Cisco VoIP & Legacy SM/NIM Interface Cards: Bridging Circuit-Switched and Packet-Switched Networks

An intermediate guide to Cisco Network Interface Modules (NIMs) and Service Modules (SMs) that connect legacy telecom infrastructure — serial WAN, T1/E1, ISDN PRI/BRI — to modern IP networks via ISR routers acting as voice gateways.

Table of Contents

• Chapter 1: Foundations of Modular Networking: Why Routers Need Voice and WAN Interface Cards
• Chapter 2: Serial WAN Interfaces: NIM-4T, NIM-16A, and NIM-24A for Legacy Connectivity
• Chapter 3: T1 and E1 Digital Trunk Lines: NIM-8MFT-T1/E1 and Channelized Voice/Data
• Chapter 4: ISDN Interfaces — PRI and BRI for Digital Voice Signaling
• Chapter 5: The VoIP Gateway — CUBE, CUCM, and SIP Trunk Migration from Legacy NIMs

Chapter 1: Foundations of Modular Networking: Why Routers Need Voice and WAN Interface Cards

Learning Objectives

• Explain the purpose of Network Interface Modules (NIMs) and how they differ from legacy HWIC/WIC/NM form factors
• Describe why Cisco ISR 4000 series routers use modular slot architectures for WAN and voice connectivity
• Distinguish between circuit-switched (PSTN/TDM) and packet-switched (IP/SIP) network paradigms
• Define foundational telecom terms including DS0, TDM, PBX, PSTN, and ISDN and explain their role in voice networking

The Evolution of Cisco Router Modularity

From Fixed-Port Routers to Modular ISR Platforms

In the early days of enterprise networking, routers shipped with a fixed set of ports soldered directly to the motherboard. If you needed a serial WAN connection, you bought a router with serial ports. If you needed an ISDN BRI connection, you bought a different router with BRI ports. This was roughly equivalent to buying a separate kitchen appliance for every task: one machine to blend, another to chop, another to mix. It worked, but it was expensive, inflexible, and filled your rack (or countertop) quickly.

Cisco recognized this problem in the 1990s and began designing routers with empty expansion slots that could accept standardized interface cards. This modular approach meant a single router chassis could serve as a WAN edge device, a voice gateway, a security appliance, or all three simultaneously, depending on which cards you installed. The concept was transformative: instead of replacing an entire router when your connectivity requirements changed, you simply swapped out a card.

This philosophy reached maturity with Cisco’s Integrated Services Router (ISR) product line. An ISR is a router designed from the ground up to host multiple service modules, consolidating what previously required a rack full of separate devices into a single platform. The “integrated services” in the name refers to this ability to combine routing, switching, voice, security, and WAN optimization in one chassis. [Source: Cisco ISR 4000 datasheets]

Form Factor Progression: WIC to HWIC to NM to SM to NIM

As router platforms evolved, so did the physical form factors of their expansion cards. Understanding this progression is important because you will encounter references to all of these in documentation, configuration guides, and existing deployments:

The key trend across this progression is increasing bandwidth, density, and manageability. Early WICs provided a handful of megabits per second per slot. Today’s NIMs deliver up to 2 Gbps per slot, a 250-fold improvement that reflects the dramatic increase in enterprise traffic volumes over the past two decades. [Source: router-switch.com comparisons]

Figure 1.1: Evolution of Cisco Router Interface Card Form Factors

timeline
    title Cisco Interface Card Form Factor Evolution
    1990s : WIC
          : "WAN Interface Card"
          : "~8 Mbps bus bandwidth"
          : "Cisco 1700, 2600, 3600"
    2000s : HWIC / EHWIC
          : "High-Speed WAN Interface Card"
          : "~400 Mbps bus bandwidth"
          : "ISR G1/G2 (1800-3900)"
    2000s : NM
          : "Network Module"
          : "Larger form factor"
          : "Cisco 2600, 3700, 3800"
    2010s : SM / SM-X
          : "Service Module Extended"
          : "Carrier card for NIMs"
          : "ISR G2 and ISR 4000"
    2015+ : NIM
          : "Network Interface Module"
          : "Up to 2 Gbps per slot"
          : "ISR 4000 series (current)"

A critical point for anyone planning hardware purchases: these form factors are not backward compatible. You cannot install a legacy HWIC card into a NIM slot on an ISR 4000, and vice versa. Each generation introduced a new physical connector, new electrical signaling, and new software interfaces. Think of it like the progression from VHS to DVD to Blu-ray: each generation improved capacity and quality, but you need the matching player for each disc format.

Cisco ISR 4000 Series NIM Slot Architecture and Compatibility Matrix

The ISR 4000 series is Cisco’s current-generation branch router platform, and the NIM is its primary expansion form factor. Each ISR 4000 model provides a specific number of NIM slots:

Each NIM slot connects to the router’s internal bus at up to 2 Gbps, providing ample throughput for even the most demanding WAN or voice applications. NIMs also support Online Insertion and Removal (OIR), meaning you can hot-swap a NIM card while the router is running without causing a full system reboot. This is a significant operational advantage in environments where downtime must be minimized. [Source: Cisco ISR 4000 datasheets]

For organizations that need even more expansion capacity, the ISR 4000 series supports the SM-X (Service Module Extended) carrier card. An SM-X carrier can host one or two additional NIMs inside its larger enclosure, effectively adding NIM slots to the router beyond the built-in count. This is particularly useful for voice gateway deployments where you might need multiple T1/E1 NIMs alongside analog voice NIMs simultaneously. [Source: Cisco ISR 4000 datasheets]

Worked Example: Sizing a Branch Office Router

Imagine you are designing the network for a 50-person branch office. The office needs:

• A WAN uplink to the corporate MPLS network (1 NIM slot for a WAN NIM)
• Four analog phone lines connected to the local phone company (1 NIM slot for a 4-port FXO NIM)
• Eight analog phones at employee desks (1 NIM slot for an 8-port FXS NIM)

That totals three NIM slots. An ISR 4321 with only two slots would be insufficient. You would select an ISR 4331 or higher, which provides three NIM slots and accommodates the entire design in a single chassis.

Figure 1.2: Branch Office NIM Slot Allocation on ISR 4331

flowchart LR
    subgraph ISR["ISR 4331 Router"]
        direction TB
        NIM1["NIM Slot 1\nWAN NIM\n(MPLS Uplink)"]
        NIM2["NIM Slot 2\n4-Port FXO NIM\n(Phone Lines)"]
        NIM3["NIM Slot 3\n8-Port FXS NIM\n(Desk Phones)"]
    end
    MPLS["Corporate\nMPLS Network"] --> NIM1
    PSTN["Phone Company\n(4 Lines)"] --> NIM2
    PHONES["8 Analog\nDesk Phones"] --> NIM3

Key Takeaway: Cisco routers evolved from fixed-port boxes to modular platforms that accept swappable interface cards. The NIM is the current-generation form factor for ISR 4000 series routers, offering up to 2 Gbps per slot, OIR support, and no backward compatibility with older form factors like HWIC or WIC. Choosing the right ISR model starts with counting how many NIM slots your design requires.

The On-Ramp Concept: Bridging Legacy and Modern Networks

Why Organizations Still Operate Legacy Telecom Infrastructure

If packet-switched IP networks are faster, cheaper, and more flexible than circuit-switched telephone networks, why does anyone still use the old technology? The answer comes down to three factors: investment protection, regulatory requirements, and reliability expectations.

Many organizations made enormous capital investments in legacy phone systems during the 1980s and 1990s. A large hospital, for example, might have a PBX phone system with thousands of extensions, paging integrations, nurse call systems, and elevator emergency phones, all wired with copper pairs to a central telephone closet. Replacing that entire infrastructure overnight is neither financially practical nor operationally safe. Instead, organizations migrate gradually, keeping legacy systems running alongside new IP-based systems during a transition period that can span years.

Additionally, certain industries face regulatory requirements that mandate traditional phone service. Emergency 911 service, for instance, has historically relied on circuit-switched connections that provide automatic location identification. While modern Enhanced 911 systems support VoIP, the transition is not yet universal. Some organizations maintain legacy PSTN connections specifically to comply with these mandates.

Finally, there is the reliability question. The traditional phone network was engineered for “five nines” availability (99.999% uptime, or roughly five minutes of downtime per year). IP networks have improved dramatically, but many IT teams still regard the PSTN as the “gold standard” for voice reliability and maintain it as a backup path.

The Router as a Protocol Translator Between Circuit-Switched and Packet-Switched Worlds

This is where Cisco’s modular router architecture becomes essential. An ISR equipped with the right NIM cards acts as an on-ramp between legacy telecom infrastructure and modern IP networks. It translates between two fundamentally different ways of carrying voice traffic.

Consider this analogy: imagine a town where half the streets are canals (boats only) and the other half are paved roads (cars only). To move goods across the town, you need a loading dock at the boundary where cargo is unloaded from boats and loaded onto trucks. The ISR router with voice NIMs serves exactly this role: it sits at the boundary between the PSTN (the canal) and the IP network (the road), converting voice traffic from one format to the other.

On the PSTN-facing side, a voice NIM physically connects to telephone lines using the same signaling and encoding that the phone network has used for decades. On the IP-facing side, the router encapsulates that same voice into IP packets using protocols like SIP (Session Initiation Protocol) or H.323 and sends them across the data network. The ISR 4000 platform with voice NIMs supports H.323, SIP, MGCP, and SCCP signaling protocols, making it compatible with virtually any VoIP system. [Source: Cisco ISR 4000 datasheets]

Figure 1.3: ISR Router as Protocol Translator Between Circuit-Switched and Packet-Switched Networks

sequenceDiagram
    participant PSTN as PSTN / TDM<br>(Circuit-Switched)
    participant ISR as Cisco ISR 4000<br>with Voice NIMs
    participant IP as IP Network<br>(Packet-Switched)

    PSTN->>ISR: Analog/T1/E1 voice signal<br>(DS0 channels, TDM)
    Note over ISR: Voice NIM receives<br>circuit-switched audio
    Note over ISR: DSP converts TDM<br>to compressed VoIP
    ISR->>IP: SIP/RTP packets<br>(encoded voice over IP)
    IP->>ISR: SIP/RTP packets<br>(inbound VoIP call)
    Note over ISR: DSP converts VoIP<br>packets to TDM audio
    ISR->>PSTN: Analog/T1/E1 voice signal<br>(DS0 channels, TDM)

Common Migration Scenarios: PSTN to SIP, Frame Relay to MPLS/SD-WAN

Enterprise networks rarely make a single leap from old technology to new. Instead, they go through staged migrations where both legacy and modern systems coexist. Here are three common scenarios where modular interface cards are essential:

Scenario 1: PSTN to SIP Trunking An organization replaces its traditional phone lines with SIP trunks from an internet telephony provider. However, some locations still have analog phones, fax machines, or alarm systems that require traditional analog connections. FXS NIMs in the ISR provide analog ports for these devices while the router handles SIP trunking over its IP WAN connection.

Scenario 2: TDM PBX to IP PBX A company operates a legacy PBX that uses T1/E1 digital trunks. They are migrating to Cisco Unified Communications Manager (an IP-based PBX) but cannot cut over all sites at once. During the transition, an ISR with T1/E1 NIMs connects to the legacy PBX on one side and the IP network on the other, allowing calls to flow between old and new systems.

Scenario 3: Frame Relay to MPLS/SD-WAN While not strictly a voice scenario, this WAN migration follows the same on-ramp logic. Branch offices with legacy Frame Relay WAN links get an ISR with a serial NIM for the old connection and an Ethernet NIM for the new MPLS or SD-WAN link. The router handles the transition transparently, and when the old circuit is finally decommissioned, you simply remove the serial NIM.

Key Takeaway: Legacy telecom infrastructure persists in enterprises due to capital investment, regulatory requirements, and reliability expectations. Cisco ISR routers with voice NIMs serve as protocol translators (on-ramps) between circuit-switched and packet-switched worlds, enabling gradual migrations rather than risky forklift upgrades.

Circuit-Switched vs Packet-Switched Fundamentals

How the PSTN Works: Dedicated Circuits, Time Slots, and Signaling

The Public Switched Telephone Network (PSTN) is the global circuit-switched telephone system that has been in continuous operation since the late 1800s. Understanding how it works is essential background for working with voice interface cards, because those cards must speak the PSTN’s language.

In a circuit-switched network, when you place a phone call, the network establishes a dedicated electrical path between your phone and the recipient’s phone. This path is reserved exclusively for your conversation for the entire duration of the call. No other traffic shares it. This is like having a private highway lane reserved for your car alone, from on-ramp to off-ramp. Nobody else can use that lane until you hang up, regardless of whether you are actively speaking or sitting in silence.

In the digital PSTN, this dedicated path takes the form of a DS0 (Digital Signal, level 0) channel. A DS0 carries exactly 64 Kbps of bandwidth, which is the product of sampling a voice signal 8,000 times per second at 8 bits per sample (8,000 x 8 = 64,000 bits/second). This sampling rate and bit depth were chosen because they are sufficient to reproduce the frequency range of human speech (300 Hz to 3,400 Hz) with acceptable quality. Every phone call on the digital PSTN occupies exactly one DS0. [Source: Cisco ISR 4000 datasheets]

Multiple DS0 channels are bundled together onto a single physical wire using Time Division Multiplexing (TDM). In TDM, each DS0 is assigned a fixed time slot on the wire. The equipment at both ends knows that, for example, time slot 5 always carries the audio for call number 5. This is like a revolving door with numbered compartments: each person waits for their assigned compartment, steps in, and gets carried to the other side in a predictable, ordered fashion.

Figure 1.4: Circuit-Switched vs Packet-Switched Voice Delivery

flowchart TD
    subgraph CS["Circuit-Switched (PSTN/TDM)"]
        direction LR
        CALLER1["Caller"] -->|"Dedicated DS0\n(64 Kbps reserved)"| PATH1["Fixed Path\nThrough Network"]
        PATH1 -->|"Same path\nentire call"| RECV1["Receiver"]
    end

    subgraph PS["Packet-Switched (IP/VoIP)"]
        direction LR
        CALLER2["Caller"] -->|"Voice Packet 1"| NET["Shared IP\nNetwork"]
        CALLER2 -->|"Voice Packet 2"| NET
        CALLER2 -->|"Voice Packet 3"| NET
        NET -->|"Packets reassembled\nin order"| RECV2["Receiver"]
    end

    CS -.->|"Both bridged by\nCisco ISR with\nVoice NIMs"| PS

The two most common TDM bundle sizes are:

A T1 line carries 24 simultaneous phone calls. An E1 line carries 30 or 31 (one time slot is typically reserved for signaling). When you install a T1/E1 NIM in a Cisco ISR, you are giving the router the ability to connect directly to these TDM circuits and participate in the PSTN. [Source: Cisco ISR 4000 datasheets]

How IP Networks Work: Packets, Routing, and Best-Effort Delivery

IP (Internet Protocol) networks take a fundamentally different approach to carrying data. Instead of reserving a dedicated path, IP networks break all communication into small chunks called packets. Each packet contains the destination address and is independently routed through the network, potentially taking a different path than the packet before it. At the destination, packets are reassembled in the correct order.

Returning to the highway analogy: if circuit switching gives you a private lane, packet switching puts your cargo into individual delivery trucks that each navigate traffic independently. Some trucks might take the freeway, others might take side streets, and they might arrive in a different order than they were sent. But the delivery is usually faster overall because the road capacity is shared efficiently among all users.

IP networks are described as best-effort, meaning they make no guarantee about delivery time, packet order, or even whether a packet will arrive at all. This might sound alarming for voice calls, but modern networks use Quality of Service (QoS) mechanisms to prioritize voice packets over less time-sensitive traffic like email or file downloads. When properly configured, VoIP quality on a well-managed IP network is indistinguishable from a traditional phone call.

The key advantage of packet switching for voice is statistical multiplexing: because bandwidth is shared dynamically, you can carry far more calls over the same physical link than circuit switching allows. In a TDM system, a DS0 is reserved for the entire duration of a call even during silence (which accounts for roughly 50-60% of a typical conversation). In a VoIP system, no bandwidth is consumed during silence, freeing capacity for other calls.

Why Both Paradigms Coexist in Enterprise Networks Today

Given that packet switching is more efficient, you might wonder why circuit switching has not disappeared entirely. Several factors sustain the coexistence:

1. Installed base: Billions of dollars of TDM equipment remain in service worldwide. Telephone company central offices, enterprise PBXs, and alarm systems still rely on circuit-switched connections.

2. Predictable quality: Circuit switching provides guaranteed bandwidth and constant latency. For voice, this means consistent call quality without the jitter and packet loss that can occasionally affect VoIP.

3. Regulatory and safety systems: As noted earlier, emergency services, elevator phones, fire alarm panels, and certain medical devices may require or strongly prefer traditional phone connections.

4. Geographic coverage: In rural or developing areas, the PSTN may be the only available communication infrastructure, with IP broadband not yet deployed.

The practical result is that enterprise networks will continue operating in a hybrid mode for years to come. This is precisely why Cisco designs ISR routers with NIM slots that can accept both legacy TDM interface cards and modern Ethernet/IP cards: the router must speak both languages fluently.

