Voice over IP


Voice over Internet Protocol, also called IP telephony, is a method and group of technologies for the delivery of voice communications and multimedia sessions over Internet Protocol networks, such as the Internet. The terms Internet telephony, broadband telephony, and broadband phone service specifically refer to the provisioning of communications services over the public Internet, rather than via the public switched telephone network, also known as plain old telephone service.

Overview

The steps and principles involved in originating VoIP telephone calls are similar to traditional digital telephony and involve signaling, channel setup, digitization of the analog voice signals, and encoding. Instead of being transmitted over a circuit-switched network, the digital information is packetized and transmission occurs as IP packets over a packet-switched network. They transport media streams using special media delivery protocols that encode audio and video with audio codecs and video codecs. Various codecs exist that optimize the media stream based on application requirements and network bandwidth; some implementations rely on narrowband and compressed speech, while others support high-fidelity stereo codecs.
The most widely speech coding standards in VoIP are based on the linear predictive coding and modified discrete cosine transform compression methods. Popular codecs include the MDCT-based AAC-LD, the LPC/MDCT-based Opus, the LPC-based SILK, μ-law and A-law versions of G.711, G.722, and an open source voice codec known as iLBC, a codec that uses only 8 kbit/s each way called G.729.
Early providers of voice-over-IP services used business models and offered technical solutions that mirrored the architecture of the legacy telephone network. Second-generation providers, such as Skype, built closed networks for private user bases, offering the benefit of free calls and convenience while potentially charging for access to other communication networks, such as the PSTN. This limited the freedom of users to mix-and-match third-party hardware and software. Third-generation providers, such as Google Talk, adopted the concept of federated VoIP. These solutions typically allow dynamic interconnection between users in any two domains of the Internet, when a user wishes to place a call.
In addition to VoIP phones, VoIP is also available on many personal computers and other Internet access devices. Calls and SMS text messages may be sent via Wi-Fi or the carrier's mobile data network. VoIP provides a framework for consolidation of all modern communications technologies using a single unified communications system.

Pronunciation

VoIP is variously pronounced as an initialism, V-O-I-P, or as an acronym /ˈvɔɪp/, as in voice. Full words, voice over Internet Protocol, or voice over IP, are sometimes used.

Protocols

Voice over IP has been implemented with proprietary protocols and protocols based on open standards in applications such as VoIP phones, mobile applications, and web-based communications.
A variety of functions are needed to implement VoIP communication. Some protocols perform multiple functions, while others perform only a few and must be used in concert. These functions include:
VoIP protocols include:

Consumer market

Mass-market VoIP services use existing broadband Internet access, by which subscribers place and receive telephone calls in much the same manner as they would via the public switched telephone network. Full-service VoIP phone companies provide inbound and outbound service with direct inbound dialing. Many offer unlimited domestic calling and sometimes international calls for a flat monthly subscription fee. Phone calls between subscribers of the same provider are usually free when flat-fee service is not available.
A VoIP phone is necessary to connect to a VoIP service provider. This can be implemented in several ways:
It is increasingly common for telecommunications providers to use VoIP telephony over dedicated and public IP networks as a backhaul to connect switching centers and to interconnect with other telephony network providers; this is often referred to as IP backhaul.
Smartphones may have SIP clients built into the firmware or available as an application download.

Corporate use

Because of the bandwidth efficiency and low costs that VoIP technology can provide, businesses are migrating from traditional copper-wire telephone systems to VoIP systems to reduce their monthly phone costs. In 2008, 80% of all new Private branch exchange lines installed internationally were VoIP. For example, in the United States, the Social Security Administration is converting its field offices of 63,000 workers from traditional phone installations to a VoIP infrastructure carried over its existing data network.
VoIP allows both voice and data communications to be run over a single network, which can significantly reduce infrastructure costs. The prices of extensions on VoIP are lower than for PBX and key systems. VoIP switches may run on commodity hardware, such as personal computers. Rather than closed architectures, these devices rely on standard interfaces. VoIP devices have simple, intuitive user interfaces, so users can often make simple system configuration changes. Dual-mode phones enable users to continue their conversations as they move between an outside cellular service and an internal Wi-Fi network, so that it is no longer necessary to carry both a desktop phone and a cell phone. Maintenance becomes simpler as there are fewer devices to oversee.
VoIP solutions aimed at businesses have evolved into unified communications services that treat all communications—phone calls, faxes, voice mail, e-mail, web conferences, and more—as discrete units that can all be delivered via any means and to any handset, including cellphones. Two kinds of service providers are operating in this space: one set is focused on VoIP for medium to large enterprises, while another is targeting the small-to-medium business market.
Skype, which originally marketed itself as a service among friends, has begun to cater to businesses, providing free-of-charge connections between any users on the Skype network and connecting to and from ordinary PSTN telephones for a charge.

