Skip to content

Instant Messaging: From ICQ to Signal

Abstract

Instant messaging turned the network into a place where people are present to one another, where a list of contacts shows who is online right now and a message arrives in seconds, not the asynchronous limbo of email. From the buddy-list era of ICQ and AIM in the late 1990s, through the multi-protocol chaos of MSN and Yahoo, to the mobile reinvention by WhatsApp and WeChat, and finally the encryption era of Signal and iMessage, instant messaging became the default mode of personal communication for billions of people, quietly overtaking SMS, email, and the phone call. Its history is a story of presence and identity, of network effects and walled gardens, and of a long migration from open, interoperable beginnings toward closed platforms, and then a partial return toward open standards and end-to-end encryption. This article complements the entries on IRC (group chat) and email (asynchronous mail), focusing on real-time personal messaging.

Roots: Talk, Chat, and the Buddy List

Real-time text predates the public internet. On 1970s time-sharing systems and PLATO, users could send messages to others logged into the same machine; PLATO’s term-talk and Talkomatic gave students direct and group chat by 1973. Unix’s talk command (and the earlier write) let two logged-in users type at each other in split-screen. IRC (1988) brought multi-user channels to the internet. But these required users to be on the same system or to know server and channel names; there was no notion of a personal, persistent list of your contacts and whether they were available.

The missing concept was presence: a real-time indicator of who in your social circle is online and reachable. Presence plus a personal buddy list is what defined instant messaging as a distinct medium.

ICQ and AIM: The Buddy-List Era

In November 1996, the Israeli company Mirabilis (founded that June by Yair Goldfinger, Sefi Vigiser, Amnon Amir, Arik Vardi and his father Yossi Vardi) released ICQ (“I Seek You”). ICQ introduced to a mass audience the features that would define the genre: a central server that tracked which of your contacts were online, a buddy list, offline messaging, and the iconic “uh-oh!” notification sound. Each user had a numeric UIN, issued in sequence starting at 100,000, so a low ICQ number became a status symbol. ICQ spread virally; AOL acquired Mirabilis on June 8, 1998, for $287 million up front plus up to $120 million in performance payments over three years.

In parallel, AOL launched AIM (AOL Instant Messenger) in May 1997, breaking it out of the walled AOL service so anyone on the internet could use it. AIM dominated North America for a decade before its shutdown on December 15, 2017, with its away messages and screen-name culture becoming a defining feature of millennial adolescence. Microsoft countered with MSN Messenger (1999, later Windows Live Messenger), which dominated much of Europe and Latin America, and Yahoo! Messenger competed as well.

This created a fragmented landscape of incompatible networks. Users often ran several clients at once, which gave rise to multi-protocol clients (Trillian, and the open-source Pidgin/Gaim and Adium) that logged into AIM, MSN, Yahoo, and ICQ simultaneously. AOL repeatedly tried to block third-party and Microsoft clients from connecting to AIM, an early skirmish in the long war between interoperability and walled gardens.

The Open Standard That Almost Won: Jabber/XMPP

Against the proprietary networks, an open alternative emerged. Jeremie Miller began work on Jabber in 1998 and released the first jabberd server on January 4, 1999: an open, decentralized, XML-based instant-messaging protocol where anyone could run a server and servers could federate, much like email. The IETF standardized it as XMPP (Extensible Messaging and Presence Protocol) in RFCs 3920 and 3921 in October 2004.

For a moment, the open model seemed ascendant. Google Talk launched on XMPP in 2005 and federated with other XMPP servers; for a time you could chat across providers. But the major platforms ultimately closed their federation: Google dropped server-to-server XMPP federation in May 2013, Facebook shut the XMPP gateway it had opened in 2010 in April 2014, and the industry consolidated around proprietary, non-interoperable services once more. XMPP survives as infrastructure (WhatsApp built its transport on a customized variant called FunXMPP, and the protocol is widely used inside corporate and gaming systems) but the dream of email-like universal IM federation faded. (The torch passed partly to Matrix, a newer open federated protocol, in the 2010s.)