  +-----------------+         +-------------------+         +------------------+
  |   PSTN / TDM    |         |   Cisco ISR 4000  |         |   IP Network     |
  |   (Circuit-     | T1/E1   |   +-------------+ |  GigE   |   (Packet-       |
  |    Switched)    |-------->|   | Voice NIM   | |-------->|    Switched)     |
  |                 |  FXO    |   +-------------+ |         |                  |
  |  Phone Lines    |-------->|   | Analog NIM  | |  SIP    |  VoIP Server     |
  |  PBX Trunks     |         |   +-------------+ |         |  (CUCM, etc.)    |
  +-----------------+         +-------------------+         +------------------+
        Legacy Side              Protocol Translator             Modern Side

Key Takeaway: Circuit-switched networks (PSTN) dedicate a fixed DS0 channel per call using TDM, providing guaranteed quality but inefficient bandwidth use. Packet-switched networks (IP) share bandwidth dynamically using packets, offering greater efficiency but requiring QoS for voice quality. Both paradigms coexist in enterprises, and Cisco ISR routers with voice NIMs bridge the gap between them.

Essential Telecom Terminology

This section provides detailed explanations of the foundational terms you will encounter throughout this textbook. Each term builds on the concepts introduced in earlier sections.

DS0: The 64 Kbps Building Block of Digital Telephony

The DS0 (Digital Signal, level 0) is the smallest unit of bandwidth in the digital telephone network. At exactly 64 Kbps, it represents a single digitized voice channel. Every phone call in the PSTN ultimately occupies one DS0, regardless of whether it traverses a T1 line in New York or an E1 line in London.

To understand why 64 Kbps: the Nyquist theorem tells us that to faithfully digitize a signal, we must sample it at twice its highest frequency. Human speech is filtered to approximately 4,000 Hz for telephony purposes, requiring a minimum sampling rate of 8,000 samples per second. Each sample is encoded as an 8-bit value using a codec called G.711, yielding 8,000 samples x 8 bits = 64,000 bits per second.

Real-world example: When you configure a T1 NIM on a Cisco ISR and allocate time slots for voice calls, you are essentially assigning DS0 channels. If your T1 has 24 DS0s and you reserve 12 for voice and 12 for data, you can handle 12 simultaneous phone calls while maintaining a 768 Kbps data link (12 x 64 Kbps). [Source: Cisco ISR 4000 datasheets]

TDM: Time Division Multiplexing and How It Packs Multiple Calls onto a Single Wire

Time Division Multiplexing (TDM) is the technique used to combine multiple DS0 channels onto a single physical circuit. Each DS0 is assigned a repeating time slot in a fixed frame structure. Equipment at both ends of the link reads and writes data in precisely synchronized time slots, ensuring each call’s audio ends up in the right channel.

Think of TDM like a conveyor belt with numbered buckets passing a loading dock. Bucket 1 always carries items for Customer 1, Bucket 2 for Customer 2, and so on. The belt runs at a constant speed, and every customer gets their turn at regular intervals, whether they have something to load or not. This “whether they have something or not” is the key inefficiency of TDM compared to packet switching: even silent time slots consume bandwidth.

TDM comes in several hierarchies:

When you install a multi-port T1/E1 NIM in a Cisco ISR, each port represents one physical TDM circuit. A 4-port T1 NIM, for instance, provides up to 96 simultaneous DS0 voice channels (4 x 24). These NIMs require a separate PVDM4 (Packet Voice Data Module, generation 4) DSP module to handle the real-time conversion between TDM audio and compressed VoIP packets. [Source: Cisco ISR 4000 datasheets]

PBX: Private Branch Exchange and Its Role in Office Phone Systems

A Private Branch Exchange (PBX) is a private telephone system used within an organization. The “exchange” part refers to its ability to switch calls: it connects internal extensions to each other without using the public phone network, and it provides shared access to a limited number of outside telephone lines.

Consider this analogy: a PBX is like a receptionist at a hotel front desk who manages a switchboard. When a guest in Room 205 wants to call Room 310, the receptionist connects them directly without involving the phone company. When a guest wants to call outside the hotel, the receptionist connects them to one of the hotel’s external phone lines. The PBX automates this process electronically.

A traditional PBX connects to the PSTN via trunk lines, which are typically T1/E1 circuits. On the internal side, it connects to individual desk phones via analog or digital station lines. The PBX handles call routing, voicemail, call forwarding, hold music, and dozens of other features.

Why this matters for NIMs: When an organization migrates from a traditional PBX to an IP-based phone system (like Cisco Unified Communications Manager), the Cisco ISR often serves as the gateway between the two. T1/E1 NIMs connect to the existing PBX trunk ports, while the router’s IP interfaces connect to the new VoIP system. This allows both systems to operate simultaneously during the migration.

FXO vs FXS: Analog Port Types for Connecting Phones and Trunk Lines

FXO (Foreign Exchange Office) and FXS (Foreign Exchange Station) are the two types of analog voice ports found on Cisco voice NIMs. They serve opposite and complementary roles, and confusing them is one of the most common mistakes in voice networking.

The key to remembering the difference: FXS provides dial tone; FXO receives dial tone.

FXS ports simulate what the phone company’s central office provides to your home: dial tone, the ability to ring the phone, and power to operate it. When you plug an analog phone into an FXS port on a Cisco NIM, the router acts like the phone company for that phone. It generates dial tone, detects when the handset is lifted (off-hook), and sends ringing voltage when an inbound call arrives.

FXO ports simulate what a telephone does when connected to the phone company. When you connect an FXO port on a Cisco NIM to a phone line from the telephone company (or a PBX extension port), the router acts like a phone on that line. It can go off-hook to seize the line, dial digits, and detect incoming ring signals.

Figure 1.5: FXO and FXS Port Roles in a Voice Gateway

flowchart LR
    CO["Phone Company\nCentral Office"] -->|"Provides\nDial Tone"| FXO["FXO Port\non Cisco NIM\n(Acts as Phone)"]
    FXO -->|"Receives\nDial Tone"| GW["Cisco ISR\nVoice Gateway"]
    GW -->|"Provides\nDial Tone"| FXS["FXS Port\non Cisco NIM\n(Acts as CO)"]
    FXS -->|"Ring Voltage\n+ Battery"| PHONE["Analog\nDesk Phone"]

    style FXO fill:#1a3a5c,stroke:#58a6ff,color:#fff
    style FXS fill:#1a3a5c,stroke:#58a6ff,color:#fff
    style GW fill:#0d1117,stroke:#58a6ff,color:#fff

Worked Example: Small Office Voice Gateway

A small office has four phone lines from the local telephone company and six analog desk phones. To convert this to VoIP while keeping the existing analog phones and phone lines:

• Install a 4-port FXO NIM to connect to the four incoming phone lines. The router receives dial tone from the phone company on these ports.
• Install an 8-port FXS NIM to connect to the six desk phones (with two spare ports for future growth). The router provides dial tone to the phones on these ports.
• The router, running Cisco voice gateway software, bridges calls between the FXO ports (PSTN-facing) and FXS ports (phone-facing), while also enabling VoIP calling over the IP network via SIP.

Voice NIMs for the ISR 4000 platform that provide analog ports (FXS, FXO, DID, E/M) include onboard DSP resources, so they do not require a separate PVDM4 module. Digital T1/E1 NIMs, by contrast, do require PVDM4 DSPs for voice processing. [Source: Cisco interface card installation guides]

Key Takeaway: DS0 is the fundamental 64 Kbps voice channel. TDM multiplexes multiple DS0s onto a single wire using time slots. A PBX is the private phone switch inside an organization. FXS ports provide dial tone to phones (like a wall jack), while FXO ports receive dial tone from the phone company (like a phone plug). Understanding these terms is essential for configuring voice NIMs on Cisco ISR routers.

Chapter Summary

This chapter established the foundational concepts needed to understand why Cisco routers use modular interface cards for voice and WAN connectivity. The evolution from fixed-port routers to the modular ISR 4000 platform reflects a broader industry shift toward flexible, software-defined infrastructure. The NIM form factor represents the current generation of this modular approach, offering up to 2 Gbps per slot, hot-swap capability, and a wide range of voice and data interface options across two to three slots per ISR 4000 model.

We explored the “on-ramp” concept: the idea that a Cisco ISR equipped with the right NIMs serves as a protocol translator between the legacy circuit-switched world (PSTN, TDM, PBX) and the modern packet-switched world (IP, SIP, VoIP). This bridging function is critical because enterprise networks do not transition overnight. Legacy telecom infrastructure persists due to capital investment, regulatory requirements, and reliability expectations, creating a sustained need for gateway routers that speak both languages.

Finally, we unpacked the essential telecom terminology that will recur throughout this textbook: DS0 as the 64 Kbps building block, TDM as the multiplexing technique that bundles DS0s onto physical circuits, PBX as the private phone switch, and FXO/FXS as the complementary analog port types. With these foundations in place, subsequent chapters will dive deeper into specific NIM types, their configuration, and the voice architectures they enable.

Key Terms

Chapter 2: Serial WAN Interfaces: NIM-4T, NIM-16A, and NIM-24A for Legacy Connectivity

Learning Objectives

• Identify the serial interface standards (RS-232, RS-449, V.35, X.21) and their physical and electrical characteristics
• Explain the historical role of serial WAN connections in leased lines, Frame Relay, HDLC, and PPP networks
• Differentiate between synchronous serial (NIM-4T) and asynchronous serial (NIM-16A/24A) NIMs and their use cases
• Describe real-world scenarios where serial NIMs remain in active use: banking, SCADA, out-of-band management

Serial WAN Standards and Physical Interfaces

Before fiber optics and Ethernet dominated wide-area networking, serial interfaces were the fundamental building blocks that connected offices, data centers, and remote sites across the globe. Understanding these standards is not merely an exercise in history — millions of serial links remain in production today, and the principles they embody still shape how we think about clocking, signaling, and point-to-point connectivity.

Think of serial interfaces like different gauges of railroad track. Each standard defines the “width of the track” (electrical signaling), the “station design” (pinouts and connectors), how far the track can stretch (cable length), and how fast the trains can travel (data rate). Just as you cannot run a high-speed bullet train on narrow-gauge track, you cannot push high-bandwidth WAN traffic over a standard designed for short-distance terminal connections.

RS-232 and RS-449: Pinouts, Cable Lengths, and Speed Limitations

RS-232 (formally TIA/EIA-232) is the most widely recognized serial standard and the one most people encounter first. Originally published in 1962, RS-232 was designed to connect a terminal — such as a teletype machine — to a modem over a short distance.

RS-232 uses unbalanced signaling, meaning each data signal is measured against a common ground wire. This makes the standard simple and inexpensive but vulnerable to electrical noise over longer distances — much like trying to have a conversation in a noisy room by speaking at normal volume. The farther apart you stand, the harder it becomes to hear.

RS-449 (TIA/EIA-449) was developed as an improvement over RS-232, addressing its distance and speed limitations. RS-449 uses balanced signaling (technically defined by the companion standards RS-422 for balanced and RS-423 for unbalanced), which transmits each signal as the voltage difference between two wires rather than between a single wire and ground.

The balanced signaling approach is like noise-canceling headphones: because both wires in a pair pick up the same interference, the receiver can subtract the noise and recover the clean signal. This is why RS-449 can reach dramatically longer distances than RS-232.

Worked Example: Choosing Between RS-232 and RS-449

A network engineer needs to connect a router to a Channel Service Unit/Data Service Unit (CSU/DSU) located 30 meters away in the same building. RS-232 is eliminated immediately — its 15-meter limit makes it unsuitable. RS-449 with balanced signaling comfortably supports this distance at speeds well above what a T1 line (1.544 Mbps) requires. The engineer selects an RS-449 cable and verifies that both devices support the standard.

Key Takeaway: RS-232 is ubiquitous but limited to short distances and low speeds due to unbalanced signaling. RS-449 extends both range and throughput through balanced signaling, making it suitable for WAN connections within a campus or equipment room.

Figure 2.1: RS-232 vs RS-449 Signaling Comparison

flowchart LR
    subgraph RS232["RS-232: Unbalanced Signaling"]
        direction TB
        A1["Data Wire"] -->|"Signal measured against"| A2["Common Ground"]
        A2 -->|"Vulnerable to noise"| A3["Max 15m / 115.2 kbps"]
    end

    subgraph RS449["RS-449: Balanced Signaling"]
        direction TB
        B1["Wire A (+)"] -->|"Voltage difference"| B3["Receiver subtracts noise"]
        B2["Wire B (-)"] -->|"Voltage difference"| B3
        B3 --> B4["Max 1,200m / 2 Mbps"]
    end

    RS232 -.->|"Improved by"| RS449

V.35 and X.21: Higher-Speed Serial for WAN Connections

While RS-232 and RS-449 were developed by American standards bodies, the ITU-T (formerly CCITT) produced its own serial interface standards optimized for telecommunications-grade WAN connections.

V.35 is perhaps the most important serial standard in WAN networking history. Despite being technically obsoleted by the ITU-T in 1988, V.35 became the de facto standard for connecting routers to CSU/DSUs and to carrier-provided equipment at T1/E1 speeds and above. The V.35 connector — a distinctive rectangular block connector with a bail-lock mechanism — remains instantly recognizable to anyone who has worked in a telecom environment.

V.35 was the preferred interface for high-speed synchronous WAN connections because it combines balanced differential signaling for data lines (reducing noise) with a robust connector designed for permanent installations. [Source: Cisco NIM-4T datasheet]

X.21 is an ITU-T standard that takes a minimalist approach compared to the pin-heavy RS-232 and V.35 connectors. X.21 uses a simple 15-pin connector and relies on the network to provide many of the functions that other standards handle with dedicated pins.

X.21 is more commonly encountered in European and Asian deployments, where public data networks adopted it as a standard interface. North American networks tended to favor V.35.

Serial Interface Comparison Summary

Key Takeaway: V.35 became the dominant standard for high-speed synchronous WAN connections due to its noise-resistant balanced signaling and robust connector design. X.21 serves a similar role in European and Asian markets. Both standards far exceed RS-232’s capabilities for WAN applications.

DTE vs DCE Clocking and Cable Selection for Serial Links

One of the most critical concepts in serial networking — and one of the most common sources of misconfiguration — is the distinction between DTE (Data Terminal Equipment) and DCE (Data Circuit-terminating Equipment).

In a serial link, one device must provide the clock signal that synchronizes data transmission between the two endpoints. Without a shared clock, the receiving device would not know when each bit begins and ends — imagine two musicians trying to play a duet without agreeing on a tempo.

• DCE (Data Circuit-terminating Equipment): Provides the clock signal. Typically this is the CSU/DSU, modem, or carrier equipment — the device that interfaces with the telecommunications network.
• DTE (Data Terminal Equipment): Receives the clock signal from the DCE. Typically this is the router or terminal — the device that generates and consumes user data.

In a lab or back-to-back router configuration (where two routers connect directly without a CSU/DSU), one router must be configured as the DCE to provide clocking. The router connected with the DCE end of a crossover serial cable must have the clock rate command configured on its serial interface.

! Router acting as DCE
interface Serial0/0/0
 clock rate 2000000
 no shutdown

How to identify DTE vs DCE cables: On Cisco serial cables, the connector label typically indicates “DTE” or “DCE.” If you are using the NIM-4T’s Smart Serial connectors, the adapter cable you attach determines whether the port behaves as DTE or DCE. This is a physical layer decision — the same NIM-4T port can function as either DTE or DCE depending on which cable is connected. [Source: Cisco NIM-4T datasheet]

Worked Example: Back-to-Back Serial Lab Setup

Two ISR 4400 routers each have a NIM-4T installed. To connect them directly for lab testing, the engineer uses a V.35 crossover cable kit consisting of one DTE cable and one DCE cable. The router connected to the DCE cable is configured with clock rate 2000000 to provide a 2 Mbps clock. The router connected to the DTE cable requires no clock rate configuration — it derives timing from the received clock signal.

Key Takeaway: In every synchronous serial link, one side must be DCE (providing the clock) and the other DTE (receiving the clock). On the NIM-4T, the attached cable determines the DTE/DCE role, and the DCE side must be configured with clock rate. Getting this wrong is one of the most common causes of serial link failures.

Figure 2.2: DTE/DCE Clocking Relationship in a Serial Link

sequenceDiagram
    participant Router as Router (DTE)
    participant Cable as Serial Cable
    participant CSU as CSU/DSU (DCE)
    participant Carrier as Carrier Network

    CSU->>Router: Provides clock signal
    Note over Router,CSU: DCE always provides timing reference
    Router->>CSU: Sends data synchronized to clock
    CSU->>Router: Sends data synchronized to clock
    CSU->>Carrier: Interfaces with telecom network

    Note over Router: No "clock rate" command needed
    Note over CSU: Clock source for the link

Historical WAN Protocols over Serial Links

Serial interfaces were merely the physical medium — the railroad track. The protocols that ran over them defined how data was packaged, addressed, and delivered. Understanding these protocols provides essential context for why serial NIMs like the NIM-4T still exist and what they were designed to support.

Leased Lines and Dedicated Point-to-Point Circuits

A leased line (also called a dedicated circuit or private line) is a permanent, always-on connection between two sites, provisioned by a telecommunications carrier. Unlike a phone call that is set up and torn down, a leased line is always available — think of it as renting a private highway lane that only your organization can use.

Leased lines were provisioned at standard speeds defined by the digital signal hierarchy:

A typical leased-line deployment connected a router at a branch office to a router at headquarters via a pair of CSU/DSUs and a carrier-provided circuit. Each router’s serial interface connected to a local CSU/DSU, which in turn connected to the carrier’s network. The serial link between router and CSU/DSU used V.35 or RS-449, with the CSU/DSU acting as DCE and providing clock to the router (DTE).

