Voice over frame relay had been around for a few years…

In the early 1990s, I worked in the marketing group of the data networking division of Nortel (then called Northern Telecom). We built terrific X.25 packet switching systems that were optimized for large phone company-style implementations (DPN-100) of lots and lots of analog or low speed access circuits connected over a connection-less higher speed (T-1 (1 Mbps) was the fastest WAN link for the longest time) backbone.

In 1993, the company announced the Magellan Passport, a new platform designed for broadband wide area networks. It could support frame relay, ATM and direct digital connection to the PBX, and did high performance cells (proprietary) over the WAN link to another location. Nortel eventually built in great functionality right into the Passport too. For example, instead of just transmitting the voice circuits to another location, it could interpret the call control signaling data (over the D-channel of the T1 PRI circuit) and could direct specific circuits to specific locations where other Passport switches would convert the cells back into voice circuits for the PBX to terminate on the called party's phone. It became a billion dollar business by the 2000 timeframe. Nortel pioneered the use of silence suppression (removing the silence in human conversations – you talk, I listen – and substituting white noise at the listening ends) and generated substantial cost savings for major enterprises.

In 1996, the company acquired Micom, a specialist in low cost and low speed frame relay access devices that could network dozens of smaller offices into enterprise networks. These FRADs were cheaper, lower performance not as intelligent as the Passport. They required a frame relay circuit between each of the offices, and did not do silence suppression, although they did do high density voice compression/decompression.

Micom had a network of sales VARs, which complemented the Passport which was sold direct to major enterprises. Micom specialized in Voice over Frame Relay implementations. At the time of the acquisition, they were just in the process of introducing a Voice over IP product called VIP. This was a DSP-based PC card that would be installed in a PC. Similar to the Micom FRAD, these PCs would connect enterprise locations over an IP network. The enterprise had to configure each to be aware of the other locations of VIP installations… and the PC architecture really didn't deliver strong reliability, but it saved a fortune in LD expenses in the mid 1990s.

The skuttlebut at the time of the acquisition was that Cisco was considering the acquisition of Micom, but since Nortel was already working with Micom to integrate their platform into the Nortel Passport portfolio, it was a concern that Nortel technology would fall into the competitors' hands. Hence, the argument that 'we have to buy them before the other guys do.' [I was involved with the Internet access business unit at that time.]

Cisco, had a dominant position in routers in 1996 (and 1995, and 1994…. and today). Their routers did not support voice interfaces in those days as the application was not well understood as a 'data applicaiton'. You might recall, that before 1995, there were many protocols in the enterprise LAN: AppleTalk, IP, DECnet, ATM, IPX and it was not yet clear how IP would dominate. However it was clear that routers as a product category were going to get voice interfaces, particularly as Nortel had validated the market and the technology.