Quality of service

Communication on the IP network is perceived as less reliable in contrast to the circuit-switched public telephone network because it does not provide a network-based mechanism to ensure that data packets are not lost, and are delivered in sequential order. It is a best-effort network without fundamental Quality of Service guarantees. Voice, and all other data, travels in packets over IP networks with fixed maximum capacity. This system may be more prone to data loss in the presence of congestion than traditional circuit switched systems; a circuit switched system of insufficient capacity will refuse new connections while carrying the remainder without impairment, while the quality of real-time data such as telephone conversations on packet-switched networks degrades dramatically. Therefore, VoIP implementations may face problems with latency, packet loss, and jitter.
By default, network routers handle traffic on a first-come, first-served basis. Fixed delays cannot be controlled as they are caused by the physical distance the packets travel. They are especially problematic when satellite circuits are involved because of the long distance to a geostationary satellite and back; delays of 400–600 ms are typical. Latency can be minimized by marking voice packets as being delay-sensitive with QoS methods such as DiffServ.
Network routers on high volume traffic links may introduce latency that exceeds permissible thresholds for VoIP. Excessive load on a link can cause congestion and associated queueing delays and packet loss. This signals a transport protocol like TCP to reduce its transmission rate to alleviate the congestion. But VoIP usually uses UDP not TCP because recovering from congestion through retransmission usually entails too much latency. So QoS mechanisms can avoid the undesirable loss of VoIP packets by immediately transmitting them ahead of any queued bulk traffic on the same link, even when the link is congested by bulk traffic.
VoIP endpoints usually have to wait for completion of transmission of previous packets before new data may be sent. Although it is possible to preempt a less important packet in mid-transmission, this is not commonly done, especially on high-speed links where transmission times are short even for maximum-sized packets. An alternative to preemption on slower links, such as dialup and digital subscriber line, is to reduce the maximum transmission time by reducing the maximum transmission unit. But since every packet must contain protocol headers, this increases relative header overhead on every link traversed.
The receiver must resequence IP packets that arrive out of order and recover gracefully when packets arrive too late or not at all. Packet delay variation results from changes in queuing delay along a given network path due to competition from other users for the same transmission links. VoIP receivers accommodate this variation by storing incoming packets briefly in a playout buffer, deliberately increasing latency to improve the chance that each packet will be on hand when it is time for the voice engine to play it. The added delay is thus a compromise between excessive latency and excessive dropout, i.e. momentary audio interruptions.
Although jitter is a random variable, it is the sum of several other random variables which are at least somewhat independent: the individual queuing delays of the routers along the Internet path in question. Motivated by the central limit theorem, jitter can be modeled as a gaussian random variable. This suggests continually estimating the mean delay and its standard deviation and setting the playout delay so that only packets delayed more than several standard deviations above the mean will arrive too late to be useful. In practice, the variance in latency of many Internet paths is dominated by a small number of relatively slow and congested bottleneck links. Most Internet backbone links are now so fast that their delays are dominated by the transmission medium and the routers driving them do not have enough buffering for queuing delays to be significant.
A number of protocols have been defined to support the reporting of quality of service and quality of experience for VoIP calls. These include RTP Control Protocol extended reports, SIP RTCP summary reports, H.460.9 Annex B, H.248.30 and MGCP extensions.
The RTCP extended report VoIP metrics block specified by is generated by an IP phone or gateway during a live call and contains information on packet loss rate, packet discard rate, packet loss/discard burst metrics, network delay, end system delay, signal/noise/echo level, mean opinion scores and R factors and configuration information related to the jitter buffer. VoIP metrics reports are exchanged between IP endpoints on an occasional basis during a call, and an end of call message sent via SIP RTCP summary report or one of the other signaling protocol extensions. VoIP metrics reports are intended to support real-time feedback related to QoS problems, the exchange of information between the endpoints for improved call quality calculation and a variety of other applications.