Leased lines were simple and reliable but expensive — you paid for the full bandwidth of the circuit whether you used it or not. This led to the development of shared WAN technologies like Frame Relay.

Key Takeaway: Leased lines provided dedicated, always-on point-to-point connectivity at guaranteed bandwidth. Their simplicity and reliability made them the gold standard for WAN connectivity, but their cost drove the adoption of shared technologies like Frame Relay.

Figure 2.3: Leased Line Point-to-Point Topology

flowchart LR
    R1["Branch Router\n(DTE)"] -->|"V.35 / RS-449"| CSU1["CSU/DSU\n(DCE)"]
    CSU1 -->|"T1 Circuit\n1.544 Mbps"| Cloud["Carrier\nNetwork"]
    Cloud -->|"T1 Circuit\n1.544 Mbps"| CSU2["CSU/DSU\n(DCE)"]
    CSU2 -->|"V.35 / RS-449"| R2["HQ Router\n(DTE)"]

    style Cloud fill:#1a3a5c,stroke:#58a6ff,color:#fff
    style R1 fill:#0d2137,stroke:#58a6ff,color:#fff
    style R2 fill:#0d2137,stroke:#58a6ff,color:#fff
    style CSU1 fill:#0d2137,stroke:#58a6ff,color:#fff
    style CSU2 fill:#0d2137,stroke:#58a6ff,color:#fff

Frame Relay: Virtual Circuits over Shared Serial Infrastructure

Frame Relay solved the cost problem of leased lines by allowing multiple customers to share the same physical serial infrastructure while maintaining logical separation. Instead of a dedicated physical circuit between every pair of sites, Frame Relay created virtual circuits identified by Data Link Connection Identifiers (DLCIs).

Think of Frame Relay like a shared office building with private mailboxes. Everyone uses the same lobby and hallways (shared infrastructure), but each tenant has their own labeled mailbox (DLCI) that ensures mail reaches the right recipient. You pay rent based on how much space you need, not for the entire building.

Each Frame Relay virtual circuit was assigned:

• A DLCI — a locally significant number identifying the virtual circuit at each end
• A Committed Information Rate (CIR) — the guaranteed minimum bandwidth the carrier would deliver
• A Burst Rate — the maximum bandwidth available when the network was not congested

Frame Relay was enormously popular from the early 1990s through the 2010s, particularly for connecting branch offices to headquarters over serial WAN links. A single serial interface on a router could support multiple DLCIs, each connecting to a different remote site — dramatically reducing the number of physical connections needed.

! Frame Relay configuration example on a serial interface
interface Serial0/0/0
 encapsulation frame-relay
 frame-relay interface-dlci 102
 frame-relay interface-dlci 103

Real-World Context: Frame Relay was heavily used in SCADA (Supervisory Control and Data Acquisition) networks, where industrial control systems at remote sites like water treatment plants, electrical substations, and oil pipelines needed reliable, low-latency connections to central monitoring stations. SCADA telemetry traveled over Frame Relay virtual circuits identified by DLCIs, with serial interfaces providing the physical connectivity. [Source: Cisco SCADA design guide]

Key Takeaway: Frame Relay allowed multiple virtual circuits to share physical serial infrastructure, dramatically reducing WAN costs compared to leased lines. DLCIs provided logical separation, and CIR guaranteed minimum bandwidth. Frame Relay over serial links was the backbone of branch-office and SCADA connectivity for two decades.

Figure 2.4: Frame Relay Virtual Circuits over Shared Infrastructure

flowchart TD
    HQ["HQ Router\nSerial0/0/0"] -->|"DLCI 102"| FR["Frame Relay\nCarrier Cloud"]
    HQ -->|"DLCI 103"| FR
    HQ -->|"DLCI 104"| FR
    FR -->|"DLCI 201"| BR1["Branch A Router"]
    FR -->|"DLCI 301"| BR2["Branch B Router"]
    FR -->|"DLCI 401"| BR3["Branch C Router"]

    style FR fill:#1a3a5c,stroke:#58a6ff,color:#fff
    style HQ fill:#0d2137,stroke:#58a6ff,color:#fff
    style BR1 fill:#0d2137,stroke:#58a6ff,color:#fff
    style BR2 fill:#0d2137,stroke:#58a6ff,color:#fff
    style BR3 fill:#0d2137,stroke:#58a6ff,color:#fff

HDLC and PPP: Encapsulation Protocols for Serial WAN Links

When two routers connect over a serial link, they need a Layer 2 protocol to frame the data — to mark where each packet begins and ends and to provide basic error detection. The two most common protocols for this purpose were HDLC and PPP.

HDLC (High-Level Data Link Control) is the default encapsulation on Cisco serial interfaces. Cisco’s implementation of HDLC adds a proprietary protocol type field, which means Cisco HDLC is not compatible with other vendors’ HDLC implementations. If both ends of the serial link are Cisco routers, HDLC is the simplest choice — it requires no configuration beyond the default.

PPP (Point-to-Point Protocol) is an open-standard alternative that provides several features HDLC lacks:

PPP is the protocol of choice whenever the serial link connects equipment from different vendors, when authentication is required, or when multiple physical links need to be bonded together for additional bandwidth.

! PPP configuration with CHAP authentication
interface Serial0/0/0
 encapsulation ppp
 ppp authentication chap

Key Takeaway: HDLC is the simplest encapsulation for Cisco-to-Cisco serial links and is the default on Cisco serial interfaces. PPP is the multi-vendor, feature-rich alternative that adds authentication, multilink bonding, and network layer negotiation. Choose HDLC for simplicity in single-vendor environments; choose PPP when interoperability or security features are needed.

How These Protocols Compared to Modern MPLS and SD-WAN

The serial-era WAN protocols — leased lines, Frame Relay, HDLC, and PPP — were eventually supplanted by MPLS (Multiprotocol Label Switching) and, more recently, SD-WAN (Software-Defined Wide Area Network). Understanding this evolution helps explain why serial NIMs persist in certain environments.

The progression from leased lines to Frame Relay to MPLS to SD-WAN represents a steady movement toward shared infrastructure, software control, and lower cost. However, this progression has not been universal. Industries that require deterministic latency (where “the packet arrives in exactly X milliseconds, every time”), air-gap security (physical isolation from the internet), or regulatory compliance with standards that specify serial connectivity continue to operate serial WAN links.

Key Takeaway: MPLS and SD-WAN have replaced serial WAN protocols for most enterprise applications, offering greater flexibility and lower cost. However, legacy serial connections persist where deterministic latency, physical security isolation, or regulatory requirements make them irreplaceable.

Figure 2.5: WAN Technology Evolution Timeline

timeline
    title Evolution of WAN Technologies
    1980s : Leased Lines (T1/E1)
          : Dedicated point-to-point circuits
          : Serial interfaces (V.35, RS-232)
    1990s : Frame Relay
          : Virtual circuits with DLCIs
          : Shared serial infrastructure
    2000s : MPLS
          : Label switching over Ethernet/Fiber
          : Carrier-managed VPNs
    2010s-Present : SD-WAN
                  : Software-defined overlays
                  : Multiple transport types
                  : Commodity broadband + LTE

Synchronous Serial: NIM-4T

The NIM-4T is a 4-port synchronous serial Network Interface Module designed for the Cisco ISR 4000 Series routers (ISR 4400 platforms). It represents Cisco’s current-generation hardware for connecting to legacy serial WAN infrastructure and is the module you will encounter whenever a modern router needs to interface with serial-based services.

NIM-4T Hardware Specifications and Supported Interface Types

The NIM-4T provides four synchronous serial ports using Smart Serial connectors — a compact, high-density connector that uses adapter cables to convert to the specific serial standard required for each connection.

[Source: Cisco NIM-4T datasheet]

The Smart Serial connector design is worth understanding. Rather than building separate NIMs for each serial standard (one for V.35, one for RS-232, and so on), Cisco designed the NIM-4T with a universal connector. You then attach a Smart Serial-to-V.35 cable, a Smart Serial-to-RS-232 cable, or whichever adapter is appropriate for your environment. The cable also determines whether the port operates as DTE or DCE. This approach keeps the NIM compact (four ports fit in a single NIM slot) while supporting the full range of serial standards.

Think of the Smart Serial connector like a universal travel adapter. The NIM-4T is the “plug” and the adapter cable converts it to the local “outlet” standard — V.35 in one installation, X.21 in another.

Worked Example: Determining Required Cables

A network engineer is installing a NIM-4T in an ISR 4431 to connect to two legacy Frame Relay circuits. The carrier’s CSU/DSU equipment has V.35 DCE connectors. The engineer needs:

• Two Smart Serial-to-V.35 DTE cables (because the CSU/DSU is DCE, the router must be DTE)
• No clock rate configuration (the CSU/DSU provides clocking as DCE)

For the remaining two ports, which will connect back-to-back to lab routers, the engineer needs:

• One Smart Serial-to-V.35 DTE cable and one Smart Serial-to-V.35 DCE cable (one port on each pair acts as DCE)
• The clock rate command on the DCE-connected port

Key Takeaway: The NIM-4T packs four synchronous serial ports into a single NIM slot using Smart Serial connectors with adapter cables. It supports all major serial standards (RS-232, RS-449, RS-530, V.35, X.21) and runs HDLC, PPP, and Frame Relay — but not X.25 or bisync. It requires IOS XE 3.12+ with an IP Base license.

Configuration for Point-to-Point WAN and Frame Relay Connections

Configuring the NIM-4T for basic WAN connectivity follows standard Cisco IOS XE serial interface configuration patterns. Below are two common scenarios.

Scenario 1: Point-to-Point Leased Line with HDLC (Default)

interface Serial0/1/0
 description Link to HQ via T1 leased line
 ip address 10.1.1.1 255.255.255.252
 no shutdown

Because HDLC is the default encapsulation on Cisco serial interfaces, no encapsulation command is needed. The CSU/DSU provides clocking, so no clock rate is required on the router.

Scenario 2: Frame Relay with Multiple DLCIs

interface Serial0/1/0
 no ip address
 encapsulation frame-relay
 no shutdown
!
interface Serial0/1/0.102 point-to-point
 description Frame Relay PVC to Branch-A
 ip address 10.1.102.1 255.255.255.252
 frame-relay interface-dlci 102
!
interface Serial0/1/0.103 point-to-point
 description Frame Relay PVC to Branch-B
 ip address 10.1.103.1 255.255.255.252
 frame-relay interface-dlci 103

In this configuration, the physical interface carries Frame Relay encapsulation, and point-to-point subinterfaces map to individual DLCIs. Each subinterface gets its own IP address and can participate in routing as if it were a separate link.

Key Takeaway: NIM-4T configuration follows standard IOS XE serial interface practices. HDLC is default and requires minimal configuration. Frame Relay uses subinterfaces mapped to DLCIs to create multiple logical connections over a single physical port.

Connecting to ISP Equipment and Legacy Banking/POS Networks

The NIM-4T finds its most common real-world applications in environments where serial connectivity is mandated by external requirements — situations where the network engineer does not get to choose the interface type.

Banking and Point-of-Sale (POS) Networks

Many financial institutions continue to use serial connections for POS terminal aggregation and ATM (Automated Teller Machine) communications. These connections often run over dedicated leased lines or Frame Relay circuits to ensure:

• Deterministic latency: Transaction authorization must complete within strict time windows (often under 2 seconds). Serial leased lines provide consistent, predictable latency because the bandwidth is dedicated and the path is fixed.
• Physical isolation: Serial connections to the banking network are not routable from the internet, providing an inherent air gap that satisfies PCI DSS (Payment Card Industry Data Security Standard) physical security requirements.
• Regulatory compliance: Some financial regulators in certain jurisdictions still specify serial connectivity in their technical standards, having validated security controls around this architecture over decades.

ISP and Carrier Handoffs

In some regions, particularly in developing markets and rural areas, the local ISP or carrier may only offer serial (T1/E1) handoffs. The NIM-4T allows a modern ISR 4400 router to terminate these legacy circuits while running current IOS XE software with its full feature set, including modern security, Quality of Service, and management capabilities.

Real-World Analogy: Using a NIM-4T in a modern ISR 4400 is like installing a record player input on a modern high-fidelity amplifier. The amplifier has all the latest digital capabilities, but it can also play the vinyl records that still exist in your collection — and some of those records contain music that has never been digitized.

Key Takeaway: The NIM-4T enables modern ISR 4400 routers to connect to legacy serial WAN infrastructure that persists in banking, POS, carrier handoffs, and regulated environments. It bridges the gap between current-generation routing platforms and the serial-based networks that industries continue to depend on.

Asynchronous Serial: NIM-16A and NIM-24A

While the NIM-4T handles synchronous serial WAN connections, the NIM-16A and NIM-24A serve an entirely different purpose: asynchronous serial connectivity for terminal server, console aggregation, and out-of-band management applications.

The distinction between synchronous and asynchronous serial is fundamental:

Synchronous serial is like two factories with synchronized conveyor belts running at the same speed — goods flow continuously and efficiently. Asynchronous serial is like two people passing notes back and forth — each note (byte) has a “here it comes” marker (start bit) and an “I’m done” marker (stop bit), and the timing between notes can vary.

Console Server and Modem Aggregation Use Cases

The NIM-16A and NIM-24A transform a Cisco ISR 4000 router into a console server (also called a terminal server) — a device that provides network-based access to the serial console ports of other equipment.

NIM-16A Specifications:

NIM-24A Specifications:

[Source: Cisco NIM-16A/24A configuration guides]

Important limitation: Neither the NIM-16A nor the NIM-24A supports SLIP, PPP, or async routing. These modules are strictly for terminal server applications — they provide character-mode access to connected devices, not data-link-layer WAN connectivity. This is a critical distinction that differentiates them from the NIM-4T.

A single ISR 4000 platform can support up to 200 asynchronous ports across multiple NIM-16A and NIM-24A modules, enabling large-scale console aggregation. [Source: Cisco NIM-16A/24A configuration guides]

Console Server Example:

A data center has 48 network devices (switches, routers, firewalls) that each have a serial console port. Two NIM-24A modules installed in a single ISR 4431 provide 48 async ports — one for each device’s console. Each async port is configured with a specific line number, and engineers access remote consoles using reverse telnet or reverse SSH from anywhere on the network:

! Access console of device connected to async line 1
! From any SSH client on the network:
ssh -l user router-ip 2001
! Port 2001 = 2000 + line number (1)

This eliminates the need to physically walk to each device or maintain separate console cables at every rack.

Key Takeaway: The NIM-16A (16 ports, 115.2 kbps) and NIM-24A (24 ports, 256 kbps) provide asynchronous RS-232 connectivity for console server and terminal aggregation — not for WAN data links. They do not support SLIP, PPP, or routing protocols. A single platform can scale to 200 async ports.

Out-of-Band Management for Remote Infrastructure

Out-of-band (OOB) management is the practice of maintaining a management path to network equipment that is completely independent of the production data network. When the primary network fails — due to a routing loop, a misconfiguration, or a fiber cut — OOB management provides an alternate path for engineers to reach the affected devices and restore service.

The NIM-16A and NIM-24A are purpose-built for OOB management architectures. The typical deployment connects each critical device’s serial console to an async port on the NIM, and the router hosting the NIM connects to the management network via a separate path (often a cellular modem or a dedicated management VLAN).

Real-World Scenario: Remote Branch OOB Management

A regional bank has 30 branch offices, each with a router, a switch, and a firewall. Each branch has one NIM-16A in its ISR 4431 router. The serial console ports of all three devices at each branch connect to async ports on the NIM-16A. A cellular modem provides an alternate management path to the ISR 4431.

When a branch’s primary WAN circuit goes down:

1. The central NOC (Network Operations Center) connects to the branch ISR 4431 via the cellular backup link
2. The engineer uses reverse telnet/SSH through the NIM-16A to reach the console of the branch router, switch, or firewall
3. The engineer diagnoses and resolves the issue remotely, without dispatching a technician

Without OOB management, a network failure at a remote branch in a rural area might require a multi-hour drive for a technician — costing time, money, and extended outage duration.

Key Takeaway: OOB management via the NIM-16A/24A provides a lifeline to remote equipment when the primary network is down. By connecting device console ports to async NIM ports and providing an alternate network path (such as cellular), engineers can diagnose and resolve outages remotely without dispatching technicians.

Figure 2.6: Out-of-Band Management Architecture

flowchart TD
    NOC["Central NOC\nEngineer"] -->|"SSH over cellular\n(backup path)"| CELL["Cellular\nModem"]
    CELL --> ISR["ISR 4431\nwith NIM-16A"]
    ISR -->|"Async Port 1\nReverse SSH :2001"| SW["Branch Switch\nConsole Port"]
    ISR -->|"Async Port 2\nReverse SSH :2002"| RTR["Branch Router\nConsole Port"]
    ISR -->|"Async Port 3\nReverse SSH :2003"| FW["Branch Firewall\nConsole Port"]
    ISR ~~~|"Primary WAN\n(DOWN)"| WAN["Production\nWAN Link"]

    style WAN fill:#5c1a1a,stroke:#ff5858,color:#fff
    style NOC fill:#0d2137,stroke:#58a6ff,color:#fff
    style ISR fill:#1a3a5c,stroke:#58a6ff,color:#fff
    style CELL fill:#0d2137,stroke:#58a6ff,color:#fff
    style SW fill:#0d2137,stroke:#58a6ff,color:#fff
    style RTR fill:#0d2137,stroke:#58a6ff,color:#fff
    style FW fill:#0d2137,stroke:#58a6ff,color:#fff

Async Dial-Up Backup and Reverse Telnet Applications

Beyond console aggregation, the NIM-16A and NIM-24A support two additional use cases that keep them relevant in specific environments.

