Network Voice Protocol

Network Voice Protocol. It was an early experimental protocol developed for transmitting digitized speech over packet-switched networks, representing a pioneering effort in voice over IP technology. Created by researchers working on the ARPANET, it demonstrated the feasibility of real-time audio communication across a distributed network. This work laid crucial groundwork for later Internet telephony and modern VoIP systems, influencing the development of subsequent audio transmission standards.

Overview

The protocol was designed to facilitate bidirectional voice conversations between hosts on the ARPANET, treating audio as a continuous stream of digital packets. Its development was led by researchers like Danny Cohen at the University of Southern California's Information Sciences Institute, who was instrumental in early packet voice experiments. The work was part of a broader DARPA initiative exploring real-time applications, connecting sites such as the MIT Lincoln Laboratory and Stanford Research Institute. It proved that a shared, best-effort network could support latency-sensitive traffic, a concept vital for future multimedia applications.

Technical details

The system operated by digitizing audio input using pulse-code modulation or continuously variable slope delta modulation, then segmenting the data into packets for transmission. It utilized the underlying NCP and later TCP protocols for transport, incorporating sequence numbers and timestamps to manage packet arrival and playback. A significant challenge was compensating for variable network delay and packet loss, which researchers addressed through buffering techniques. The protocol also defined control messages for call setup and management, establishing a basic framework for session control that informed later work on protocols like SIP.

Development and history

Initial experiments began in the early 1970s under the sponsorship of the DARPA, with the first successful conversation occurring in 1973 between USC and the MIT Lincoln Laboratory. Key figures included Danny Cohen, Steve Casner, and Vint Cerf, who explored the integration of real-time voice with the existing ARPANET infrastructure. These activities were closely related to parallel projects like the Packet Radio Network and SATNET, which examined voice transmission in other challenging environments. The work was formally documented in RFC 741, published in 1977, which detailed the protocol's specifications and operational experience.

Implementation and usage

Practical implementations ran on specialized hardware, including PDP-11 minicomputers equipped with custom codec and digital signal processing boards. The software was integrated with early TCP/IP stacks, allowing voice sessions between major ARPANET nodes. Primary users were the research teams at institutions like the USC Information Sciences Institute, Stanford Research Institute, and BBN Technologies. Usage remained confined to experimental demonstrations and technical validations, as the computational and bandwidth requirements were substantial for the era. These tests provided critical data on network performance and quality of service for real-time data.

Legacy and influence

The project directly inspired subsequent DARPA research into packet voice, which evolved into the Network Voice Protocol's successor, the Internet Stream Protocol. Its concepts influenced the design of later commercial voice over IP systems and foundational standards for multimedia networking developed by the IETF. The experimental work on managing latency and jitter informed the development of protocols such as the RTP and the RSVP. This early proof-of-concept is historically recognized as a seminal step toward integrating real-time communication into the Internet Protocol Suite, paving the way for applications like Skype and Zoom.

Category:Network protocols Category:Voice over IP Category:ARPANET Category:History of the Internet