DSL and ATM

DSL modems typically provide Ethernet connections to local equipment, but inside they may actually be Asynchronous Transfer Mode modems. They use ATM Adaptation Layer 5 to segment each Ethernet packet into a series of 53-byte ATM cells for transmission, reassembling them back into Ethernet frames at the receiving end.
Using a separate virtual circuit identifier for audio over IP has the potential to reduce latency on shared connections. ATM's potential for latency reduction is greatest on slow links, because worst-case latency decreases with increasing link speed. A full-size Ethernet frame takes 94 ms to transmit at 128 kbit/s but only 8 ms at 1.5 Mbit/s. If this is the bottleneck link, this latency is probably small enough to ensure good VoIP performance without MTU reductions or multiple ATM VCs. The latest generations of DSL, VDSL and VDSL2, carry Ethernet without intermediate ATM/AAL5 layers, and they generally support IEEE 802.1p priority tagging so that VoIP can be queued ahead of less time-critical traffic.
ATM has substantial header overhead: 5/53 = 9.4%, roughly twice the total header overhead of a 1500 byte Ethernet frame. This "ATM tax" is incurred by every DSL user whether or not they take advantage of multiple virtual circuits – and few can.

Layer 2

Several protocols are used in the data link layer and physical layer for quality-of-service mechanisms that help VoIP applications work well even in the presence of network congestion. Some examples include:
The quality of voice transmission is characterized by several metrics that may be monitored by network elements and by the user agent hardware or software. Such metrics include network packet loss, packet jitter, packet latency, post-dial delay, and echo. The metrics are determined by VoIP performance testing and monitoring.

PSTN integration

A VoIP media gateway controller works in cooperation with a media gateway and connects the digital media stream, so as to complete creating the path for voice as well as data media. They include the interfaces for connecting the standard PSTN networks with the ATM and Inter Protocol networks. The Ethernet interfaces are also included in the modern systems, which are specially designed to link calls that are passed via the VoIP.
E.164 is a global FGFnumbering standard for both the PSTN and PLMN. Most VoIP implementations support E.164 to allow calls to be routed to and from VoIP subscribers and the PSTN/PLMN. VoIP implementations can also allow other identification techniques to be used. For example, Skype allows subscribers to choose "Skype names" whereas SIP implementations can use URIs similar to email addresses. Often VoIP implementations employ methods of translating non-E.164 identifiers to E.164 numbers and vice versa, such as the Skype-In service provided by Skype and the ENUM service in IMS and SIP.
Echo can also be an issue for PSTN integration. Common causes of echo include impedance mismatches in analog circuitry and acoustic coupling of the transmit and receive signal at the receiving end.

Number portability

and mobile number portability also impact VoIP business. In November 2007, the Federal Communications Commission in the United States released an order extending number portability obligations to interconnected VoIP providers and carriers that support VoIP providers. Number portability is a service that allows a subscriber to select a new telephone carrier without requiring a new number to be issued. Typically, it is the responsibility of the former carrier to "map" the old number to the undisclosed number assigned by the new carrier. This is achieved by maintaining a database of numbers. A dialed number is initially received by the original carrier and quickly rerouted to the new carrier. Multiple porting references must be maintained even if the subscriber returns to the original carrier. The FCC mandates carrier compliance with these consumer-protection stipulations.
A voice call originating in the VoIP environment also faces challenges to reach its destination if the number is routed to a mobile phone number on a traditional mobile carrier. VoIP has been identified in the past as a Least Cost Routing system, which is based on checking the destination of each telephone call as it is made, and then sending the call via the network that will cost the customer the least. This rating is subject to some debate given the complexity of call routing created by number portability. With GSM number portability now in place, LCR providers can no longer rely on using the network root prefix to determine how to route a call. Instead, they must now determine the actual network of every number before routing the call.
Therefore, VoIP solutions also need to handle MNP when routing a voice call. In countries without a central database, like the UK, it might be necessary to query the GSM network about which home network a mobile phone number belongs to. As the popularity of VoIP increases in the enterprise markets because of least cost routing options, it needs to provide a certain level of reliability when handling calls.