The Mobile Reinvention: WhatsApp, WeChat, and the SMS Killers

The smartphone changed everything. SMS text messaging was ubiquitous but expensive, especially for international and group messages, and capped at 160 characters. The carriers’ own attempt at a richer successor (MMS, picture messages billed per item at several times the SMS price) had already failed to displace it. On February 24, 2009, Jan Koum and Brian Acton (former Yahoo engineers) incorporated WhatsApp, which sent messages over the phone’s data connection for free, used the phone number itself as the identity (no usernames, no buddy lists to build, your address book was your contact list), and worked across iPhone and Android. This combination was explosive. WhatsApp displaced paid SMS in much of the world and reached hundreds of millions of users; Facebook acquired it on February 19, 2014, for about $19 billion, its largest acquisition. WhatsApp passed 2 billion active users in February 2020.

In China, WeChat (Weixin), started as a Tencent Guangzhou project in October 2010 under Allen Zhang and released by Tencent in 2011, evolved beyond messaging into a “super-app” (WeChat Pay from 2013, mini-programs from 2017, government services, and social feeds), passing a billion monthly active users in 2018 and becoming inseparable from daily life and a model the rest of the industry has chased since. Apple’s iMessage (October 2011, with iOS 5) quietly hijacked the SMS habit on iPhones: messages between Apple devices became encrypted internet messages (the “blue bubbles”) while falling back to green-bubble SMS for others, a design that produced both seamless Apple-to-Apple messaging and a now-famous social-status divide.

The Encryption Era

As messaging carried ever more of human communication, end-to-end encryption (E2EE) (where only the sender and recipient can read messages, not even the service provider) moved from a fringe demand to a baseline expectation, accelerated by the 2013 Snowden revelations of mass surveillance.

The enabling technology was the Signal Protocol, developed by Moxie Marlinspike and Trevor Perrin at Open Whisper Systems out of the earlier TextSecure protocol. It combines a Double Ratchet algorithm with prekeys and an Extended Triple Diffie-Hellman handshake, giving forward secrecy and continuous key renewal for messages sent while the recipient is offline. The Signal app itself is the privacy advocate’s choice; since February 2018 it has been run by the nonprofit Signal Foundation, seeded with $50 million from WhatsApp co-founder Brian Acton. The protocol’s reach went far beyond Signal: WhatsApp finished rolling it out across all message types on April 5, 2016, bringing end-to-end encryption to a billion users at once, and it underpins encryption in Google Messages (RCS) and Facebook Messenger’s secret conversations.

Meanwhile RCS (Rich Communication Services), a carrier-backed successor to SMS, slowly rolled out as a richer default texting standard; on September 16, 2024 Apple added RCS support to the iPhone with iOS 18, narrowing (though not closing) the green-bubble divide.

Dead End: Google’s Messaging Graveyard

No company illustrates the failure modes of instant messaging better than Google, which launched and killed an extraordinary parade of messaging products without ever building a durable rival to WhatsApp or iMessage: Google Talk (2005, on open XMPP), Google Wave (announced May 2009, an ambitious and confusing collaboration tool, development halted August 2010), Google Buzz (2010, a privacy fiasco), Hangouts (2013), Allo and Duo (2016), Google Chat, and a long-running, fragmented push into RCS via Google Messages. The pattern, overlapping products, abrupt cancellations, abandoned open standards (Google’s 2013 retreat from XMPP federation broke cross-provider chat for many), and constant rebranding, became a running industry joke and a case study in how organizational incoherence, not technology, can squander a platform opportunity.

The deeper lesson echoes the whole history of the medium: messaging is governed by network effects, and network effects punish fragmentation. The winners (WhatsApp, WeChat, iMessage) each picked one identity model and one network and stuck with it; the losers fragmented their own users across incompatible apps. The open-federation alternative (XMPP, and later Matrix) kept the better architecture but never overcame the head start and lock-in of the closed platforms, leaving regulators (notably the EU’s Digital Markets Act, which began pressing large messengers toward interoperability in the mid-2020s) to try to force open what the market had closed.

📚 Sources