not want Bob to be able to affect your data but you manage the finance servers and his users need to access them.

Chain replication might look like:

A <--push/pull--> B <--push/pull--> C --push--> D
(From Bob's perspective, it is just C --pull--> D)

Hub/Spoke
One server acts as a hub with the others as spokes.
This will reduce replication time in a large network.

C <--push/pull--> A
C <--push/pull--> B
C --push--> D

(Bob's configuration is the same is both cases)


Most important:
Do not configure replication in a loop - this will potentially cause big problems.
e.g. If A <-> B <-> C then do not configure C <-> A
OR if A <-> B and A <-> C then do not configure B <-> C

Answer 2

Replication overview
Where multiple WINS servers are used on your network, they can be configured to replicate records in their databases to other servers. This process is shown in the following figure. Two WINS servers, WINS-A and WINS-B, are both configured to fully replicate their records with each other.