Reverse Telnet (and Reverse SSH)

Reverse telnet is a technique where an incoming TCP connection to a specific port on the router is “reversed” — instead of providing a login prompt on the router itself, the connection is forwarded to the device attached to the corresponding async serial port. The convention uses port numbers starting at 2000 plus the async line number.

Configuration for reverse telnet on an async line:

line 1
 transport input telnet ssh
 no exec
 modem InOut

The no exec command prevents the async line from presenting a router login prompt, ensuring that connections are passed through to the attached device.

Modem Aggregation and Dial-Up Backup

While dial-up internet access has largely vanished from consumer use, analog modems remain relevant in specific industrial and telecommunications scenarios:

• Dial-backup for critical circuits: When a primary WAN link fails, the router can automatically dial out through an attached modem to establish a backup connection. Though slow (typically 33.6 kbps to 56 kbps), this provides basic connectivity for management traffic and critical alerts.
• Modem pools for remote access: Some legacy applications — particularly in utilities and government — still use dial-in modem pools for remote maintenance access. The NIM-16A/24A can aggregate modems attached to multiple async ports.

Key Takeaway: Reverse telnet/SSH enables remote console access to devices connected to NIM-16A/24A async ports via TCP port mapping. Modem aggregation and dial-up backup, while declining, persist in critical infrastructure and industrial environments where alternative backup paths are unavailable.

Why Async Serial Persists in Industrial and Rural Deployments

Given the availability of Ethernet, Wi-Fi, and cellular connectivity, one might wonder why anyone would continue to deploy asynchronous serial interfaces. The answer lies in several converging factors.

SCADA and Industrial Control Systems

SCADA systems monitor and control physical processes — water flow, electrical grid status, pipeline pressure, factory machinery. Many SCADA field devices (Remote Terminal Units, or RTUs) communicate over RS-232 serial connections using protocols like Modbus RTU and DNP3. These devices are often installed in harsh environments with 20-to-30-year lifecycles, and replacing them requires not just new hardware but recertification of the entire control system.

The NIM-16A and NIM-24A enable a Cisco router to serve as a gateway between IP networks and serial SCADA devices, translating between TCP/IP and serial protocols. A single NIM-24A can aggregate serial connections from 24 field devices, providing each with network accessibility through the router. [Source: Cisco SCADA design guide]

Air-Gap Security Requirements

Serial connections are inherently point-to-point and non-routable. Unlike an Ethernet port connected to a switched network, a serial port connects to exactly one device. There is no ARP table to poison, no MAC address table to overflow, no VLAN to hop. This physical isolation — the “air gap” — satisfies security requirements in environments where any exposure to a shared network is considered unacceptable risk.

Banking networks, military installations, and classified government systems frequently mandate serial connectivity for this reason. [Source: Cisco SCADA design guide]

Rural and Remote Deployments

In areas without reliable broadband, cellular, or fiber infrastructure, serial connections over existing copper telephone lines may be the only available communications medium. Agricultural monitoring, remote weather stations, and distant pipeline monitoring points often fall into this category. The async serial NIM enables these sites to communicate, even if only at dial-up speeds, using infrastructure that has been in place for decades.

Summary of Persistence Factors:

Key Takeaway: Asynchronous serial persists because of long-lived industrial equipment, air-gap security requirements, regulatory mandates, and the absence of alternatives in remote locations. The NIM-16A and NIM-24A provide the bridge between these serial-dependent environments and modern IP networks.

Chapter Summary

Serial WAN interfaces represent a foundational layer of networking that continues to serve critical roles despite the dominance of Ethernet and IP-based WANs. This chapter covered the journey from physical standards to practical deployment:

Serial standards (RS-232, RS-449, V.35, X.21) define the electrical, mechanical, and procedural characteristics of serial connections. V.35 emerged as the dominant standard for high-speed synchronous WAN links, while RS-232 remains ubiquitous for lower-speed asynchronous applications. The DTE/DCE relationship governs clocking, with the DCE side always providing the timing reference.

Historical WAN protocols (leased lines, Frame Relay, HDLC, PPP) ran over these serial standards to provide enterprise connectivity for decades. Frame Relay’s virtual circuit model reduced costs by sharing infrastructure, while HDLC and PPP provided the Layer 2 framing. These protocols have largely given way to MPLS and SD-WAN, but they persist in specific industries.

The NIM-4T provides four synchronous serial ports on the ISR 4400 platform, supporting all major serial standards through Smart Serial adapter cables. It handles HDLC, PPP, and Frame Relay at up to 8 Mbps per port, serving banking, POS, SCADA, and legacy carrier environments.

The NIM-16A and NIM-24A provide 16 and 24 asynchronous RS-232 ports respectively, purpose-built for console server, out-of-band management, and terminal aggregation applications. They do not support WAN routing protocols — their value lies in providing remote access to device consoles and aggregating serial connections from SCADA field devices.

Together, these NIMs ensure that the Cisco ISR 4000 platform can bridge the gap between modern IP networking and the serial-dependent infrastructure that will remain in production for years to come.

Key Terms

Chapter 3: T1 and E1 Digital Trunk Lines: NIM-8MFT-T1/E1 and Channelized Voice/Data

Learning Objectives

By the end of this chapter, you will be able to:

• Explain the T1 (1.544 Mbps, 24 DS0 channels) and E1 (2.048 Mbps, 30 DS0 channels) digital trunk standards
• Describe what Multi-Flex Trunk (MFT) means and how it enables flexible channelization on the NIM-8MFT-T1/E1
• Differentiate between channelized T1/E1 (individual DS0 voice channels) and unchannelized (single data pipe) operation
• Configure basic T1/E1 controller settings on a Cisco ISR router for voice and data applications

3.1 T1 Standard: North American Digital Hierarchy

Before VoIP existed, telephone companies needed a way to carry multiple phone calls over a single copper pair between central offices. The answer was Time-Division Multiplexing (TDM) — interleaving digitized voice samples from many calls into a single high-speed stream. The T1 circuit, developed by Bell Labs in the 1960s, became the foundation of the North American digital telephone network and remains in widespread use today for connecting enterprise PBX systems, voice gateways, and WAN links.

Think of a T1 line like a revolving door with 24 compartments. Each compartment (channel) carries one person (one phone call) through the door, and the door spins 8,000 times per second. Everyone gets their turn in strict rotation — that is TDM in a nutshell.

3.1.1 T1 Frame Structure: DS1 and DS0 Channels

A T1 circuit operates at 1.544 Mbps and is formally designated as a DS1 (Digital Signal Level 1) in the North American digital hierarchy. It is built from 24 DS0 channels, each running at 64 Kbps.

Here is how the math works:

Each T1 frame is exactly 193 bits long: 24 channels x 8 bits = 192 payload bits, plus 1 framing bit. The network transmits 8,000 frames per second, which matches the Nyquist sampling rate for voice (sampling at 8 kHz to capture frequencies up to 4 kHz).

Figure 3.1: T1 Frame Structure — 193 Bits per Frame

flowchart LR
    F["Framing Bit\n(1 bit)"] --> DS1["DS0 #1\n(8 bits)"]
    DS1 --> DS2["DS0 #2\n(8 bits)"]
    DS2 --> DS3["DS0 #3\n(8 bits)"]
    DS3 --> dots1["..."]
    dots1 --> DS23["DS0 #23\n(8 bits)"]
    DS23 --> DS24["DS0 #24\n(8 bits)"]

    subgraph Total["T1 Frame = 193 bits"]
        F
        DS1
        DS2
        DS3
        dots1
        DS23
        DS24
    end

    Rate["8,000 frames/sec\n= 1.544 Mbps"] ~~~ Total

    style F fill:#d4a017,stroke:#333,color:#000
    style DS1 fill:#2196F3,stroke:#333,color:#fff
    style DS2 fill:#2196F3,stroke:#333,color:#fff
    style DS3 fill:#2196F3,stroke:#333,color:#fff
    style DS23 fill:#2196F3,stroke:#333,color:#fff
    style DS24 fill:#2196F3,stroke:#333,color:#fff
    style Rate fill:#333,stroke:#333,color:#fff

A DS0 (Digital Signal Level 0) is the fundamental building block — a single 64 Kbps channel that carries one digitized voice call using Pulse Code Modulation (PCM). Every phone call on the public switched telephone network ultimately occupies one DS0.

Worked Example: How Many Calls Fit on a T1?

A business orders a T1 from their local carrier to connect their Cisco voice gateway to the PSTN. How many simultaneous phone calls can this T1 carry?

• The T1 has 24 DS0 channels.
• Each DS0 carries one G.711 voice call at 64 Kbps.
• If all 24 channels are allocated for voice: 24 simultaneous calls.
• If the business uses ISDN PRI signaling, one channel (DS0 #24) is reserved for the D-channel (signaling), leaving 23 bearer (B) channels for voice calls.

3.1.2 ESF and D4/SF Framing Formats

T1 circuits use one of two framing formats that define how the framing bits are organized across multiple frames:

D4/Superframe (SF):

• Groups 12 consecutive frames into one superframe
• The 12 framing bits (one per frame) follow a fixed pattern used for frame synchronization
• Uses AMI (Alternate Mark Inversion) line coding (discussed below)
• Older format, still found on some legacy circuits but largely superseded by ESF

Extended Superframe (ESF):

• Groups 24 consecutive frames into one extended superframe
• The 24 framing bits are divided into three functions:
  • 6 bits for CRC-6 error checking (allows the circuit to detect bit errors without taking it out of service)
  • 6 bits for frame synchronization
  • 12 bits for a Facilities Data Link (FDL) — a 4 Kbps management channel that carriers use to monitor circuit performance remotely
• Uses B8ZS line coding (discussed below)
• The modern standard for T1 circuits

Real-world analogy: Think of D4/SF as an old postal system where letters are just numbered sequentially — if one gets lost, you might not notice. ESF is like modern package tracking with checksums and a side channel for delivery status updates.

3.1.3 AMI and B8ZS Line Coding

Line coding determines how binary 1s and 0s are represented as electrical signals on the copper wire. T1 circuits require the signal to have enough transitions (voltage changes) for the receiving equipment to maintain clock synchronization.

AMI (Alternate Mark Inversion):

• Binary 1s alternate between positive and negative voltage pulses
• Binary 0s are represented as zero voltage (no pulse)
• Problem: a long string of consecutive zeros produces no signal transitions, causing the receiver to lose clock sync
• AMI handles this by “robbing” the least significant bit of every 6th frame for signaling (called robbed-bit signaling), but this degrades data integrity

B8ZS (Bipolar with 8-Zero Substitution):

• Same as AMI for normal data
• When 8 consecutive zeros occur, B8ZS replaces them with a special pattern that includes intentional bipolar violations (two consecutive pulses of the same polarity)
• The receiver recognizes these violations as a substitution code and restores the original eight zeros
• This guarantees sufficient signal transitions for clock recovery without robbing any bits
• B8ZS is required for ESF framing and for carrying clear-channel 64 Kbps data (such as ISDN)

3.1.4 How Telcos Provision and Deliver T1 Lines

When a business orders a T1 circuit, the telephone company provisions it as follows:

1. Local loop: A copper pair (or pairs) runs from the carrier’s central office to the customer premises. For distances beyond about 6,000 feet (1.8 km), repeaters amplify the signal.
2. Demarcation point (demarc): The carrier installs a smartjack — a small device at the customer premises that terminates the circuit and provides loopback testing capability.
3. CSU/DSU: The Channel Service Unit / Data Service Unit connects to the smartjack and performs signal regeneration, line coding, and framing. On the NIM-8MFT-T1/E1, the CSU/DSU is integrated into the module — no external device is needed.
4. Circuit testing: The carrier verifies the circuit using loopback tests and performance monitoring via the FDL (on ESF circuits).

Figure 3.3: T1 Circuit Delivery — Physical Path from Carrier to Router

flowchart LR
    CO["Carrier\nCentral Office"] -->|"Copper Pair\n(Local Loop)"| REP["Repeaters\n(if > 6,000 ft)"]
    REP -->|"Copper Pair"| SJ["Smartjack\n(Demarc Point)"]
    SJ -->|"RJ-48C"| NIM["NIM-8MFT-T1/E1\n(Integrated CSU/DSU)"]
    NIM --> ISR["Cisco ISR\n4000 Router"]

    style CO fill:#6a1b9a,stroke:#333,color:#fff
    style REP fill:#555,stroke:#333,color:#fff
    style SJ fill:#d4a017,stroke:#333,color:#000
    style NIM fill:#2196F3,stroke:#333,color:#fff
    style ISR fill:#2e7d32,stroke:#333,color:#fff

Key Takeaway: A T1 circuit delivers 1.544 Mbps over copper using 24 DS0 channels at 64 Kbps each, with ESF framing and B8ZS line coding as the modern standard. The framing bit adds 8 Kbps of overhead, bringing the total to 1.544 Mbps. ESF provides CRC-6 error detection and a management data link that D4/SF lacks.

3.2 E1 Standard: International Digital Hierarchy

While North America standardized on the T1/DS1 format, the rest of the world adopted the E1 standard defined by the ITU-T (International Telecommunication Union). E1 circuits are the backbone of telephone networks across Europe, Asia, Africa, the Middle East, and South America. If your organization has offices outside North America, you will almost certainly encounter E1 circuits.

The key difference is capacity: an E1 carries 30 voice channels compared to T1’s 24, running at a higher aggregate rate of 2.048 Mbps.

3.2.1 E1 Frame Structure: 32 Time Slots

An E1 frame contains 32 time slots (numbered TS0 through TS31), each carrying 8 bits. At 8,000 frames per second:

Unlike T1, where the framing bit is separate from the payload, E1 dedicates specific time slots to overhead:

This gives 30 bearer channels for voice or data, with TS0 for framing and TS16 for signaling. When ISDN PRI is used on an E1, TS16 carries the D-channel, and the remaining 30 time slots serve as B-channels — providing 30 simultaneous calls versus T1’s 23.

Figure 3.2: E1 Frame Structure — 32 Time Slots

flowchart LR
    TS0["TS0\nFraming"] --> TS1["TS1\nBearer"]
    TS1 --> TS2["TS2–TS15\n(14 Bearer\nChannels)"]
    TS2 --> TS16["TS16\nSignaling\n(D-Channel)"]
    TS16 --> TS17["TS17–TS31\n(15 Bearer\nChannels)"]

    subgraph E1Frame["E1 Frame = 256 bits (32 x 8 bits)"]
        TS0
        TS1
        TS2
        TS16
        TS17
    end

    Rate["8,000 frames/sec\n= 2.048 Mbps\n30 Bearer Channels"] ~~~ E1Frame

    style TS0 fill:#d4a017,stroke:#333,color:#000
    style TS1 fill:#2196F3,stroke:#333,color:#fff
    style TS2 fill:#2196F3,stroke:#333,color:#fff
    style TS16 fill:#e74c3c,stroke:#333,color:#fff
    style TS17 fill:#2196F3,stroke:#333,color:#fff
    style Rate fill:#333,stroke:#333,color:#fff

Real-world analogy: If a T1 is a 24-compartment revolving door, an E1 is a 32-compartment door where two compartments are reserved for the building manager (framing and signaling) and the other 30 are for people (calls).

3.2.2 CRC-4 Multiframe Alignment and Error Detection

E1 circuits use CRC-4 (Cyclic Redundancy Check, 4-bit) for error detection. The CRC-4 multiframe structure groups 16 consecutive E1 frames into one multiframe:

• Frames 0–7 form submultiframe I
• Frames 8–15 form submultiframe II
• A 4-bit CRC is calculated over submultiframe I and transmitted in the TS0 bits of submultiframe II (and vice versa)
• This allows both ends to detect bit errors on the circuit continuously

CRC-4 is the standard framing mode for E1 circuits worldwide. Some older installations may use non-CRC4 framing, but this is increasingly rare. When configuring an E1 controller on a Cisco router, you specify the framing as either crc4 or no-crc4.

3.2.3 HDB3 Line Coding

E1 circuits use HDB3 (High Density Bipolar 3-zeros maximum) line coding, which serves the same purpose as B8ZS on T1 — ensuring sufficient signal transitions for clock recovery.

• HDB3 is based on AMI but substitutes a special code whenever 4 consecutive zeros occur (compared to B8ZS’s threshold of 8 zeros)
• The substitution includes an intentional bipolar violation that the receiver detects and decodes
• HDB3 is more aggressive than B8ZS at eliminating long zero runs, which reflects the European standard’s stricter requirements for clock density

3.2.4 Regional Deployment

E1 is the dominant digital trunk standard outside North America and Japan:

When deploying Cisco ISR routers in international locations, you must select the correct NIM variant (T1 or E1) and configure the controller accordingly. The NIM-8MFT-T1/E1 supports both standards, but each physical port must be configured for one or the other — you cannot mix T1 and E1 on the same port.

Key Takeaway: E1 circuits carry 2.048 Mbps across 32 time slots, with 30 usable bearer channels (TS0 for framing, TS16 for signaling). CRC-4 multiframing provides continuous error detection, and HDB3 line coding ensures clock synchronization. E1 is the standard outside North America, offering 25% more voice capacity than T1.