Emergency calls

A telephone connected to a land line has a direct relationship between a telephone number and a physical location, which is maintained by the telephone company and available to emergency responders via the national emergency response service centers in form of emergency subscriber lists. When an emergency call is received by a center the location is automatically determined from its databases and displayed on the operator console.
In IP telephony, no such direct link between location and communications end point exists. Even a provider having hardware infrastructure, such as a DSL provider, may know only the approximate location of the device, based on the IP address allocated to the network router and the known service address. Some ISPs do not track the automatic assignment of IP addresses to customer equipment.
IP communication provides for device mobility. For example, a residential broadband connection may be used as a link to a virtual private network of a corporate entity, in which case the IP address being used for customer communications may belong to the enterprise, not being the IP address of the residential ISP. Such off-premises extensions may appear as part of an upstream IP PBX. On mobile devices, e.g., a 3G handset or USB wireless broadband adapter, the IP address has no relationship with any physical location known to the telephony service provider, since a mobile user could be anywhere in a region with network coverage, even roaming via another cellular company.
At the VoIP level, a phone or gateway may identify itself with a Session Initiation Protocol registrar by its account credentials. In such cases, the Internet telephony service provider knows only that a particular user's equipment is active. Service providers often provide emergency response services by agreement with the user who registers a physical location and agrees that emergency services are provided to that address only if an emergency number is called from the IP device.
Such emergency services are provided by VoIP vendors in the United States by a system called Enhanced 911, based on the Wireless Communications and Public Safety Act of 1999. The VoIP E911 emergency-calling system associates a physical address with the calling party's telephone number. All VoIP providers that provide access to the public switched telephone network are required to implement E911, a service for which the subscriber may be charged. "VoIP providers may not allow customers to "opt-out" of 911 service."
The VoIP E911 system is based on a static table lookup. Unlike in cellular phones, where the location of an E911 call can be traced using assisted GPS or other methods, the VoIP E911 information is accurate only if subscribers, who have the legal responsibility, keep their emergency address information current.

Fax support

Sending faxes over VoIP networks is sometimes referred to as Fax over IP. Transmission of fax documents was problematic in early VoIP implementations, as most voice digitization and compression codecs are optimized for the representation of the human voice and the proper timing of the modem signals cannot be guaranteed in a packet-based, connection-less network. A standards-based solution for reliably delivering fax-over-IP is the T.38 protocol.
The T.38 protocol is designed to compensate for the differences between traditional packet-less communications over analog lines and packet-based transmissions which are the basis for IP communications. The fax machine may be a standard device connected to an analog telephone adapter, or it may be a software application or dedicated network device operating via an Ethernet interface. Originally, T.38 was designed to use UDP or TCP transmission methods across an IP network. UDP provides near real-time characteristics due to the "no recovery rule" when a UDP packet is lost or an error occurs during transmission.
Some newer high end fax machines have built-in T.38 capabilities which are connected directly to a network switch or router. In T.38 each packet contains a portion of the data stream sent in the previous packet. Two successive packets have to be lost to actually lose data integrity.

Power requirements

Telephones for traditional residential analog service are usually connected directly to telephone company phone lines which provide direct current to power most basic analog handsets independently of locally available electrical power.
IP Phones and VoIP telephone adapters connect to routers or cable modems which typically depend on the availability of mains electricity or locally generated power. Some VoIP service providers use customer premises equipment with battery-backed power supplies to assure uninterrupted service for up to several hours in case of local power failures. Such battery-backed devices typically are designed for use with analog handsets.
Some VoIP service providers implement services to route calls to other telephone services of the subscriber, such a cellular phone, in the event that the customer's network device is inaccessible to terminate the call.
The susceptibility of phone service to power failures is a common problem even with traditional analog service in areas where many customers purchase modern telephone units that operate with wireless handsets to a base station, or that have other modern phone features, such as built-in voicemail or phone book features.

