BYOC Explained: Use Your Own SIP Trunks

Ask a contact centre manager in Amman, Riyadh or Manama what their dialer costs and you will usually get two numbers: the seat price, and a per-minute charge nobody can fully explain. The second number is where bundled telecom lives. BYOC removes it.

This guide explains what bring your own carrier means operationally, what changes on the day you connect your own trunk, and what to check before you move a live campaign onto it. It is written for BPOs, collections teams and telecom resellers who already hold carrier relationships and do not want to give them up to buy a dialer.

What BYOC actually means

Most dialer platforms buy wholesale minutes from a carrier, mark them up, and sell them back to you as included minutes or platform usage. You never see the wholesale rate, you cannot negotiate directly, and if you find a better rate for a specific destination you cannot use it without changing platform.

BYOC inverts that. You enter your own SIP trunk credentials in the admin panel, the platform validates connectivity, codec support and authentication, and calls go out over your carrier at your rates. The platform charges a flat fee based on seats or usage tier. Carrier minutes are billed by your provider directly, and the platform takes no markup on them and has no visibility into them.

Any SIP-compatible provider works: BroadNet, a regional telco, a Tier 1 carrier, or your own telecom infrastructure. If it speaks SIP over UDP, TCP or TLS, it connects. There are no preferred carrier requirements and no limit on the number of trunks a tenant may hold.

Keeping the contracts you already signed

For a reseller or a BPO, the carrier relationship is often the business, not an expense line. It carries the rates, the routes to the destinations your clients actually call, and years of commercial history. A platform that requires you to abandon it is asking you to hand over your margin.

Under BYOC those agreements stay exactly as they are. You add trunks rather than replace suppliers, and you can hold several at once with routing rules that pick between them by destination prefix, cost, quality score, load balancing or time of day. Regional traffic can go through a regional carrier for local rates while other destinations go elsewhere, evaluated in priority order per call.

Local caller ID is part of BYOC, not separate from it

Owning the trunk is half the answer. The other half is the number you present on it. Caller IDs are assigned per trunk and per campaign, and numbers are grouped into named caller-ID pools with a selection mode chosen per pool: match the recipient's area code for local presence, round-robin, sequential or random. The dialer applies the mode at dial time, before the call is originated.

Two controls matter most for teams calling across several Gulf and Levant markets. Per-number daily call caps let you ramp a new number on your own schedule instead of pushing full campaign volume through it on day one. Deactivation takes a number out of rotation the moment you decide, and traffic continues across the rest of the pool. CLI verification checks ownership and registration status, so the number you present is one you are entitled to present.

In a multi-tenant deployment each tenant gets its own caller-ID pools with isolated rotation and mapping, which is what keeps one client's number strategy from affecting another's. Give every new tenant its own caller IDs rather than sharing a trunk identity across clients.

Two media nodes, and what happens when one goes

A carrier-grade trunk pointed at a single machine is still a single machine. Media in DialerBee runs on two FreeSWITCH media nodes. A SIP load balancer distributes calls round-robin across them, so carrier traffic does not depend on one node staying up, and browser-based agent audio is load balanced over WebRTC rather than piling onto whichever node answered first.

The agent desktop holds a WebSocket to a media node. If that node becomes unavailable, the client fails over to the other node at reconnect, so an agent comes back on the surviving node instead of waiting for the failed one. Readiness endpoints check the dependencies a service actually needs before it reports ready, and reconcilers correct what gets left behind: orphaned calls, stuck agents, and calls that were in flight across a restart.

Trunk-level failover sits beside node-level resilience. Trunk health is monitored in real time through ASR, ACD and PDD, and if a trunk fails or quality drops below the threshold you configured, calls move to your backup carrier without manual intervention.

What to check before you migrate a live campaign

Moving a running campaign onto your own trunk is not difficult, but it rewards a dry run. Work through this list with your carrier before the switch rather than during it.

• Signalling and transport: SIP over UDP, TCP or TLS, with SRTP if you encrypt media.
• Authentication: IP access lists, digest authentication, or TLS client certificates, and which one your carrier prefers.
• Codecs: G.711 in both variants, G.729 and Opus are negotiated per trunk. Confirm what your carrier supports on the destinations you call most.
• Routing rules: the prefixes, costs, quality scores and time-of-day conditions that decide which trunk a call takes, in priority order.
• Failover: which trunk is the backup, and the quality threshold that triggers the move.
• Caller ID: which pool each campaign uses, the selection mode, the daily cap per number, and proof of ownership for the numbers you present.
• Reporting: per-call carrier cost tracking and per-trunk comparison, so the saving is visible rather than assumed.

