Trunking and Carrier Interconnection

Trunking and Carrier Interconnection

Every telephone call that crosses carrier boundaries — which is most calls — requires an interconnection between the originating and terminating carrier. This article covers how carriers connect to each other, the types of trunking used to carry voice traffic, and how the interconnection framework shapes the routing decisions visible in NPA/NXX data.

What Is a Trunk?

A trunk is a communication path between two switching systems. In traditional telecom, a trunk was a physical circuit — a copper pair or a channel within a T1 line — dedicated to carrying one voice call. In modern networks, a trunk is a logical connection between two SIP endpoints capable of carrying multiple simultaneous calls over a shared IP link.

The term persists from the telephone industry’s earliest days, when operators physically connected calls using trunk lines between switchboards.

TDM Trunking

Time-Division Multiplexing (TDM) trunking is the legacy method, still in use but declining.

T1/PRI

A T1 line carries 24 channels at 1.544 Mbps. Each channel is a 64 kbps DS0 capable of carrying one voice call. When used for voice with ISDN signaling, a T1 is configured as a PRI (Primary Rate Interface): 23 voice channels (B channels) plus one signaling channel (D channel).

PRIs connect PBX systems to carriers and carriers to each other. Signaling uses Q.931 (ISDN) on the D channel, or SS7 ISUP on dedicated signaling links for carrier-to-carrier trunks.

DS3/T3

A DS3 carries 672 channels (28 T1s) at 44.736 Mbps. Used for high-capacity carrier-to-carrier interconnections. Increasingly rare as IP takes over.

Advantages of TDM

  • Dedicated bandwidth per call — no contention or packet loss
  • Extremely well-understood technology with decades of operational experience
  • Deterministic latency

Disadvantages of TDM

  • Expensive — dedicated circuits must be provisioned whether used or not
  • Inflexible — adding capacity means ordering more circuits (weeks or months lead time)
  • Geographically constrained — requires physical circuit between endpoints

SIP Trunking

SIP trunking replaces TDM circuits with SIP-based connections over IP networks. Instead of 23 fixed voice channels on a PRI, a SIP trunk can carry as many simultaneous calls as bandwidth allows.

How SIP Trunks Work

A SIP trunk is a logical connection between a customer’s SBC or PBX and a carrier’s SBC. Calls are set up using SIP INVITE messages, and voice audio flows as RTP packets. There is no dedicated circuit per call — bandwidth is shared and allocated dynamically.

Advantages of SIP Trunking

  • Cost: Typically 30-60% cheaper than equivalent PRI capacity
  • Flexibility: Add or remove call capacity instantly, no circuit provisioning
  • Geographic independence: SIP trunks can terminate anywhere with IP connectivity
  • Feature-rich: Supports HD voice, video, multiple codecs, and advanced routing
  • Consolidation: A single IP connection can replace dozens of PRIs

Carrier SIP Interconnection

Between carriers, SIP interconnection happens at the SBC level. Two carriers establish a SIP peering relationship by:

  1. Exchanging technical parameters (IP addresses, supported codecs, capacity)
  2. Configuring SBCs to send and receive SIP traffic
  3. Establishing routing tables mapping NPA/NXX destinations to the peer
  4. Signing an interconnection agreement covering rates, quality metrics, and terms

Interconnection Models

Direct Peering

Two carriers connect their networks directly — either physically at a colocation facility or virtually over a dedicated IP link. Direct peering provides the best call quality (fewest hops) and usually the best rates.

Large carriers (AT&T, Verizon, T-Mobile, Lumen) have extensive direct peering arrangements with each other. Mid-size carriers peer with as many partners as economically justified.

Transit / Tandem

When two carriers do not have a direct connection, calls are routed through a transit carrier (IP) or tandem switch (TDM). The transit carrier has interconnections with both the originating and terminating carriers and bridges the gap.

Transit adds cost (the transit carrier charges per minute) and latency (additional network hops). But it provides universal reachability — a small carrier with transit arrangements can reach any destination without needing hundreds of direct peering agreements.

Bill-and-Keep vs. Sender-Pays

Interconnection economics vary by model:

  • Bill-and-keep: Each carrier absorbs its own costs; no intercarrier payments for termination. Increasingly common for VoIP and IP interconnection.
  • Sender-pays (access charges): The originating carrier pays the terminating carrier for call completion. Traditional in the PSTN, especially for calls terminating on ILEC networks with regulated access charges.
  • Revenue share: The terminating carrier pays the originating carrier (or vice versa) a share of the call revenue. Common in international voice and premium services.

Points of Presence (PoPs)

A Point of Presence (PoP) is a physical location where a carrier has network equipment and can interconnect with other carriers. Carriers establish PoPs in major cities and at carrier hotel facilities where many carriers converge.

The number and location of a carrier’s PoPs affects their routing capabilities. A carrier with PoPs in 50 cities can offer lower-latency paths to more destinations than one with PoPs in 5 cities.

Interconnection and NPA/NXX Data

The interconnection relationships between carriers are reflected in routing data:

  • Carrier diversity within an NPA: When you see multiple carriers holding NXX assignments in an area code, each carrier must be reachable through either direct peering or transit
  • Top carriers per NPA: The top carriers table on each area code page shows which carriers hold the most NXX assignments — these are the carriers an originating carrier most needs direct peering with to efficiently serve that area code
  • Rate center coverage: A carrier’s presence in a rate center (holding NXX assignments there) typically means they have a PoP or interconnection in the corresponding LATA

Further Reading