Overview
Second-generation mobile networks represented a fundamental architectural shift: moving from analogue FM radio to digital transmission. The Global System for Mobile Communications (GSM) standard, developed by ETSI and first commercially deployed by Radiolinja in Finland in 1991, became the dominant global standard. In the United States, the competing IS-95 standard (branded cdmaOne), developed by Qualcomm and using a completely different multiple-access technique, offered an alternative digital path that would influence the 3G era profoundly.
The move to digital brought three immediate benefits: better spectral efficiency (more calls per unit of spectrum), encrypted voice (making eavesdropping with a simple radio scanner impossible), and the ability to carry non-voice data — enabling SMS, the unexpected killer application that transformed how people communicate.
GSM Architecture and Radio Interface
GSM divides spectrum into 200 kHz channels and then further subdivides each channel in time using Time Division Multiple Access (TDMA), creating eight time slots per channel frame. Each active voice call is assigned one time slot, receiving radio access eight times per frame (every 4.615 ms). The voice codec used is the Regular Pulse Excitation – Long Term Prediction (RPE-LTP) algorithm, operating at 13 kbps, which reduces the raw audio to a digital bitstream that can be protected with forward error correction before transmission.
The radio modulation scheme is Gaussian Minimum Shift Keying (GMSK), a continuous-phase modulation chosen for its spectral efficiency and constant envelope properties — important for battery life, as a constant-envelope signal allows a more efficient power amplifier. GSM operates in several frequency bands: the original 900 MHz and 1800 MHz bands in Europe, and 850 MHz / 1900 MHz in North America.
The SIM (Subscriber Identity Module) card — a standardised removable smart card carrying subscriber authentication credentials — was a GSM invention. The concept of separating subscriber identity from the handset itself enabled the modern mobile market as we know it.
IS-95 and CDMA
The IS-95 standard took a fundamentally different approach to multiple access. Rather than dividing spectrum by frequency or time, Code Division Multiple Access (CDMA) allows all users to transmit simultaneously across the entire 1.25 MHz channel, distinguishing each user's signal through the use of orthogonal spreading codes. Each user's data is multiplied by a unique pseudo-random code sequence running at a much higher chip rate (1.2288 Mcps), spreading the signal across the bandwidth. The receiver uses the same code to despreading and recover the original data.
CDMA offered several theoretical advantages: soft handoff (a mobile could communicate with multiple base stations simultaneously, improving coverage at cell edges), inherent resistance to multipath interference via rake receivers, and variable-rate speech coding that reduced transmitted power during silence. IS-95 became the basis for the CDMA2000 3G family.
2.5G: GPRS and EDGE
The 2.5G evolution layer added packet-switched data capability to the GSM infrastructure without requiring new spectrum. General Packet Radio Service (GPRS), standardised in 1997 and deployed from 2000, allowed multiple time slots to be aggregated for a single data connection, achieving theoretical peaks of 171.2 kbps (using all eight slots) though practical speeds were typically 40–80 kbps. Crucially, GPRS introduced an always-on paradigm: the mobile maintained a packet data connection even when no data was being transferred, paying only for data volume rather than connection time.
Enhanced Data rates for GSM Evolution (EDGE), sometimes called 2.75G, upgraded the radio modulation for data channels from GMSK to 8-PSK (8-Phase Shift Keying), tripling the number of bits per symbol. Combined with improved coding schemes, EDGE achieved theoretical peaks of 384 kbps and consistent real-world speeds of 100–200 kbps — sufficient for basic web browsing and email. EDGE also served as a 3G fallback standard for operators unable to deploy WCDMA infrastructure.
Security Architecture
2G introduced proper cryptographic security to mobile communications. GSM authentication uses a challenge-response protocol based on the A3 algorithm and a 128-bit secret key (Ki) stored on the SIM. The network sends a 128-bit random number (RAND); the SIM computes a signed response (SRES) using Ki and RAND; if the network's expected SRES matches, authentication succeeds. The session encryption key (Kc) is derived from the same process using the A8 algorithm, and voice/data is encrypted over the air using the A5 stream cipher.
- A5/1: The strong cipher used in Europe and North America. 64-bit key, later shown to be vulnerable to precomputed rainbow table attacks.
- A5/2: A deliberately weakened export cipher, officially deprecated following demonstrated cryptanalysis.
- A5/3 (KASUMI): A later, stronger cipher based on the MISTY1 algorithm, introduced as part of the 3GPP security upgrade.
- Notable weakness: Authentication was one-way — the network authenticated the handset, but the handset did not authenticate the network. This enabled IMSI-catcher (fake base station) attacks.
SMS: The Unexpected Revolution
Short Message Service (SMS) was designed as a minor feature of the GSM specification, originally intended for network operators to send status messages to handsets. The first person-to-person SMS was sent in December 1992. Carrier billing models and the social appeal of asynchronous text communication drove explosive adoption through the late 1990s, and by the mid-2000s SMS had become the world's most widely used data application — generating more revenue than voice calls in many markets. Its influence on the design of instant messaging, social media, and the general brevity of digital communication is incalculable.