Then test at low volume on a real campaign for a day before you move the rest. Listen to the calls. A trunk that looks healthy on a dashboard and sounds poor to a customer is a trunk you have not finished testing.

Sizing the trunk for predictive dialing

A predictive campaign places several calls for every free agent, so trunk capacity has to be sized against your peak rather than your average. Fifty agents running at a three-to-one over-dial ratio can ask for a hundred and fifty concurrent channels at the top of the hour. Under-provision and calls simply fail at exactly the moment the campaign is busiest. Agree the channel count with your carrier at expected peak concurrency plus roughly twenty percent of headroom.

Calls per second is the limit people forget. Carriers commonly cap a trunk at somewhere between five and thirty new calls per second, and a dialer bursts hardest at campaign start and at shift login, which is when a whole team becomes available at once. Either raise the carrier limit or throttle the dialer's outbound rate to sit under it, and test the burst rather than the steady state. The full parameter list, including codecs and DTMF, is in the SIP trunk setup walkthrough.

Five failures that show up in the first week

Almost every BYOC problem reported after go-live is one of these five, and each is cheaper to find in a test call than in a client review.

• One-way audio. Nearly always NAT traversal or a firewall. Confirm the RTP port range is open and that the media nodes are reachable from the carrier's network in both directions.
• Recordings missing. The media path has bypassed the recording leg. Verify that a recording exists and plays back end to end before the first live campaign, not after the first dispute.
• Caller ID rejected. Carriers drop calls presenting a number that is not provisioned on the trunk. Every DID you intend to present has to be authorised on the carrier account first.
• Capacity surprises. Sized against average concurrency rather than peak, as above.
• Codec mismatch. If the trunk offers only what the carrier does not accept on that destination, calls fail cleanly and confusingly. Confirm the negotiated codec per destination during setup.

BYOC pays off fastest when you already hold carrier contracts at negotiated rates, run high monthly volume, need routing and caller-ID control at the carrier level, dial across several countries, or run tenants who bring their own trunks. A small team with no existing carrier relationship may reasonably start on a bundled arrangement and move to BYOC when volume justifies it, which does not require changing platform.

Who benefits most

Telecom resellers gain the most, because BYOC and white-label together let them run the platform under their own brand on their own carrier, per tenant, with isolated trunk configuration, billing and reporting. BPOs come next: different clients often require different carriers, and per-tenant trunks let a client's traffic stay on the carrier that client's contract names.

Collections and high-volume outbound teams benefit for a simpler reason. When per-minute cost is paid at negotiated carrier rates instead of a marked-up platform rate, the economics of a high-attempt campaign change, and routing by destination keeps them that way as the list mix shifts.

Frequently asked questions

What does BYOC mean?

Bring your own carrier. You connect your own SIP trunks to the dialer instead of buying minutes bundled into the platform, so you keep your carrier relationships and your negotiated rates and the platform adds no per-minute markup.

Which SIP providers can I connect?

Any SIP-compatible provider: BroadNet, regional telcos, Tier 1 carriers, or your own telecom infrastructure. If it speaks SIP over UDP, TCP or TLS, it connects. There are no provider restrictions and no preferred carrier requirement.

How does DialerBee charge if there is no per-minute fee?

The platform is priced per agent, with no per-minute platform fees. Carrier minutes are billed by your own provider at your own rates. The platform has no visibility into or markup on your carrier costs.

What happens if a trunk goes down mid-campaign?

Trunk health is monitored in real time through ASR, ACD and PDD. If a trunk fails or quality drops below the threshold you set, calls fail over to your backup carrier without manual intervention, and agents keep dialling.

Can each client have its own carrier?

Yes. Trunk configuration, routing rules, caller-ID assignment, billing and reporting are per tenant, so a BPO or a reseller can route each client through the carrier that client's contract names, under their own brand.

Does BYOC change how caller ID works?

It puts it in your hands. Caller IDs are assigned per trunk and per campaign from named pools, with local presence, round-robin, sequential or random selection, per-number daily caps, manual deactivation and ownership verification on the numbers you present.

Related reading

• Bring Your Own Carrier
• Platform Reliability
• Caller-ID Pool Control
• White-Label Platform