I decided to use dbca because it's easier compared with srvctl command line.
Indeed when you use DBCA to create a service you do not need to add the corresponding TNS entry in your tnsnames.ora on the cluster nodes: DBCA does it all for you in the background.
Of course you can use SRVCTL to create a service, but you also need to manually add the TNS entry: moreover when you use SRVCTL, after you've created the new service you need to start it with a separate command, while DBCA starts it when you click FINISH button.
So let's start logging as oracle user from one of your nodes, and type
dbca
It first shows the Welcome window where you need to select your type of database: Oracle Real Application Cluster database or Oracle single instance database.
Next you need to choose the operation to be performed. Choose Services Management and click NEXT button.
At the step 2 select your database. In my case the screen shows my only BILLING database
The step 3 shows the current services available for the database chosen. In this screen you can add, remove or modify services. Today we want to add a new service so click the ADD button.
After clicking the ADD button you can choose a name for next service: in my case I called it SIM_TAF. Press OK button after you enter your service name.
Now configure your option for the new service. I choose it should have only a preferred instance where to run (BILLING3) and two available instances to switch over in case of problems (BILLING1 and BILLING2). Choose also your TAF (Transparent Application Failover) policy:
transparent application failover is a client-side feature that allows for clients to reconnect to surviving databases in the event of a failure of a database instance. Notifications are used by the server to trigger TAF callbacks on the client-side. Your options could be: Basic, in which the application connects to a backup node only after the primary connection fails. This approach has low overhead, but the end user experiences a delay while the new connection is created, or Pre-Connect, where the application simultaneously connects to both a primary and a backup node. This offers faster failover, because a pre-spawned connection is ready to use. But the extra connection adds everyday overhead by duplicating connections. For my environment Basic TAF policy was good. Click FINISH.
A summary screen will appear. Click OK
The configuration of the new service will start
Click NO if you don't want to perform other operation.
That's all.