Freeze one failure and its ordered path
Record an authorized attempt's timestamp and timezone, direction, country, site, access type, trunk, channel inventory, SBC or PBX, protected calling and called numbers, Call-ID, first SIP response or timeout and media result. Preserve one closely matched successful control where available.
Do not replace certificates, broaden firewall policy, reroute numbers or add channels until the first missing boundary is known. Portal inventory, transport reachability, accepted signaling, routing, answer and media are different states.
Prove access and the customer demarcation
Identify whether the failing site uses customer Internet access or ordered BVPN, Satellite or SD-WAN access, then confirm the current demarcation and SBC placement. Retain link, address, route, quality and loss evidence in both directions.
A working data path does not prove voice quality or SIP reachability. Conversely, a channel or trunk alarm does not establish that the underlying access is down. Compare the current order and Orange incident record with the exact site rather than another location's topology.
Follow certificate, TLS and SIP state separately
On an Internet-access design, inspect the customer-managed SBC identity and the certificate supplied to Orange: name, chain, validity, key ownership and rotation history. Preserve the network, TCP, TLS and SIP results as separate timestamps. Do not expose private keys or API tokens.
Then follow OPTIONS or the authorized INVITE through the expected peers, transport, number format and first final response. A successful certificate upload or portal save is not evidence that the current SBC presents that certificate or that Orange accepts the call.
Distinguish capacity from routing rejection
Compare active calls with the ordered Voice Channels and the selected access commitment. One channel carries one call, but a no-channel condition, call-start limit, access impairment, identity rejection and unreachable destination can present different SIP outcomes. Retain the first provider and PBX response instead of increasing capacity from a symptom alone.
After a bounded correction, retest one normal call and the relevant capacity boundary. Do not use an emergency call to exercise load or failover.
Trace country-specific numbers and emergency routing
Confirm that the number is active or ported for the exact country and site, then retain delivered Request-URI, asserted identity, PBX normalization and final destination. For outbound failure, prove dial-plan classification, allowed calling scope, presented identity and Orange result.
Emergency service remains country and order specific. Verify the address or zone, number-to-location mapping, route, ANI, callback and provider-approved validation method without placing an unscheduled live emergency call.
Accept signaling, media and restoration independently
For the corrected path, retain provisional and final SIP responses, SDP offer and answer, negotiated codec and DTMF, media addresses, packets and decoded audio in both directions. If fax or encrypted media is ordered, test that exact workflow rather than inferring it from voice.
Recheck portal or API inventory, incident closure, new-call failover and restoration against the signed design. Remove temporary traces and access changes. No Orange trunk, access path, call, media stream, failover event or emergency workflow was validated for this diagnostic.
Sources & applicability
Primary references for the technical details above. Operational examples and planning checklists are VoIP.info editorial guidance.
- Orange Business Talk ↗
- Business Talk global service description — January 2026 ↗
- Business Talk and Contact Center service documents ↗
- Business Talk Digital API reference ↗
Examples require adaptation to your topology. No live PBX or hardware testing is claimed.