3.3 Multi-Flex Trunk (MFT) and Channelization

Now that you understand the T1 and E1 standards, the next question is: how do you actually use those 24 or 30 channels? This is where Multi-Flex Trunk (MFT) technology comes in. MFT gives you the flexibility to assign each DS0 channel to different services — voice calls, data connections, or ISDN signaling — on the same physical T1 or E1 circuit.

Think of MFT like a modular office building. You have a fixed number of rooms (DS0 channels), but you can designate each room for whatever purpose you need: some for phone operators, some for data processing, some for meetings. You can rearrange the assignments as your needs change.

3.3.1 What MFT Means: Flexible Allocation of DS0 Channels

Multi-Flex Trunk (MFT) refers to the ability of a Cisco NIM to flexibly allocate individual DS0 time slots within a T1 or E1 circuit to different functions. Rather than dedicating the entire circuit to a single purpose, MFT lets you:

• Assign some DS0s to carry voice calls (via channel-associated signaling or ISDN PRI)
• Assign other DS0s to carry data (as a serial WAN link)
• Reassign channels as business requirements change — without swapping hardware

This flexibility is the “multi” and “flex” in Multi-Flex Trunk. On older equipment, a T1 card was either a voice card or a data card. MFT modules combine both capabilities in a single NIM.

3.3.2 Channelized Mode: Individual DS0s for Voice Calls

In channelized mode, the T1 or E1 is divided into individual DS0 channels, each carrying a separate voice call or signaling path. This is the traditional mode of operation for voice trunking.

There are three primary ways to channelize a T1/E1 on Cisco IOS-XE:

1. CAS (Channel-Associated Signaling) via ds0-group:

• Each DS0 or group of DS0s is assigned a signaling type such as E&M (Ear and Mouth, for tie trunks), FXS (Foreign Exchange Station, to connect analog phones), or FXO (Foreign Exchange Office, to connect to analog PSTN lines)
• Signaling travels in-band, using robbed-bit signaling on T1 or TS16 on E1
• Configured with the ds0-group command under the controller

2. ISDN PRI via pri-group:

• Uses Common Channel Signaling (CCS) — one dedicated DS0 carries all signaling (the D-channel) while the remaining DS0s are B-channels for bearer traffic
• T1 PRI: 23B + 1D (timeslots 1-23 for voice, timeslot 24 for signaling)
• E1 PRI: 30B + 1D (timeslots 1-15 and 17-31 for voice, timeslot 16 for signaling)
• Configured with the pri-group command

3. Data channel groups via channel-group:

• One or more DS0s are bonded into a serial data interface
• Used for WAN connectivity (Frame Relay, PPP, HDLC)
• Configured with the channel-group command

Worked Example: CAS Voice Trunking on T1

A company wants to connect 12 analog tie-trunk circuits between two offices using E&M wink-start signaling on a T1. The configuration assigns DS0s 1 through 12 to voice:

controller T1 0/2/0
 framing esf
 linecode b8zs
 ds0-group 1 timeslots 1-12 type e&m-wink-start

This creates a voice port for timeslots 1–12. The remaining timeslots (13–24) are unused and available for other assignments.

3.3.3 Unchannelized Mode: Single Data Pipe

In unchannelized mode, all DS0 channels on the T1 or E1 are bonded together into a single high-speed serial data link. No individual channel assignments are made — the entire circuit acts as one pipe.

• T1 unchannelized: all 24 DS0s bonded = 1.536 Mbps usable data rate (the framing overhead reduces the total from 1.544 Mbps)
• E1 unchannelized: all 30 bearer time slots bonded = 1.920 Mbps usable data rate (TS0 and TS16 remain as overhead)

This mode is used when the T1 or E1 serves purely as a WAN data link — for example, connecting two routers over a leased line using PPP or HDLC encapsulation.

controller T1 0/2/0
 framing esf
 linecode b8zs
 channel-group 0 timeslots 1-24

This creates a single serial interface (Serial0/2/0:0) carrying all 24 timeslots as one data stream.

3.3.4 Mixed Mode: Splitting Between Voice and Data

The real power of MFT is mixed mode — splitting the same T1 or E1 between voice and data simultaneously. This is where MFT truly earns its name.

Worked Example: Mixed Voice and Data on a Single T1

A branch office needs 12 voice channels for PSTN calls and a data link for WAN connectivity, but only has one T1 circuit from the carrier. With MFT, they can split the T1:

• Timeslots 1–12: Voice calls using E&M wink-start signaling
• Timeslots 13–24: Data link for WAN traffic

controller T1 0/2/0
 framing esf
 linecode b8zs
 ds0-group 1 timeslots 1-12 type e&m-wink-start
 channel-group 1 timeslots 13-24

This creates both a voice port (for the 12 E&M channels) and a serial interface (Serial0/2/0:1 for the data channel group) on the same physical T1. The voice and data traffic coexist on the same circuit without interference because TDM keeps each timeslot strictly separated.

Figure 3.4: MFT Channelization Modes — T1 Example (24 DS0 Channels)

flowchart TD
    T1["T1 Circuit\n24 DS0 Channels"]

    T1 --> CH["Channelized Mode"]
    T1 --> UN["Unchannelized Mode"]
    T1 --> MX["Mixed Mode"]

    CH --> CAS["CAS (ds0-group)\n24 individual voice channels\nE&M / FXS / FXO signaling"]
    CH --> PRI["ISDN PRI (pri-group)\n23 B-channels + 1 D-channel\nCommon channel signaling"]

    UN --> DATA["Single Serial Link\n(channel-group)\nAll 24 DS0s bonded\n= 1.536 Mbps data pipe"]

    MX --> VOICE["DS0s 1–12\nVoice (ds0-group)\n12 voice channels"]
    MX --> WANDATA["DS0s 13–24\nData (channel-group)\n768 Kbps WAN link"]

    style T1 fill:#2196F3,stroke:#333,color:#fff
    style CH fill:#2e7d32,stroke:#333,color:#fff
    style UN fill:#d4a017,stroke:#333,color:#000
    style MX fill:#6a1b9a,stroke:#333,color:#fff
    style CAS fill:#2e7d32,stroke:#333,color:#fff
    style PRI fill:#2e7d32,stroke:#333,color:#fff
    style DATA fill:#d4a017,stroke:#333,color:#000
    style VOICE fill:#2e7d32,stroke:#333,color:#fff
    style WANDATA fill:#d4a017,stroke:#333,color:#000

Key Takeaway: Multi-Flex Trunk (MFT) allows flexible allocation of individual DS0 channels on a T1 or E1 to voice (via ds0-group or pri-group) or data (via channel-group). Mixed mode lets a single circuit carry both voice and data simultaneously, maximizing the value of each T1/E1 line.

3.4 NIM-8MFT-T1/E1 Hardware and Configuration

The NIM-8MFT-T1/E1 is Cisco’s Network Interface Module that brings T1 and E1 connectivity to the ISR 4000 series routers. It is the hardware platform that implements everything discussed in this chapter — T1/E1 framing, line coding, channelization, and MFT flexibility.

3.4.1 NIM-8MFT-T1/E1 Specifications

The NIM-8MFT-T1/E1 is a modular card that installs into a NIM slot on Cisco ISR 4321, 4331, 4351, 4431, and 4451 routers. Key specifications include:

The integrated CSU/DSU is a significant advantage. On older platforms, you needed a separate external CSU/DSU device between the router and the telco smartjack. The NIM-8MFT-T1/E1 handles signal regeneration, line coding, and loopback testing directly on the module.

Drop-and-insert multiplexing is an advanced feature that allows the router to extract (drop) specific timeslots from an incoming T1/E1 for local processing while passing the remaining timeslots through (insert) to another T1/E1 port. This is useful in tandem switching scenarios.

Real-world analogy: The NIM-8MFT-T1/E1 is like a multi-bay mail sorting station. Each bay (port) connects to a different mail route (T1/E1 circuit). Within each bay, you can sort individual letters (DS0 channels) to different destinations — some to the voice system, some to the data network, some passed straight through to another bay.

3.4.2 Controller Configuration Commands for T1 and E1

Configuring a T1 or E1 port on IOS-XE follows a consistent pattern: you enter the controller configuration mode, set framing and line coding, then define how the timeslots are used.

Network Clock Synchronization

On ISR 4000 series routers, you must first enable network clock synchronization so the router can recover timing from the T1/E1 circuit:

Router(config)# network-clock synchronization automatic

This tells the router to automatically select the best available clock source from its T1/E1 interfaces. Without this command, the router may experience clock slips that cause bit errors and call quality issues.

T1 Controller — ISDN PRI Configuration

This is the most common configuration for connecting a Cisco voice gateway to the PSTN via T1:

Router(config)# controller T1 0/2/0
Router(config-controller)# framing esf
Router(config-controller)# linecode b8zs
Router(config-controller)# pri-group timeslots 1-24 voice-dsp

Explanation:

• controller T1 0/2/0 — enters configuration for the T1 controller in slot 0, subslot 2, port 0
• framing esf — sets Extended Superframe framing (must match what the carrier provisions)
• linecode b8zs — sets B8ZS line coding (must match the carrier)
• pri-group timeslots 1-24 voice-dsp — allocates all 24 timeslots as an ISDN PRI group; timeslot 24 automatically becomes the D-channel, leaving 23 B-channels for voice calls

E1 Controller — ISDN PRI Configuration

For an E1 circuit, the configuration is similar but reflects the E1 framing, line coding, and timeslot numbering:

Router(config)# controller E1 0/2/0
Router(config-controller)# framing crc4
Router(config-controller)# linecode hdb3
Router(config-controller)# pri-group timeslots 1-15,17-31 voice-dsp

Explanation:

• framing crc4 — sets CRC-4 multiframe alignment
• linecode hdb3 — sets HDB3 line coding
• pri-group timeslots 1-15,17-31 — allocates 30 bearer timeslots as an ISDN PRI; note that TS0 (framing) and TS16 (D-channel) are excluded from the range because IOS-XE handles them automatically

T1 Controller — CAS Voice Configuration

For Channel-Associated Signaling with E&M wink-start:

Router(config)# controller T1 0/2/0
Router(config-controller)# framing esf
Router(config-controller)# linecode b8zs
Router(config-controller)# ds0-group 1 timeslots 1-12 type e&m-wink-start

T1 Controller — Unchannelized Data Configuration

For a pure WAN data link:

Router(config)# controller T1 0/2/0
Router(config-controller)# framing esf
Router(config-controller)# linecode b8zs
Router(config-controller)# channel-group 0 timeslots 1-24

Configuration Summary Table

3.4.3 Troubleshooting T1/E1 Alarms: Red, Yellow, and Blue

T1 and E1 circuits use a standardized alarm system to indicate fault conditions. Understanding these alarms is critical for troubleshooting — they tell you where the problem is.

Red Alarm (LOF/LOS)

• What it means: The local end has lost the incoming signal. Either there is a Loss of Signal (LOS) — no electrical signal is being received — or a Loss of Frame (LOF) — a signal is present but the framing pattern cannot be detected.
• Common causes: Cut cable, failed CSU/DSU, incorrect framing configuration (e.g., you configured ESF but the carrier provisioned D4/SF), unplugged cable.
• Where the problem is: Between the far end and the local end — the local receiver is not getting a valid signal.

Yellow Alarm (RAI — Remote Alarm Indication)

• What it means: The far end is telling you that it is receiving errors from your direction. The remote side is sending a Remote Alarm Indication back to alert you that something is wrong with the signal you are transmitting.
• Common causes: Line coding mismatch (e.g., you configured B8ZS but the carrier expects AMI), excessive bit errors from cable degradation, clock synchronization issues.
• Where the problem is: The local transmitter or the cable between the local end and the far end. Your signal is reaching the far end, but it is degraded or misconfigured.

Blue Alarm (AIS — Alarm Indication Signal)

• What it means: An Alarm Indication Signal is being received — an unframed all-ones signal that indicates an upstream failure. The device sending the AIS has detected a fault further upstream and is sending the blue alarm downstream to notify all equipment that the circuit is out of service.
• Common causes: Carrier network equipment failure between central offices, upstream T1/E1 multiplexer failure.
• Where the problem is: Upstream in the carrier’s network, not at your premises. Contact the carrier.

Real-world analogy: Think of alarms as traffic signals on a highway:

• Red alarm = the road ahead is completely blocked (you cannot see any traffic coming toward you)
• Yellow alarm = oncoming drivers are flashing their lights to warn you that your headlights are off (the other end sees a problem with your signal)
• Blue alarm = the highway department has put up a “road closed” sign miles upstream (the problem is in the carrier’s network)

Alarm Summary Table

Figure 3.5: T1/E1 Alarm Troubleshooting — Fault Location by Alarm Color

flowchart TD
    ALARM["T1/E1 Alarm Detected"]
    ALARM --> RED{"Red Alarm?\n(LOF/LOS)"}
    ALARM --> YELLOW{"Yellow Alarm?\n(RAI)"}
    ALARM --> BLUE{"Blue Alarm?\n(AIS)"}

    RED -->|"Yes"| REDACT["Fault: Local Receive Path\n\nCheck:\n- Physical cable connection\n- Framing config mismatch\n- Local CSU/DSU hardware"]

    YELLOW -->|"Yes"| YELLOWACT["Fault: Local Transmit Path\n\nCheck:\n- Line coding mismatch\n- Clock source config\n- Cable quality/degradation"]

    BLUE -->|"Yes"| BLUEACT["Fault: Upstream Carrier Network\n\nAction:\n- Contact service provider\n- Carrier equipment failure\n- Nothing to fix locally"]

    style ALARM fill:#555,stroke:#333,color:#fff
    style RED fill:#e74c3c,stroke:#333,color:#fff
    style YELLOW fill:#f39c12,stroke:#333,color:#000
    style BLUE fill:#2196F3,stroke:#333,color:#fff
    style REDACT fill:#e74c3c,stroke:#333,color:#fff
    style YELLOWACT fill:#f39c12,stroke:#333,color:#000
    style BLUEACT fill:#2196F3,stroke:#333,color:#fff

Verification Commands

To check the status of a T1 or E1 controller and view any active alarms:

Router# show controller T1 0/2/0

This command displays:

• Current framing and line coding settings
• Alarm status (red, yellow, blue, or no alarms)
• Error counters (CRC errors, framing errors, line code violations)
• Clock source and synchronization status
• Loopback status

For E1 controllers:

Router# show controller E1 0/2/0

Worked Example: Diagnosing a Yellow Alarm

A network engineer configures a new T1 PRI on an ISR 4331 and sees a yellow alarm. The show controller T1 0/2/0 output shows:

T1 0/2/0 is up.
  Framing is ESF, Line Code is B8ZS
  ...
  Receiver has loss of signal.
  Far end has Yellow alarm.

Wait — both a loss-of-signal and a yellow alarm? This suggests:

1. The local end is not receiving a signal (red alarm condition locally)
2. The far end is also reporting errors (yellow alarm from far end)

The engineer checks the physical cable and finds it is connected to the wrong port on the smartjack. After reseating the cable to the correct port, both alarms clear and the controller shows “no alarms.”

The lesson: always check the physical layer first. Most T1/E1 alarm conditions trace back to cabling, incorrect port connections, or mismatched framing/line coding settings between the router and the carrier.

Key Takeaway: The NIM-8MFT-T1/E1 provides up to 8 T1 or E1 ports with integrated CSU/DSU, supporting channelized, unchannelized, and mixed-mode operation. Configuration follows a controller-based model in IOS-XE using framing, linecode, and channel assignment commands. Red, yellow, and blue alarms indicate where a fault lies — local receive path, local transmit path, or upstream carrier network, respectively.

Chapter Summary

This chapter covered the two dominant digital trunk standards in telecommunications and how Cisco’s NIM-8MFT-T1/E1 brings them to the ISR 4000 platform:

• T1 circuits deliver 1.544 Mbps over 24 DS0 channels, using ESF framing and B8ZS line coding as the modern standard. Each DS0 carries 64 Kbps — enough for one digitized voice call. T1 is the standard in North America.

• E1 circuits deliver 2.048 Mbps across 32 time slots, with 30 usable bearer channels (TS0 for framing, TS16 for signaling). CRC-4 multiframing provides error detection, and HDB3 line coding ensures clock synchronization. E1 is the international standard used outside North America.

• Multi-Flex Trunk (MFT) allows flexible allocation of DS0 channels to different services — voice, data, or signaling — on the same physical circuit. Channelized mode assigns individual DS0s to voice calls, unchannelized mode bonds all channels into a single data pipe, and mixed mode supports both simultaneously.

• The NIM-8MFT-T1/E1 supports 1, 2, 4, or 8 T1/E1 ports with integrated CSU/DSU, configured via controller commands in IOS-XE. Troubleshooting relies on understanding the alarm hierarchy: red (local receive failure), yellow (far-end reports your transmit errors), and blue (upstream carrier failure).

