I have a 400 user network spread across several branches across the U.S. We have two NOCs, one at Corporate and on at a DR facility. Corporate has over 100 employees while the branches have anywhere from 4 to 50. Corporate and DR have Cisco routers managed by our WAN provider. Branches have bridges in lieu of routers so as to reduce costs and take advantage of their "cloud firewall and routing" service. We use HP ProCurve switches throughout the org. Our WAN has two planes: Normal and Priority, where Priority is only for voice-related packets.
We use two VLANs, one for Voice traffic (200) and one for everything else (50). We use an Avaya IP Office setup with voice servers at Corporate and DR. Within the phone system configuration I can say, definitively, that all voice and signaling traffic is configured to be marked for Expedited Forwarding (46). All branch offices connect to the DR phone server and use Corporate's only for backup. All traffic from the voice servers to phones is properly marked for EF, including both voice (UDP/RTP) and Signaling (TCP). However, only voice (UDP/RTP) is being marked and prioritized properly. Signaling (TCP) is not being marked EF, though it should be. I know this because a Wireshark packet capture made at a ProCurve switch port for a specific phone confirms this.
I have seen the documentation specific to configuring HP ProCurve switches for interoperability with an Avaya phone system. I believe I have followed it properly. Still, question I have is this, "Do I have my switches configured properly to mark all traffic as EF for switch ports enabled the Voice VLAN 200?" I have reviewed so much information that I am starting to believe this might be a problem or even a bug within the phones or phone system itself. We are having sporadic problems with phones continuing to ring even after they've been answered and phones slow to get dial-tone, which all related to Signalling. I do have some other statements I implemented for attempting to prioritize traffic for certain IP addresses. It is not working properly and probably will not as our WAN provider doesn't have anything setup for such prioritization. Still, removing these entries does not remedy the problems I just described.
Here is one of my switch configs. They all look virtually identical with the except of unique identifiers. On this particular switch I am monitoring port 18 via port 1 with Wireshark so as to capture packets for a phone.
My HP ProCurve 2610 config (and yes, startup and running configs are identical and the switch was rebooted after last config modification):