To configure this replication to be a P2P one, open the Properties dialog for the replication and go to the Subscription Options screen. Select True for the Allow P2P Subscriptions value and save the settings. Now, when you right-click the replication name you'll see a new option-Configure P2P Topology. Select this to bring up P2P configuration wizard.

Continue in the wizard till you reach a topology configuration screen. You can start adding nodes in this screen for P2P. Right click an empty part of the screen and select Add a new Peer Node. Select a database server (with right permissions) and then select the database name to use for replication. Make sure that you increment the Peer Originator ID for each node.

Once all the nodes are in place, right click any node and select Add a new Peer Connection. This brings up a rubber-banding arrow that moves with your mouse and attaches to other nodes.

Click on a node to create a P2P replication between that and the origin. Perform this for each node so that they talk to each other. You can also speed this up by selecting Connect to all displayed nodes which creates the topology for a server to all others. To create a mesh, perform this step with all nodes.

Continue on and enter the appropriate security information for access to each node's database. Once done, complete the wizard and wait till the topology is actually setup. Now right click Replication in any server and bring up Replication Monitor. This will show you whether the replication is correctly setup or not. To see the effect of replication, make some changes (insert, updates, deletes) in any table that you selected for replication in any server. Very soon you'll see the same change occuring in all other nodes as well.

Some very important things to keep in mind while setting up P2P replication.

1. P2P does not use Snapshots to create the 'base' image for databases across nodes. It is up to you to ensure that the databases start off with identical schemas and data before you setup the replication.

2. Primary conflicts can occur here. You need to ensure that if a table has data getting inserted at any node, primary keys are unique across nodes. This can be done by:

a. Using GUIDs as PKs
b. Using composite keys with one part of the key being unique to the server
c. Using Identity columns with start and step values to ensure no conflicts. For instance, one on server, the identity field can be 1+2 (so that it uses 1, 3, 5...) and on another it uses 2+2 (so that it uses 2, 4, 6 ...).

3. P2P is available only in the Enterprise Edition of SQL Server 2008.

Once you setup P2P it becomes very easy to manage data changes across multiple locations. This is a great new feature of SQL Server 2008 and can make the life of a database admin much better.