The Mobile Phone ecosystem

Any sufficiently advanced technology is indistinguishable from magic - Arthur C Clarke

Technology has advanced so rapidly in the past few decades that the basic workings of something as simple as a telephone call are incomprehensible to 99% of the modern world. This is a layman's attempt to map that process in a post-copper world.

To the Tower!

We begin with the pocket computer that almost everyone over the age of 15 carries daily here in the United States. If you were to compare an iPhone 16 to a high-end PC from the 1980s or early 90s – say, a 386DX you'd find that the device in your pocket is 3000 times faster at raw computation alone. This is before we even mention data transmission with the blazing-fast speeds of 14.4kbps over a modem and hardwired copper phone line – versus modern cellular technology which reaches over 5 Gbps, or around 350,000 times faster.

This pocket supercomputer uses a built-in radio to communicate with cell towers throughout the US between 600 megahertz and 6 gigahertz and something called Orthogonal Frequency Division Multiplexing, or OFDM.

OFDM works by using multiple frequencies simultaneously to carry as much data as possible to these radio towers previously mentioned. The frequencies are spaced on the radio spectrum so that the transmissions don't interfere with each other when transmitting.

The other part of this wireless magic is Quadrature Amplitude Modulation (QAM). Carriers use different types of QAM but they all essentially work the same, they adjust the amplitude(signal strength) and phase(timing) of the radio wave to get more data into each transmission. Without going too in depth or bullying you into studying for your Extra Amateur Radio License, just think of this as sending a stronger signal and tweaking the time at which data is affixed to the radio wave.

What about phone Calls now?

It's possible I've already lost you, and you're trying to figure out what any of this has to do with making a phone call. But this next part is key. Whereas in traditional copper landlines – and even early cellular calls – your actual voice was carried over a call using continuous analog sound waves. Today your voice is turned into binary data and transmitted using TCP/IP. This transformation of an analog voice to ones and zeros happens in your phone before it ever hits that built-in radio we talked about.

This transition from analog to fully digital is important for understanding what is being transmitted to the tower. First and foremost the normal TCP/IP handshake needs to happen. Cellphones are assigned three key things that allow them to communicate with a tower and start transmitting data: an ipv6 address, a SIM card (or eSIM) and an IMEI (International Mobile Equipment Identity). These are absolutely essential before any of that communication can take place – whether placing a phone call or streaming a YouTube tutorial explaining how to repair a chainsaw or leaking sink.

The IPv6 address is easy, it's assigned within the framework of the IEEE using a variety of techniques that vary amongst the carriers. Two of the most common methods are DHCPv6 (Dynamic Host Configuration Protocol version 6) or SLAAC (Stateless Address Auto-configuration). This IPv6 address allows the TCP/IP traffic to begin to flow between the carrier and the phone. From here the IMEI and SIM we talked about are used to validate the phone and the current subscription/access. Unlike an IP address which can change every time the phone connects to a tower, the IMEI is tied to the device from the factory and cannot be changed without some serious wizardry.

So the phone searches for the nearest or strongest broadcast signal from the towers that it can communicate, and then sends traffic. The phone is now communicating with the tower via radio waves. This all happens on a Baseband Unit, eNodeB (4G LTE) or gNodeB(5G) which reserves frequencies based on a Radio Resource control request (RRC). This allows the phone onto the Radio Access Network (RAN). After the connection between phone and tower is locked the IMEI and SIM information is sent on the carriers core network and checked for validity. Once authenticated the phone will transmit and receive TCP/IP traffic using the same frequency and timing from the above RRC process.

Backhaul

Now that we have a stream of data going from the phone to the tower and onto the core network, let’s examine how that traffic flows. The node (eNodeB, gNodeB, etc.) needs to transport data from the tower back to a Mobile Telephone Switching Office (MTSO).

In the majority of cases, this data moves from the node/tower via a fiber-optic connection provided by a local Internet Service Provider (ISP). In this context, we will refer to these local fiber providers as Type 2 carriers, as the cellular carrier is the customer leasing the backhaul infrastructure.

This first hop from the tower to the backbone is where many of the speed issues in 5G begin. If the tower your phone is connected to has a Network Interface Device (NID) running on a 1 Gbps SFP (Small Form-factor Pluggable) module, then all customers connected to that tower share that same 1 Gbps link—introducing congestion and bottlenecks.

As your data moves beyond the tower, it is wrapped with VLAN tagging (802.1Q). Virtual LANs (VLANs) are added to each packet as it enters the Type 2 network, allowing it to be routed from the tower across the backhaul network to the MTSO.

At the MTSO, these VLAN tags are removed (popped), and the traffic is then fed into the carrier’s core network using a different internal VLAN assigned by the carrier.

Another option for transporting traffic from the tower to the MTSO is Multiprotocol Label Switching (MPLS), a technology commonly used in large Wide-Area Networks (WANs) and ISPs. Unlike VLAN tagging, MPLS assigns a label added to the packet instead of a VLAN tag. These MPLS labels allow for more efficient routing across the backhaul network.

Whereas VLAN tags operate at Layer 2 (Data Link Layer) and are added to the Ethernet frame, MPLS functions between the Network (Layer 3) and Transport (Layer 4) layers in the OSI model. This positioning enables more efficient traffic engineering and quality of service (QoS) compared to VLAN-based routing.

One Point of Clarification – Voice traffic is transmitted using the User Datagram Protocol (UDP). While I’ve referenced TCP when discussing TCP/IP as a whole, voice traffic itself is connectionless and relies on Quality of Service (QoS) to prioritize voice traffic over data.

While routes and connections (such as session setup) may involve TCP, Streaming video/ data traffic are typically transmitted using TCP, whereas voice and FaceTime type traffic is exclusively transmitted using UDP at Layer 4.