Key Terms

Chapter 4: ISDN Interfaces — PRI and BRI for Digital Voice Signaling

Learning Objectives

By the end of this chapter, you will be able to:

• Explain ISDN PRI signaling (23B+D for T1-PRI, 30B+D for E1-PRI) and its role in enterprise voice connectivity
• Explain ISDN BRI signaling (2B+D) and its use cases for small offices and backup WAN links
• Differentiate between NT (Network Termination) and TE (Terminal Equipment) modes on BRI interfaces
• Describe how the NIM-8CE1T1-PRI and NIM-2BRI-NT/TE modules terminate ISDN circuits and convert calls to VoIP

4.1 ISDN Fundamentals: The Digital Telephone Standard

What ISDN Is and Why It Was Developed

Integrated Services Digital Network (ISDN) is a set of international telecommunications standards for transmitting voice, video, and data simultaneously over the traditional public switched telephone network (PSTN). Developed in the 1980s by the ITU-T (International Telecommunication Union — Telecommunication Standardization Sector), ISDN was designed to replace the analog “last mile” with end-to-end digital connectivity.

Think of the analog telephone network like a dirt road: it gets the job done, but only one kind of traffic can travel it at a time, and quality degrades over distance. ISDN paved that dirt road, turning it into a multi-lane highway where voice, data, and signaling each have their own dedicated lane. The result was faster call setup, clearer audio, and the ability to carry data alongside voice — all on the same copper pair already running to your building.

ISDN matters in the context of Cisco voice infrastructure because thousands of enterprises still rely on ISDN trunks to connect their Private Branch Exchange (PBX) or Cisco Unified Communications Manager (CUCM) to the PSTN. Even as SIP trunking grows, ISDN PRI circuits remain the backbone of voice connectivity in call centers, hospitals, hotels, and government agencies worldwide.

B-Channels vs. D-Channel

ISDN organizes its bandwidth into two types of channels:

The B-channel (Bearer) is where the actual conversation travels. Each B-channel provides 64 Kbps of bandwidth — exactly one uncompressed digitized voice call using the G.711 codec. If you need two simultaneous calls, you need two B-channels.

The D-channel (Delta) is the control plane. It carries signaling messages: “set up a call to this number,” “the far end is ringing,” “the caller hung up.” The D-channel uses a protocol called Q.931 to manage call setup, maintenance, and teardown. The D-channel never carries voice — it is purely the traffic cop directing calls onto available B-channels.

Real-world analogy: Imagine a restaurant kitchen. The B-channels are the burners on the stove — each one cooks one dish (carries one call). The D-channel is the expediter who reads incoming orders, assigns them to open burners, and tells the waitstaff when dishes are ready. Without the expediter (D-channel), the kitchen has no way to coordinate which burner handles which order.

Q.931 Call Setup Signaling

Q.931 is the ITU-T Layer 3 signaling protocol used on the ISDN D-channel. It manages the entire lifecycle of a call through a well-defined sequence of messages:

1. SETUP — The calling side sends a SETUP message containing the called number, calling number, and bearer capability (voice, data, etc.).
2. CALL PROCEEDING — The switch acknowledges it received the SETUP and is processing the request.
3. ALERTING — The far-end phone is ringing.
4. CONNECT — The called party answered; a B-channel is now allocated for the conversation.
5. CONNECT ACK — Acknowledgment that the connection is established.
6. DISCONNECT — One side hangs up, initiating teardown.
7. RELEASE / RELEASE COMPLETE — The B-channel is freed for reuse.

This structured signaling is a major advantage over analog lines, where “signaling” consists of crude voltage changes and in-band tones that are slow and error-prone.

Worked Example — Q.931 Call Flow:

Calling Side (Router)          PSTN Switch           Called Side
      |                            |                      |
      |--- SETUP ----------------->|                      |
      |<-- CALL PROCEEDING --------|                      |
      |                            |--- SETUP ----------->|
      |                            |<-- ALERTING ---------|
      |<-- ALERTING ---------------|                      |
      |                            |<-- CONNECT ----------|
      |<-- CONNECT ----------------|                      |
      |--- CONNECT ACK ----------->|                      |
      |                            |                      |
      |========= B-channel (voice) allocated =============|
      |                            |                      |
      |--- DISCONNECT ------------>|                      |
      |<-- RELEASE ----------------|                      |
      |--- RELEASE COMPLETE ------>|                      |

In this sequence, the entire call setup — from dialing to hearing ringback — typically completes in 1 to 3 seconds. Compare that to analog lines where call setup can take 10 seconds or more due to pulse/tone dialing and in-band signaling delays.

Figure 4.1: Q.931 Call Setup and Teardown Sequence

sequenceDiagram
    participant Caller as Calling Side (Router)
    participant PSTN as PSTN Switch
    participant Called as Called Side

    Caller->>PSTN: SETUP (called number, ANI, bearer capability)
    PSTN-->>Caller: CALL PROCEEDING
    PSTN->>Called: SETUP
    Called-->>PSTN: ALERTING (ringing)
    PSTN-->>Caller: ALERTING
    Called-->>PSTN: CONNECT (answered)
    PSTN-->>Caller: CONNECT
    Caller->>PSTN: CONNECT ACK

    Note over Caller,Called: B-channel allocated — voice conversation active

    Caller->>PSTN: DISCONNECT
    PSTN-->>Caller: RELEASE
    Caller->>PSTN: RELEASE COMPLETE
    Note over Caller,Called: B-channel freed for reuse

ISDN vs. Analog Advantages

Key Takeaway: ISDN replaced analog signaling with a structured, digital framework of B-channels for user traffic and a D-channel for call control using Q.931. This separation of voice and signaling was a foundational step toward modern telephony — and the same B-channel/D-channel architecture still underpins PRI trunks in enterprise networks today.

4.2 PRI — Primary Rate Interface

T1-PRI: 23B+D = 1.544 Mbps

Primary Rate Interface (PRI) is the “enterprise-grade” flavor of ISDN. In North America and Japan, PRI rides on a T1 circuit — a 1.544 Mbps digital link carried over two pairs of copper or fiber. The T1 frame is divided into 24 time slots of 64 Kbps each:

• 23 B-channels carry voice calls (time slots 1–23)
• 1 D-channel carries Q.931 signaling (time slot 24)

This means a single T1-PRI supports 23 simultaneous phone calls. Need more capacity? Add another T1-PRI. In a multi-PRI configuration, the D-channel on one span can even control B-channels on adjacent spans through a feature called NFAS (Non-Facility Associated Signaling), squeezing out an extra B-channel per additional span.

Real-world analogy: A T1-PRI is like a 24-lane toll plaza. Twenty-three lanes are open for cars (voice calls), and one lane is reserved for the toll booth supervisor (signaling). The supervisor directs which cars go into which lanes, collects information, and manages the flow — but never carries passengers.

E1-PRI: 30B+D = 2.048 Mbps

Outside North America, the E1 standard prevails. An E1 circuit runs at 2.048 Mbps and divides into 32 time slots:

• 30 B-channels carry voice calls (time slots 1–15 and 17–31)
• 1 D-channel carries signaling (time slot 16)
• 1 framing/sync channel (time slot 0)

A single E1-PRI therefore supports 30 simultaneous calls — seven more than T1-PRI.

Figure 4.2: T1-PRI vs E1-PRI Time Slot Structure

graph TD
    subgraph T1["T1-PRI (1.544 Mbps — 24 Time Slots)"]
        T1B["Time Slots 1-23: 23 B-Channels (Voice)"]
        T1D["Time Slot 24: D-Channel (Signaling)"]
    end

    subgraph E1["E1-PRI (2.048 Mbps — 32 Time Slots)"]
        E1F["Time Slot 0: Framing / Sync"]
        E1B1["Time Slots 1-15: 15 B-Channels (Voice)"]
        E1D["Time Slot 16: D-Channel (Signaling)"]
        E1B2["Time Slots 17-31: 15 B-Channels (Voice)"]
    end

    T1B --- T1D
    E1F --- E1B1 --- E1D --- E1B2

NIM-8CE1T1-PRI: CE Designation and Capabilities

The NIM-8CE1T1-PRI is Cisco’s high-density channelized voice module for the ISR 4000 and Catalyst 8000 series routers. Let’s break down the naming:

• NIM — Network Interface Module (the form factor for ISR 4000/Cat 8000)
• 8C — 8 channelized ports
• E1T1 — Each port supports either E1 or T1 (software-selectable)
• PRI — Designed for ISDN PRI operation

The “CE” in the designation stands for “channelized E1/T1” — meaning the module can break each physical port into individual 64 Kbps time slots (channels), which is exactly what PRI requires to allocate B-channels and D-channels.

Key hardware features:

With 8 ports of E1-PRI, a single NIM-8CE1T1-PRI card can terminate up to 240 simultaneous calls (8 x 30 B-channels). For T1-PRI, the maximum is 184 calls (8 x 23 B-channels). This makes it suitable for medium-to-large enterprise deployments where a router serves as an ISDN-to-VoIP gateway.

The integrated CSU/DSU is worth highlighting. In older deployments, an external CSU/DSU box sat between the telco’s demarcation point and the router, adding cost and a potential failure point. The NIM-8CE1T1-PRI eliminates that box by building the CSU/DSU directly onto the module.

Enterprise Use Cases

Call Centers: A 200-seat call center might deploy four T1-PRI spans (92 B-channels) to handle peak call volumes. The NIM-8CE1T1-PRI provides all four spans from a single module, keeping the router chassis compact.

Hotels: A 500-room hotel typically provisions PRI trunks scaled to 10–15% of room count. Two E1-PRI spans (60 B-channels) handle peak checkout-morning call volumes. The hotel’s Cisco router terminates these PRI circuits and converts calls to SIP for delivery to the IP-PBX.

Hospitals: Healthcare facilities require reliable voice for code alerts, nurse call systems, and external lines. PRI is preferred over SIP trunking in some regulated environments because the dedicated circuit guarantees bandwidth — there is no contention with internet traffic.

Government/Military: Many government agencies mandate circuit-switched voice for security or regulatory compliance. ISDN PRI with encryption provides a known, auditable path from endpoint to endpoint.

Key Takeaway: PRI is the workhorse of enterprise ISDN voice — T1-PRI delivers 23 calls per span in North America, while E1-PRI delivers 30 calls per span elsewhere. The NIM-8CE1T1-PRI module packs up to 8 channelized ports with integrated CSU/DSU into a single NIM slot, making it the go-to card for high-density ISDN-to-VoIP gateway deployments on the ISR 4000 and Catalyst 8000 platforms.

4.3 BRI — Basic Rate Interface

BRI Structure: 2B+D = 144 Kbps Total

Where PRI is the enterprise trunk, Basic Rate Interface (BRI) is the access-line equivalent — designed for individual desks, small offices, or low-bandwidth connections. A BRI circuit provides:

• 2 B-channels at 64 Kbps each (128 Kbps total bearer capacity)
• 1 D-channel at 16 Kbps (signaling)

The total user bandwidth is 144 Kbps, though the physical line rate is 192 Kbps when overhead framing bits (48 Kbps) are included.

Real-world analogy: If PRI is a 24-lane toll plaza, BRI is a two-lane country road with a small guard shack. Two cars (calls) can travel simultaneously, and the guard (D-channel) manages traffic. It is modest — but perfectly adequate for a small office that only needs two concurrent phone lines.

Both B-channels can be used independently — for example, one voice call and one data session — or bonded together using Multilink PPP (MLPPP) to create a single 128 Kbps data link for backup WAN connectivity.

NT vs. TE Mode

BRI introduces an important concept not found in PRI: the distinction between NT (Network Termination) and TE (Terminal Equipment) modes. This determines which side of the ISDN connection the device emulates.

NT mode — The router behaves like the telephone company. It provides clocking, power (phantom feed), and Layer 1 activation to the devices connected to it. Use NT mode when the router is acting as a mini-switch connecting ISDN phones or terminal adapters directly.

TE mode — The router behaves like a telephone or terminal. It receives clocking and power from the network side. Use TE mode when the router connects upstream to a telco ISDN line or an NT1 device.

Real-world analogy: Think of an electrical outlet and a plug. NT mode is the outlet — it supplies power and sets the rules. TE mode is the plug — it conforms to whatever the outlet provides. You cannot plug two outlets together or two plugs together; one side must be NT and the other must be TE.

The ISDN reference model defines several reference points that clarify where NT and TE sit:

   TE1 ----S---- NT2 ----T---- NT1 ----U---- Network
   (phone)      (PBX)        (line term)    (telco CO)

Figure 4.3: ISDN BRI Reference Model — NT and TE Reference Points

flowchart LR
    TE1["TE1\n(ISDN Phone)"]
    NT2["NT2\n(PBX)"]
    NT1["NT1\n(Line Termination)"]
    CO["Telco Central\nOffice"]

    TE1 -- "S interface\n(4-wire)" --> NT2
    NT2 -- "T interface\n(4-wire)" --> NT1
    NT1 -- "U interface\n(2-wire)" --> CO

    style TE1 fill:#1a3a5c,stroke:#58a6ff,color:#ffffff
    style NT2 fill:#1a3a5c,stroke:#58a6ff,color:#ffffff
    style NT1 fill:#1a3a5c,stroke:#58a6ff,color:#ffffff
    style CO fill:#0d2137,stroke:#58a6ff,color:#ffffff

On a Cisco router with a BRI-NT/TE module, you configure the mode to match your deployment:

• Router connects to telco BRI line → configure as TE (the router is terminal equipment receiving service)
• Router connects directly to ISDN phones → configure as NT (the router provides the ISDN service)

NIM-2BRI-NT/TE Hardware

The NIM-2BRI-NT/TE is Cisco’s BRI module for the ISR 4000 and Catalyst 8000 platforms. Key specifications:

The dual-mode capability (NT/TE per port) is what makes this module versatile. Port 0 might connect upstream to the telco in TE mode, while Port 1 connects downstream to a legacy ISDN phone in NT mode — all on the same card.

Use Cases for BRI

European and Japanese Small Offices: BRI was widely deployed in Europe and Japan, where the telco infrastructure favored ISDN over analog lines earlier than in North America. A small law firm in Germany with 5 employees might have two BRI lines (4 B-channels = 4 concurrent calls) — sufficient for daily operations. The NIM-2BRI-NT/TE on an ISR 4331 terminates these lines and converts calls to SIP for the office IP phone system.

Video Conferencing (Legacy): Early video conferencing systems (Polycom ViewStation, Tandberg) used bonded BRI channels (2 x 64 = 128 Kbps) for compressed video calls. While modern systems use IP, some legacy equipment still relies on ISDN BRI for video endpoints in courtrooms, telemedicine suites, and government facilities.

Dial Backup for WAN Links: BRI provides an inexpensive backup path when a primary WAN circuit fails. The router monitors the primary link and, upon detecting failure, automatically dials out over the BRI D-channel to establish a 64 or 128 Kbps backup connection. This “dial-on-demand routing” (DDR) kept remote sites connected during outages at minimal recurring cost, since ISDN BRI was billed per-minute only when active.

Worked Example — BRI Dial Backup Scenario:

A retail branch office has a primary 10 Mbps Metro Ethernet WAN link. If it fails, the ISR 4321 router with a NIM-2BRI-NT/TE module dials the headquarters router over BRI:

1. Primary link goes down — router detects loss of keepalives.
2. Router activates BRI interface, sends Q.931 SETUP on D-channel.
3. Two B-channels connect (128 Kbps bonded via MLPPP).
4. Critical traffic (point-of-sale transactions, voice) routes over the BRI backup.
5. When the primary link recovers, the router tears down the BRI call to stop per-minute charges.

This pattern was especially popular in the 2000s and early 2010s. While LTE/5G backup has largely replaced BRI for this role, many existing deployments remain in service.

Figure 4.4: BRI Dial Backup — State Transitions

stateDiagram-v2
    [*] --> PrimaryActive: Primary WAN link up
    PrimaryActive --> DetectFailure: Loss of keepalives
    DetectFailure --> BRIDialing: Activate BRI interface
    BRIDialing --> BRIConnected: Q.931 SETUP / CONNECT\n(128 Kbps MLPPP)
    BRIConnected --> PrimaryRecovery: Primary link restored
    PrimaryRecovery --> PrimaryActive: Tear down BRI call\n(stop per-minute charges)
    BRIConnected --> BRIConnected: Critical traffic flows\nover BRI backup

Key Takeaway: BRI provides two B-channels and a D-channel for small-scale ISDN connectivity. The NT/TE mode distinction determines whether the router acts as the network provider or the terminal device. The NIM-2BRI-NT/TE module supports both modes per port, making it suitable for small office voice termination, legacy video conferencing, and dial backup on ISR 4000 and Catalyst 8000 routers.

4.4 ISDN-to-VoIP Conversion on the Router

How the Router Terminates ISDN and Bridges to SIP

A Cisco ISR equipped with a NIM-8CE1T1-PRI or NIM-2BRI-NT/TE serves as a media gateway — it sits at the boundary between the circuit-switched ISDN world and the packet-switched IP world. Here is what happens when an inbound ISDN call arrives and must be delivered as a VoIP call:

1. Physical Layer: The NIM module receives the T1/E1/BRI electrical signal and recovers clocking.
2. ISDN Layer 2 (Q.921/LAPD): The D-channel data link is established. LAPD frames carry Q.931 messages reliably.
3. ISDN Layer 3 (Q.931): A SETUP message arrives on the D-channel with the called number (DNIS) and calling number (ANI).
4. Dial-Peer Matching: The router’s call routing engine matches the called number against configured dial peers to determine where to send the call.
5. VoIP Call Setup: The router initiates a SIP INVITE (or H.323 Setup) toward the VoIP destination — typically CUCM, a SIP proxy, or a SIP trunk provider.
6. Media Transcoding: The PVDM4 DSP module converts the ISDN B-channel’s G.711 PCM audio into the negotiated VoIP codec (G.711, G.729, etc.) and encapsulates it in RTP packets.
7. Call Established: Voice flows bidirectionally — ISDN PCM on one side, RTP/UDP/IP on the other — with the router performing real-time conversion.

