The above config assumes the voice and data are in the same clocking domain. If the two services come from the same provider, then they generally are.
If they come from different vendors, we quite often see clock slips on one of the interfaces due to the sligh timing differences. In this case, you have to decide which interface can handle the errors better - in many cases you have to put up with what are essentially controlled slips on the data interface (since the majority of data apps are generally more resilient to the odd error that voice/fax/modem services).
The later model VWIC2 cards actually allow data interfaces to run independantly to a voice port, so the router backplane derives clocking from the voice interface and the data interface is in it's own local clocking domain.
Refer here for some more details -
http://www.cisco.com/en/US/docs/ios/12_2/12_2z/12_2zj/feature/guide/gthwecan.html#wp1068593
The independent keyword expands on the clock source internal and clock source line to specify that the port can operate on an independent clocking domain. Port 0 is the default primary clock source, and port 1 is the default secondary clock source and is loop-timed. With independent clocking enabled, this dependency no longer exists, so the independent keyword means that both ports can be independently clocked.
Note When independent clocking is configured, the controller will support only one channel group and no voice applications.
If more than one channel group is configured, the following error message occurs:
channel-group 2 timeslots 3
%Channel-group already created.
%Only 1 channel-group can be configured with
independent clocking.
%Insufficient resources to create channel
group
When you are configuring the clock source independent and no clock source independent, the channel group has to be removed from this configuration.
