Data to Voice inter-VLAN routing is only required for softphones since the RTP media must make it over to the physical phone (and back). All other features are client to server in nature: Deskphone Control is CTI to CUCM, IM&P is XMPP to CUP, VVM is to Unity Connection, etc.
If your corporate security policy forbids easy access to the Voice VLAN you can use a Trusted Relay Point for the softphones. The TRP is just an MTP which acts as a proxy for the RTP media. It can be combined with IOS ZBFW to open pin holes for RTP media; or, be as simple as being permitted in an ACL. A second advantage to the TRP is that it can apply appropriate QoS markings if you do not want to trust or mark at the access layer for the Data VLAN.
One footnote: When Jabber adds CUVA-style deskphone video you would also need inter-VLAN routing access for the Cisco Audio Stream Tunnel (CAST) protocol and video RTP traffic. CAST works by seeing the IP of the phone through the PC's CDP driver and then connecting to the phone. When the phone makes/receives a call it uses CAST to get video SDP into the call but does not proxy the media. The CAST client (e.g. Jabber) sends/receives the video RTP from the Data VLAN while the phone sends/receives the audio from the Voice VLAN. TRP would only work here if you applied it to the deskphone which likely isn't desirable.
Please remember to rate helpful responses and identify helpful or correct answers.