Real-world analogy: The router is a simultaneous interpreter at a United Nations meeting. One delegate speaks French (ISDN/TDM), and the interpreter translates in real time to English (SIP/RTP) for the other delegate. Neither side needs to understand the other’s language — the interpreter handles everything.

Figure 4.5: ISDN-to-VoIP Conversion — Media Gateway Processing Pipeline

flowchart TD
    A["1. Physical Layer\nNIM receives T1/E1/BRI signal\nand recovers clocking"] --> B["2. ISDN Layer 2 (Q.921/LAPD)\nD-channel data link established"]
    B --> C["3. ISDN Layer 3 (Q.931)\nSETUP message with DNIS and ANI"]
    C --> D["4. Dial-Peer Matching\nRouter matches called number\nto configured dial peers"]
    D --> E["5. VoIP Call Setup\nSIP INVITE sent to\nCUCM or SIP proxy"]
    E --> F["6. Media Transcoding (PVDM4)\nG.711 PCM converted to\nnegotiated VoIP codec"]
    F --> G["7. Call Established\nBidirectional voice:\nISDN PCM <--> RTP/UDP/IP"]

    style A fill:#1a3a5c,stroke:#58a6ff,color:#ffffff
    style B fill:#1a3a5c,stroke:#58a6ff,color:#ffffff
    style C fill:#1a3a5c,stroke:#58a6ff,color:#ffffff
    style D fill:#1a3a5c,stroke:#58a6ff,color:#ffffff
    style E fill:#1a3a5c,stroke:#58a6ff,color:#ffffff
    style F fill:#1a3a5c,stroke:#58a6ff,color:#ffffff
    style G fill:#1a3a5c,stroke:#58a6ff,color:#ffffff

Dial-Peer Configuration for ISDN Voice Ports

Cisco routers use dial peers to route voice calls. A dial peer is a configuration object that matches incoming or outgoing calls based on number patterns and directs them to a specific destination. Two types are relevant:

• POTS dial peer — Points to a physical voice port (the ISDN interface). Handles calls entering or leaving the ISDN side.
• VoIP dial peer — Points to an IP address or DNS name. Handles calls entering or leaving the SIP/H.323 side.

Worked Example — PRI to SIP Gateway Configuration:

! === Define the ISDN switch type (must match the telco) ===
isdn switch-type primary-ni

! === Configure the T1 controller ===
controller T1 0/1/0
 framing esf
 linecode b8zs
 pri-group timeslots 1-24

! === Configure the ISDN serial interface (created by pri-group) ===
interface Serial0/1/0:23
 no ip address
 isdn switch-type primary-ni
 isdn incoming-voice voice
 no shutdown

! === POTS dial peer — matches inbound ISDN calls ===
dial-peer voice 100 pots
 description Inbound-from-PSTN
 incoming called-number .
 direct-inward-dial
 port 0/1/0:23

! === VoIP dial peer — routes calls to CUCM via SIP ===
dial-peer voice 200 voip
 description Outbound-to-CUCM
 destination-pattern 1...
 session protocol sipv2
 session target ipv4:10.1.1.100
 codec g711ulaw

Let’s walk through this configuration:

1. isdn switch-type primary-ni — Tells the router which ISDN variant the telco switch uses. primary-ni is National ISDN (common in North America). This must match the provider’s configuration, or the D-channel will not come up.

2. controller T1 0/1/0 — Configures the physical T1 port on slot 0, subslot 1, port 0. Framing is set to ESF and line coding to B8ZS (standard for T1 in North America).

3. pri-group timeslots 1-24 — Tells the controller to treat all 24 time slots as a PRI group, automatically creating the serial interface Serial0/1/0:23 (the D-channel is always the last time slot).

4. dial-peer voice 100 pots — A POTS dial peer that matches any inbound call on the ISDN port. The incoming called-number . line uses a wildcard (. means “match any digit string”) to accept all incoming calls. The direct-inward-dial keyword passes the called number through for DID routing.

5. dial-peer voice 200 voip — A VoIP dial peer that matches four-digit extensions starting with “1” (destination-pattern 1...). Calls matching this pattern are sent via SIP to the CUCM server at 10.1.1.100 using G.711 u-law codec.

ISDN Switch Types and Regional Significance

The isdn switch-type command is one of the most critical settings in an ISDN deployment. It determines the Q.931 variant the router uses to communicate with the telco switch. A mismatch means the D-channel will never activate, and no calls will flow.

For BRI, the equivalent commands use the basic- prefix:

Tip: When commissioning a new ISDN circuit, always confirm the switch type with the service provider. It is the single most common cause of “ISDN won’t come up” troubleshooting calls.

For BRI NT mode, the additional command isdn protocol-emulate network is required on the interface. This tells the router to act as the network side, providing clocking and Layer 1 activation to connected TE devices:

interface BRI0/2/0
 isdn switch-type basic-net3
 isdn protocol-emulate network
 no shutdown

Trunk Groups for Multi-PRI Management

When a site has multiple PRI spans, configuring individual dial peers for each serial interface becomes unwieldy. Trunk groups simplify this by bundling multiple PRI interfaces into a single logical entity:

trunk group PSTN-TRUNKS
 max-calls voice 46
 
voice-port 0/1/0:23
 trunk-group PSTN-TRUNKS

voice-port 0/1/1:23
 trunk-group PSTN-TRUNKS

dial-peer voice 100 pots
 trunk-group PSTN-TRUNKS
 incoming called-number .
 direct-inward-dial

Now the router treats both PRI spans as a single pool. Inbound calls land on whichever span has available B-channels, and outbound calls automatically hunt across spans for a free channel. This is especially valuable in call center environments where even distribution across trunks prevents any single span from becoming a bottleneck.

Translation Rules for Number Format Conversion

ISDN trunks from the PSTN often deliver called and calling numbers in formats that don’t match the internal dial plan. Translation rules rewrite numbers as they enter or leave the router:

! Strip the leading "1" from 11-digit inbound numbers to get 10-digit format
voice translation-rule 1
 rule 1 /^1\(.*\)/ /\1/

voice translation-profile STRIP-1
 translate called 1

dial-peer voice 100 pots
 translation-profile incoming STRIP-1

This rule converts an inbound called number like 12125551234 to 2125551234 before the router attempts dial-peer matching — ensuring it aligns with the internal 10-digit numbering plan.

Monitoring with show isdn status

The show isdn status command is the first tool to reach for when troubleshooting ISDN connectivity. It displays the state of all three ISDN layers:

Router# show isdn status
Global ISDN Switchtype = primary-ni
ISDN Serial0/1/0:23 interface
        dsl 0, interface ISDN Switchtype = primary-ni
    Layer 1 Status:
        ACTIVE
    Layer 2 Status:
        TEI = 0, Ces = 1, SAPI = 0, State = MULTIPLE_FRAME_ESTABLISHED
    Layer 3 Status:
        0 Active Layer 3 Call(s)
    Active dsl 0 CCBs = 0
    The Free Channel Mask:  0xFFFFFF7F
    Number of L2 Discards = 0, L2 Session ID = 1

Here is how to interpret the output:

The Free Channel Mask is a hexadecimal bitmask showing available B-channels. 0xFFFFFF7F means all 23 B-channels are free (bit 7 cleared = D-channel, not a B-channel). As calls are established, bits flip to 0, indicating occupied channels.

Additional useful verification commands:

Worked Example — Troubleshooting a Down PRI:

A network engineer reports that no calls are going through on a T1-PRI circuit. Here is the diagnostic process:

1. Run show isdn status. Layer 1 shows DEACTIVATED.
2. This indicates a physical problem. Run show controllers T1 0/1/0.
3. The controller output shows AIS (Alarm Indication Signal) — the far end is sending an all-ones pattern, indicating the telco has a problem.
4. Contact the service provider. They confirm a fiber cut between the central office and the customer premises.
5. Once the provider restores the circuit, Layer 1 transitions to ACTIVE, Layer 2 to MULTIPLE_FRAME_ESTABLISHED, and calls begin flowing.

Had Layer 1 been ACTIVE but Layer 2 stuck at TEI_ASSIGNED, the troubleshooting path would shift to verifying the isdn switch-type configuration, since a switch-type mismatch is the most common cause of Layer 2 failures.

Key Takeaway: The Cisco ISR acts as a media gateway, terminating ISDN circuits on one side and speaking SIP on the other. Dial peers are the routing engine that connects these two worlds. Correct switch-type configuration is essential — a mismatch prevents the D-channel from establishing. The show isdn status command provides a quick, layered view of ISDN health and is the starting point for any troubleshooting effort.

Chapter Summary

ISDN brought structured, digital signaling to voice telecommunications through a clean separation of bearer traffic (B-channels) and call control (D-channel using Q.931). This chapter covered the two ISDN service levels and their corresponding Cisco NIM hardware:

• PRI delivers enterprise-grade capacity — 23 B-channels over T1 or 30 B-channels over E1. The NIM-8CE1T1-PRI provides up to 8 channelized ports with integrated CSU/DSU, supporting up to 240 concurrent calls on a single module. It is the standard choice for call centers, hotels, hospitals, and government facilities where reliable PSTN trunking is required.

• BRI serves small offices and specialized applications with 2 B-channels per line. The NIM-2BRI-NT/TE supports both NT and TE modes per port, enabling the router to connect upstream to a telco line (TE) or downstream to ISDN phones (NT). BRI also found a second life as a dial-backup technology for WAN resilience.

• On the router, ISDN-to-VoIP conversion relies on dial peers (POTS for the ISDN side, VoIP for the SIP side), proper switch-type selection to match the provider, and PVDM4 DSPs for real-time codec transcoding. Trunk groups and translation rules simplify management at scale.

• show isdn status is the essential diagnostic command, presenting Layer 1 (physical), Layer 2 (D-channel data link), and Layer 3 (call control) health in a single view.

While SIP trunking continues to grow, ISDN PRI and BRI remain deeply embedded in enterprise networks worldwide. Understanding these interfaces — and the Cisco NIM modules that terminate them — is essential for any engineer managing voice infrastructure that bridges the TDM and IP worlds.

Key Terms

Chapter 5: The VoIP Gateway — CUBE, CUCM, and SIP Trunk Migration from Legacy NIMs

Learning Objectives:

By the end of this chapter, you will be able to:

• Describe the role of a Cisco ISR router as a voice gateway bridging PSTN/ISDN to VoIP networks
• Explain how CUBE handles SIP trunking and protocol interworking
• Describe how CUCM provides call control for NIM-equipped gateways
• Plan a migration strategy from legacy PRI/BRI/analog lines to SIP trunks

5.1 The Voice Gateway Concept

Imagine a border crossing between two countries that speak different languages. On one side, travelers communicate in one language (circuit-switched telephone signals); on the other, they speak another (IP packets). A voice gateway is the translator standing at that border, converting every word — every signal — so that both sides understand each other perfectly.

That is the fundamental job of a voice gateway: bridging the traditional circuit-switched Public Switched Telephone Network (PSTN) to the packet-switched world of Voice over IP (VoIP). Without it, an analog phone plugged into a wall jack and a softphone running on a laptop could never exchange a single word.

5.1.1 What a Voice Gateway Does

A voice gateway performs three essential functions:

1. Signaling conversion — It translates call setup and teardown messages between protocols. An ISDN PRI uses Q.931 signaling; a VoIP network uses SIP (Session Initiation Protocol) or H.323. The gateway converts one to the other so both sides agree on when a call starts, who is ringing, and when a call ends.

2. Media conversion — It digitizes analog voice (or repackages digital TDM voice) into IP packets using codecs such as G.711 or G.729, and vice versa. This includes packetization, jitter buffering, and reassembly.

3. Feature mediation — It maps telephony features (caller ID, call transfer, supplementary services) between the two worlds, handling the mismatches that inevitably arise.

5.1.2 How NIM Cards Provide Physical PSTN Interfaces

On a Cisco ISR 4000-series router acting as a voice gateway, NIM (Network Interface Module) cards provide the physical connection to the PSTN. These are the hardware that plug into the telephone company’s copper or fiber infrastructure:

The router itself handles the IP side — it has Ethernet interfaces connected to the enterprise LAN or WAN. The NIM cards face the PSTN; the Ethernet interfaces face the IP network. The router’s IOS-XE software and DSP resources sit in the middle, performing the conversion.

Figure 5.1: Voice Gateway Architecture — NIM Cards Bridging PSTN to IP

flowchart LR
    subgraph PSTN["PSTN / Carrier Network"]
        PRI["PRI Trunk\n(T1/E1)"]
        BRI["BRI Line\n(2B+D)"]
        ANALOG["Analog Lines\n(POTS)"]
    end

    subgraph ISR["Cisco ISR 4000 Router"]
        subgraph NIMs["NIM Cards (PSTN-Facing)"]
            NIM_T1["NIM-1MFT-T1/E1"]
            NIM_BRI["NIM-2BRI-NT/TE"]
            NIM_FXO["NIM-4FXO / NIM-2FXS"]
        end
        DSP["Onboard DSPs\n(Transcoding, Echo Cancel)"]
        IOSXE["IOS-XE Voice Software\n(Signaling Conversion)"]
        ETH["Ethernet Interfaces\n(IP-Facing)"]
    end

    subgraph IP["Enterprise IP Network"]
        LAN["LAN / WAN"]
        PHONES["IP Phones & Softphones"]
    end

    PRI --> NIM_T1
    BRI --> NIM_BRI
    ANALOG --> NIM_FXO
    NIM_T1 --> DSP
    NIM_BRI --> DSP
    NIM_FXO --> DSP
    DSP --> IOSXE
    IOSXE --> ETH
    ETH --> LAN
    LAN --> PHONES

5.1.3 Voice DSP Resources

Digital Signal Processors (DSPs) are specialized chips that handle the computationally intensive work of voice processing. On NIM voice cards for the ISR 4000 series, DSPs are built directly onto the NIM — no separate PVDM (Packet Voice Data Module) is required. This is a significant simplification over older platforms where you had to purchase and install PVDM cards separately.

DSPs handle:

• Transcoding — Converting between codecs. For example, if the PSTN side uses G.711 (uncompressed, 64 kbps per call) but the WAN link requires G.729 (compressed, 8 kbps per call), the DSP performs the conversion in real time.
• Echo cancellation — Removing the echo that occurs when voice signals bounce back from impedance mismatches in the analog circuit. Without echo cancellation, callers would hear their own voice reflected back with a delay.
• Codec negotiation — Working with the call signaling to select the best codec both sides support, balancing voice quality against bandwidth consumption.

Think of DSPs as specialized translators who not only convert the language but also clean up background noise and adjust the speaking pace to match the listener’s preferred speed.

5.1.4 Gateway vs. Gatekeeper vs. SBC: Role Boundaries

These three devices serve distinct roles in a voice network, and confusing them is a common source of misunderstanding:

A single Cisco ISR can serve as both a voice gateway and an SBC (via CUBE, discussed in the next section). Gatekeepers are largely legacy at this point, as SIP-based architectures have replaced most H.323 deployments.

Key Takeaway: A voice gateway bridges the circuit-switched PSTN and packet-switched IP worlds by converting signaling, media, and features. NIM cards provide the physical PSTN interfaces, onboard DSPs handle transcoding and echo cancellation, and the ISR router’s IOS-XE software orchestrates the entire conversion process.

5.2 CUBE — Cisco Unified Border Element

If the voice gateway is a translator, then CUBE is the diplomatic security detail. CUBE (Cisco Unified Border Element) acts as a Session Border Controller, sitting at the edge of the enterprise voice network and controlling every SIP session that crosses the boundary to an external network — whether that is a SIP trunk to a service provider, a connection to a cloud UC platform, or a link to a business partner’s phone system.

5.2.1 CUBE as SBC for SIP Trunking

CUBE operates as a SIP back-to-back user agent (B2BUA). This means it does not simply pass SIP messages through like a proxy. Instead, it terminates the incoming SIP session on one side and originates a completely new SIP session on the other. This architecture gives CUBE full control over the signaling and media flowing through it.

Figure 5.2: CUBE as a SIP Back-to-Back User Agent (B2BUA)

sequenceDiagram
    participant CUCM as CUCM / Internal Phone
    participant CUBE_IN as CUBE (Inbound Leg)
    participant CUBE_OUT as CUBE (Outbound Leg)
    participant SP as SIP Trunk Provider

    CUCM->>CUBE_IN: SIP INVITE (internal session)
    Note over CUBE_IN: Terminates inbound<br/>SIP session
    CUBE_IN->>CUBE_OUT: Internal routing decision
    Note over CUBE_OUT: Originates new<br/>SIP session
    CUBE_OUT->>SP: SIP INVITE (external session)
    SP-->>CUBE_OUT: 200 OK
    CUBE_OUT-->>CUBE_IN: Maps response
    CUBE_IN-->>CUCM: 200 OK
    Note over CUBE_IN,CUBE_OUT: Media flows through CUBE<br/>(topology hidden, SIP normalized)
    CUCM->>CUBE_IN: RTP Media Stream
    CUBE_OUT->>SP: RTP Media Stream (new source IP)

In practical terms, CUBE:

• Hides internal network topology — External parties see only CUBE’s IP address, never the internal phones or call control systems.
• Normalizes SIP — Different vendors implement SIP with different quirks. CUBE can modify SIP headers, translate between SIP variants, and ensure interoperability.
• Enforces security policies — CUBE can filter calls based on trusted IP lists, apply Access Control Lists (ACLs), use Class of Restriction (COR) to control who can call where, and even change the default SIP signaling port away from 5060 for additional security.
• Controls media flow — CUBE supports both media flow-through (where media passes through the router, enabling transcoding and IP address hiding) and media flow-around (where media goes directly between endpoints, reducing router load).

