Three IP addresses are recommended considering load balancing and high availability requirements regardless of the number of servers in the cluster. The IP addresses must be on the same subnet as your public network in the cluster.
Thanks for sharing such a valuable link, but still i did not get my answer why scan VIP it limited to max. 3. Why not it 4,5,6 or some other number ? why it is 3 ?
Why not do the same question differently.
The Clusterware need more than 3 SCAN IP to provide high availability and load balancing? If you understand the concept and role of SCAN I believe you will know the right answer.
For all SCAN Listener going down is necessary to fails the three nodes where are registered SCAN Listener at the same time. Even if this happens, the CRS should take less than one minute to register the the SCAN Listener on the nodes that remain active. (SCAN is a small piece of a large system that is already prepared for failures)
Remember that SCAN IP does not establish connections with Oracle RAC, but manages and redirects connections to the nodes of the cluster. The connections are established using the Virtual IP (VIP). This means that if you stop all SCAN Listener/IP all connections that are already established are not affected, because they are using the VIP.
Oracle version 11.2 use up to three IP, because the engineering team studied and understood that to ensure high availability and load balance is necessary 3 SCAN IP. (just it)
It may be that this will change in the future but for now understand that 3 is enough to guarantee Grid Computing.
We could ask another question, how many DNS servers (cluster / failover) you have in your environment to provide HA. If your answer is up to three nodes, what happens if the three DNS servers fail at the same time as it is a point of failure (SCAN/VIP works with DNS).
This type of question leads us to Grid Computing.
One way to think about grid computing is as the virtualization and pooling of IT resources—compute power, storage, network capacity, and so on—into a single set of shared services that can be provisioned or distributed, and then redistributed as needed.
Just as an electric utility uses a grid to deal with wide variations in power demands without affecting customer service levels, grid computing provides IT resources with levels of control and adaptability that are transparent to end users, but that let IT professionals respond quickly to changing computing workloads.
The entire infrastructure must be prepared to support failure. If your environment is designed to support multiple types of failures you will see that IP Scan 3 is enough.