5.2.2 SIP Trunk Configuration

A SIP trunk is, conceptually, the IP equivalent of a PRI trunk. Where a PRI carries 23 voice channels (T1) or 30 voice channels (E1) over a physical circuit, a SIP trunk carries voice sessions over an IP connection. The number of simultaneous calls is limited by bandwidth and licensing rather than physical channel counts.

Configuring CUBE for SIP trunking involves several key steps. Here is a simplified worked example:

Step 1: Enable CUBE mode on the router.

voice service voip
 mode border-element
 allow-connections sip to sip

The mode border-element command activates CUBE functionality. The allow-connections sip to sip command permits SIP calls to be routed between SIP endpoints — without this, the router would reject SIP-to-SIP calls.

Step 2: Configure a dial-peer for the inbound direction (from CUCM toward CUBE).

dial-peer voice 100 voip
 description Inbound from CUCM
 session protocol sipv2
 session target ipv4:10.1.1.10
 incoming called-number .
 voice-class codec 1
 dtmf-relay rtp-nte

Step 3: Configure a dial-peer for the outbound direction (from CUBE toward the service provider).

dial-peer voice 200 voip
 description Outbound to SIP Provider
 destination-pattern 9T
 session protocol sipv2
 session target ipv4:203.0.113.50
 voice-class codec 1
 dtmf-relay rtp-nte

In this example, calls arriving from the internal CUCM call control (at 10.1.1.10) are matched by dial-peer 100. Outbound calls to the service provider (at 203.0.113.50) are matched by dial-peer 200 based on the destination pattern 9T (any number starting with 9, which is a common access code for external lines).

5.2.3 Protocol Interworking

One of CUBE’s most valuable capabilities is protocol interworking — translating between different voice signaling protocols. CUBE supports:

In a migration scenario, protocol interworking is critical. An organization might still have active PRI trunks on NIM-1MFT-T1/E1 cards while simultaneously running SIP trunks through CUBE. The router handles both, with CUBE converting between the legacy ISDN signaling and the modern SIP signaling as needed.

5.2.4 CUBE Licensing, Capacity Planning, and High Availability

Licensing: CUBE requires a UC (Unified Communications) license package on the ISR. The number of concurrent sessions is controlled by license tiers. Capacity varies by platform — an ISR 4331 might support hundreds of sessions, while an ISR 4461 can scale to thousands.

Capacity planning involves calculating:

• Expected concurrent calls at peak hour
• Bandwidth per call (depends on codec: G.711 at ~87 kbps per call with overhead, G.729 at ~31 kbps)
• DSP resources for transcoding (if different codecs are used on each side)
• CPU headroom for SIP message processing

High availability: On the ISR 4000 series, CUBE supports an active/standby high-availability model using a virtual IP address. Two ISR routers are configured as an HA pair. Under normal conditions, one router handles all calls (active). If it fails, the standby router takes over the virtual IP and begins processing calls. Large deployments can scale up to 64,000 sessions using CUSP (Cisco Unified SIP Proxy) for load distribution across multiple CUBE instances.

Key Takeaway: CUBE acts as a SIP back-to-back user agent at the enterprise voice network boundary. It normalizes SIP, hides topology, enforces security, and performs protocol interworking — making it the recommended SBC for SIP trunk deployments and the key enabler for migrating away from legacy PRI/BRI trunks.

5.3 CUCM — Cisco Unified Communications Manager

If CUBE is the border security, CUCM (Cisco Unified Communications Manager) is the central switchboard operator — the brain that decides where every call goes, which phones ring, and what features are available.

5.3.1 CUCM’s Role as Call Control Platform

CUCM is a software-based call control platform that runs on dedicated servers (or virtual machines). It provides:

• Call routing — Determining the path a call takes based on the dialed number, time of day, caller identity, and available resources
• Phone registration — IP phones, softphones, and gateways register with CUCM, which tracks their status and capabilities
• Feature services — Hold, transfer, conference, voicemail integration, presence, and dozens of other telephony features
• User management — Associating phone lines with users, controlling permissions, and integrating with enterprise directories

Think of CUCM as the air traffic control tower for voice. Every “plane” (call) checks in with the tower, receives routing instructions, and is guided to its destination. Without it, phones would not know how to reach each other, and advanced features would be impossible.

5.3.2 How NIM-Equipped Gateways Register with CUCM

When a Cisco ISR router with NIM voice cards serves as a voice gateway, it must register with CUCM so that CUCM can route calls through it. The gateway can register using several protocols, each with different trade-offs:

MGCP is the primary recommended protocol for CUCM-managed gateways because it gives CUCM complete control over the gateway’s behavior. When an FXO port receives an incoming call, the NIM card signals the router, which in turn signals CUCM via MGCP. CUCM then instructs the gateway on exactly how to handle the call — which codec to use, where to send the media, and what features to apply.

Version requirements matter: CUCM 10.5 or later is required for ISR 44xx series gateways, and CUCM 10.5.2 or later for ISR 43xx series. The ISR must run IOS-XE 3.16 or later with the UC license package.

5.3.3 Route Patterns, Dial Plans, and Call Routing

CUCM uses route patterns to match dialed digits and determine call paths. Here is a worked example of how a call flows through the system:

Scenario: An employee at extension 4001 dials 9-1-555-0199 to reach an external number.

1. The IP phone sends the dialed digits to CUCM.
2. CUCM matches the digits against its route patterns. The pattern 9.1[2-9]XX[2-9]XXXXXX matches (the leading 9 is stripped as an access code, leaving 1-555-0199).
3. The route pattern points to a route list, which contains one or more route groups.
4. The route group contains the voice gateway (the ISR with NIM cards) as a member.
5. CUCM sends routing instructions to the gateway (via MGCP or SIP).
6. The gateway places the call out through the appropriate NIM card — for example, out a PRI trunk on a NIM-1MFT-T1/E1 card to the PSTN.

The call routing hierarchy in CUCM looks like this:

Dialed Digits
  └── Route Pattern (9.1[2-9]XX[2-9]XXXXXX)
        └── Route List (RL-PSTN-Primary)
              ├── Route Group 1 (RG-PRI-Gateway) ← primary path
              └── Route Group 2 (RG-SIP-Trunk)   ← failover path

This layered structure allows for redundancy. If the PRI gateway is unavailable, CUCM automatically fails over to the SIP trunk — a critical capability during a migration from legacy NIMs to SIP trunks.

Figure 5.3: CUCM Call Routing Hierarchy — Route Patterns to Gateways

graph TD
    DIGITS["Dialed Digits\n(9-1-555-0199)"] --> RP["Route Pattern\n9.1[2-9]XX[2-9]XXXXXX"]
    RP --> RL["Route List\nRL-PSTN-Primary"]
    RL --> RG1["Route Group 1\nRG-PRI-Gateway\n(Primary Path)"]
    RL --> RG2["Route Group 2\nRG-SIP-Trunk\n(Failover Path)"]
    RG1 --> GW["ISR Voice Gateway\n(NIM-1MFT-T1/E1)"]
    RG2 --> CUBE["CUBE SBC\n(SIP Trunk to Provider)"]
    GW --> PSTN["PSTN\n(PRI Circuit)"]
    CUBE --> SIP_PROV["SIP Trunk Provider\n(IP Network)"]

    style RG1 fill:#1a5c1a,stroke:#2ea02e,color:#fff
    style RG2 fill:#1a3a5c,stroke:#58a6ff,color:#fff
    style RL fill:#3a2a1a,stroke:#d29922,color:#fff

5.3.4 Integration with Webex Calling and Cloud UC Platforms

Modern CUCM deployments increasingly integrate with cloud UC platforms, particularly Cisco Webex Calling. In these hybrid architectures:

• CUCM continues to manage on-premises phones and gateways
• CUBE provides the SIP trunk connection to the Webex cloud
• NIM-equipped gateways provide local PSTN survivability — if the WAN link to the cloud fails, the local gateway can still route calls through its PRI or analog lines

This hybrid model is often a stepping stone in migration. Organizations can move users to cloud calling in waves while maintaining on-premises gateway resources for resilience and for users not yet migrated.

Key Takeaway: CUCM is the centralized call control brain that manages phone registration, call routing, and feature services. NIM-equipped gateways register with CUCM — typically via MGCP for maximum control — and CUCM directs calls through those gateways to the PSTN. The route pattern/route list/route group hierarchy enables flexible, redundant call routing that supports phased migration.

5.4 Migration: From Legacy NIMs to SIP Trunks

Every technology has a lifecycle, and legacy PRI, BRI, and analog PSTN interfaces are deep into their sunset phase. The question for most organizations is not whether to migrate to SIP trunks, but how to do it safely.

5.4.1 Why Companies Migrate

Three forces drive the migration from legacy NIM-based PSTN connections to SIP trunks:

Cost. A PRI circuit typically costs $300-$800 per month for 23 channels. A SIP trunk providing equivalent capacity often costs 40-60% less, with the ability to pay only for the channels actually used rather than a fixed bundle.

Flexibility. PRI trunks are tied to a physical location — the copper or fiber terminates at a specific building. SIP trunks are location-independent. An organization can consolidate all PSTN connectivity through a central CUBE, even if offices are spread across multiple cities. Adding capacity means purchasing additional SIP sessions, not waiting weeks for the carrier to provision new circuits.

PSTN sunset. Carriers worldwide are decommissioning legacy TDM infrastructure. In many markets, ISDN PRI is no longer available for new installations, and existing circuits face end-of-life dates. Organizations that wait too long may find themselves scrambling when their carrier announces discontinuation.

Think of it like transitioning from a fleet of dedicated delivery trucks (PRI — fixed routes, fixed capacity, high maintenance) to an on-demand delivery service (SIP — flexible, scalable, pay-per-use).

5.4.2 The Five-Phase Migration Approach

A well-planned migration from legacy NIMs to SIP trunks follows five phases:

Total timeline: approximately 3-5 months for a typical mid-sized deployment.

Figure 5.4: Five-Phase Migration from Legacy NIMs to SIP Trunks

flowchart LR
    P1["Phase 1\nAssess & Inventory\n(2-4 weeks)"] --> P2["Phase 2\nPrepare Gateway\n(2-4 weeks)"]
    P2 --> P3["Phase 3\nSet Up SIP Trunks\n(1-2 weeks)"]
    P3 --> P4["Phase 4\nParallel Run\n(4-8 weeks)"]
    P4 --> P5["Phase 5\nCutover &\nDecommission\n(2-4 weeks)"]

    P1 ~~~ N1["Catalog NIMs, lines,\nutilization, special circuits"]
    P2 ~~~ N2["Convert to SIP,\nupdate IOS-XE,\ndeploy CUBE"]
    P3 ~~~ N3["Configure CUBE dial-peers,\nestablish provider trunk"]
    P4 ~~~ N4["Run PRI + SIP side by side,\nvalidate quality"]
    P5 ~~~ N5["Port numbers, cut traffic,\nremove legacy NIM cards"]

    style P1 fill:#1a3a5c,stroke:#58a6ff,color:#fff
    style P2 fill:#1a3a5c,stroke:#58a6ff,color:#fff
    style P3 fill:#1a3a5c,stroke:#58a6ff,color:#fff
    style P4 fill:#1a5c1a,stroke:#2ea02e,color:#fff
    style P5 fill:#5c1a1a,stroke:#da3633,color:#fff

5.4.3 Worked Example: Phased Migration

Scenario: Acme Corp has an ISR 4351 with two NIM-1MFT-T1/E1 cards (two PRIs, 46 voice channels) and one NIM-4FXO (four analog lines for a fax machine and elevator phone). They want to migrate to SIP trunks.

Phase 1 — Assess and Inventory:

Phase 2 — Prepare the Gateway:

• Upgrade IOS-XE to latest recommended release
• Deploy CUBE on the same ISR 4351 (it can serve as both gateway and SBC)
• Convert the gateway registration to SIP (since CUBE requires SIP)
• Verify CUCM route patterns can direct calls to both PRI and SIP trunk paths

Phase 3 — Set Up SIP Trunks:

• Contract with a SIP trunk provider for 35 concurrent sessions (covering peak load with headroom)
• Configure CUBE with inbound and outbound dial-peers to the provider
• Configure CUCM with a SIP trunk to CUBE and add it to route lists as a secondary path

Phase 4 — Parallel Run:

• Route main office numbers (555-0100 range) through SIP trunk
• Keep call center numbers (555-0200 range) on PRI trunk 2
• Monitor call quality, completion rates, and failover behavior for 6 weeks
• Keep PRI trunk 1 active as failover during this period

Phase 5 — Cutover and Decommission:

• Port call center numbers to SIP provider (allow 2-4 weeks for number porting)
• Move all call center traffic to SIP trunk
• Convert fax line to a T.38 fax-over-IP session through CUBE (or retain one FXO port if analog fax is required)
• Address elevator phone — may require retaining one FXO port for regulatory compliance, or replace with a cellular emergency phone
• Cancel PRI circuits with carrier
• Remove NIM-1MFT-T1/E1 cards (or repurpose router for other tasks)

Figure 5.5: Parallel Run Architecture — Legacy PRI and SIP Trunk Coexistence

flowchart TD
    CUCM["CUCM\nCall Control"] --> RL["Route List\n(Primary + Failover)"]

    RL -->|"Main Office\n555-0100 range"| SIP_PATH["SIP Trunk Path"]
    RL -->|"Call Center\n555-0200 range"| PRI_PATH["Legacy PRI Path"]

    subgraph SIP_PATH_DETAIL["SIP Trunk Path (New)"]
        CUBE["CUBE SBC\non ISR 4351"]
        SIPPROV["SIP Trunk\nProvider"]
    end

    subgraph PRI_PATH_DETAIL["Legacy PRI Path (Retained)"]
        NIM["NIM-1MFT-T1/E1\nPRI Trunk 2"]
        CARRIER["PSTN Carrier\nPRI Circuit"]
    end

    SIP_PATH --> CUBE
    CUBE --> SIPPROV
    PRI_PATH --> NIM
    NIM --> CARRIER

    SIPPROV --> PSTN_OUT["PSTN"]
    CARRIER --> PSTN_OUT

    style SIP_PATH fill:#1a5c1a,stroke:#2ea02e,color:#fff
    style PRI_PATH fill:#5c4a1a,stroke:#d29922,color:#fff
    style PSTN_OUT fill:#1a3a5c,stroke:#58a6ff,color:#fff

5.4.4 SIP Trunk as Modern Replacement for PRI

The following comparison highlights why SIP trunks are the natural successor to PRI:

5.4.5 Decommissioning Legacy Cards

Decommissioning is not simply pulling a card out of a slot. A responsible decommission process includes:

1. Verify zero traffic — Confirm that no calls are routing through the legacy NIM cards. Use show voice call summary and show isdn status to verify.
2. Update dial plans — Remove or redirect any CUCM route patterns, route groups, and route lists that reference the legacy gateway.
3. Cancel carrier circuits — Coordinate with the carrier to disconnect PRI/BRI circuits. Ensure number porting is complete before cancellation.
4. Document the change — Update network diagrams, inventory systems, and disaster recovery plans.
5. Physically remove cards — NIM cards support OIR (Online Insertion and Removal), so they can be removed from a running router without a reboot. However, best practice is to perform removal during a maintenance window.
6. Reclaim resources — The DSP resources on the removed NIM are freed. If the ISR is no longer needed as a voice gateway at all, consider redeploying it for other network functions.

Key Takeaway: Migration from legacy NIMs to SIP trunks follows a five-phase approach: assess, prepare, set up SIP, parallel run, and cutover. The parallel-run phase is critical — it validates the new SIP infrastructure while maintaining legacy fallback. CUBE is the recommended SBC for SIP-PSTN interworking, and number porting typically requires 2-4 weeks of lead time.

Chapter Summary

This chapter traced the path from physical PSTN connections through voice gateways to modern SIP trunking:

• Voice gateways bridge the circuit-switched PSTN and packet-switched IP networks. NIM cards provide the physical PSTN interfaces, onboard DSPs handle transcoding and echo cancellation, and the ISR router orchestrates the conversion.

• CUBE operates as a SIP back-to-back user agent at the network boundary, providing topology hiding, SIP normalization, security enforcement, and protocol interworking (SIP, H.323, ISDN, analog). It is the recommended SBC for SIP trunk deployments.

• CUCM serves as the centralized call control platform, managing phone registration, call routing through route patterns and route lists, and integration with both on-premises gateways and cloud platforms like Webex Calling. Gateways register with CUCM via MGCP (primary), SIP, SCCP, or H.323.

• Migration from legacy PRI/BRI/analog NIMs to SIP trunks is driven by cost savings, flexibility, and PSTN sunset timelines. A five-phase approach — assess, prepare, set up SIP trunks, parallel run, cutover — minimizes risk while enabling a controlled transition.

The technologies covered in this chapter represent the bridge between the legacy voice world (covered in earlier chapters) and the modern cloud-based communications platforms that are increasingly dominant. Understanding how they interconnect is essential for anyone managing, troubleshooting, or evolving a Cisco voice infrastructure.

Key Terms