Security

The security concerns of VoIP telephone systems are similar to those of other Internet-connected devices. This means that hackers with knowledge of VoIP vulnerabilities can perform denial-of-service attacks, harvest customer data, record conversations, and compromise voicemail messages. Compromised VoIP user account or session credentials may enable an attacker to incur substantial charges from third-party services, such as long-distance or international calling.
The technical details of many VoIP protocols create challenges in routing VoIP traffic through firewalls and network address translators, used to interconnect to transit networks or the Internet. Private session border controllers are often employed to enable VoIP calls to and from protected networks. Other methods to traverse NAT devices involve assistive protocols such as STUN and Interactive Connectivity Establishment.
Though many consumer VoIP solutions do not support encryption of the signaling path or the media, securing a VoIP phone is conceptually easier to implement than on traditional telephone circuits. A result of the lack of encryption is that it is relatively easy to eavesdrop on VoIP calls when access to the data network is possible. Free open-source solutions, such as Wireshark, facilitate capturing VoIP conversations.
Standards for securing VoIP are available in the Secure Real-time Transport Protocol and the ZRTP protocol for analog telephony adapters, as well as for some softphones. IPsec is available to secure point-to-point VoIP at the transport level by using opportunistic encryption.
Government and military organizations use various security measures to protect VoIP traffic, such as voice over secure IP, secure voice over IP, and secure voice over secure IP. The distinction lies in whether encryption is applied in the telephone endpoint or in the network. Secure voice over secure IP may be implemented by encrypting the media with protocols such as SRTP and ZRTP. Secure voice over IP uses Type 1 encryption on a classified network, such as SIPRNet. Public Secure VoIP is also available with free GNU software and in many popular commercial VoIP programs via libraries, such as ZRTP.

Caller ID

Voice over IP protocols and equipment provide caller ID support that is compatible the public switched telephone network. Many VoIP service providers also allow callers to configure custom caller ID information.

Hearing aid compatibility

Wireline telephones which are manufactured in, imported to, or intended to be used in the US with Voice over IP service, on or after February 28, 2020, are required to meet the hearing aid compatibility requirements set forth by the Federal Communications Commission.

Operational cost

VoIP has drastically reduced the cost of communication by sharing network infrastructure between data and voice.
A single broad-band connection has the ability to transmit more than one telephone call. Secure calls using standardized protocols, such as Secure Real-time Transport Protocol, as most of the facilities of creating a secure telephone connection over traditional phone lines, such as digitizing and digital transmission, are already in place with VoIP. It is necessary only to encrypt and authenticate the existing data stream. Automated software, such as a virtual PBX, may eliminate the need of personnel to greet and switch incoming calls.

Regulatory and legal issues

As the popularity of VoIP grows, governments are becoming more interested in regulating VoIP in a manner similar to PSTN services.
Throughout the developing world, particularly in countries where regulation is weak or captured by the dominant operator, restrictions on the use of VoIP are often imposed, including in Panama where VoIP is taxed, Guyana where VoIP is prohibited. In Ethiopia, where the government is nationalising telecommunication service, it is a criminal offence to offer services using VoIP. The country has installed firewalls to prevent international calls being made using VoIP. These measures were taken after the popularity of VoIP reduced the income generated by the state owned telecommunication company.

Canada

In Canada, the Canadian Radio-television and Telecommunications Commission regulates telephone service, including VoIP telephony service. VoIP services operating in Canada are required to provide 9-1-1 emergency service.

European Union

In the European Union, the treatment of VoIP service providers is a decision for each national telecommunications regulator, which must use competition law to define relevant national markets and then determine whether any service provider on those national markets has "significant market power". A general distinction is usually made between VoIP services that function over managed networks and VoIP services that function over unmanaged networks.
The relevant EU Directive is not clearly drafted concerning obligations which can exist independently of market power, and it is impossible to say definitively whether VoIP service providers of either type are bound by them. A review of the EU Directive is under way and should be complete by 2007.

Arab states of the GCC">Gulf Cooperation Council">GCC

Oman

In Oman, it is illegal to provide or use unauthorized VoIP services, to the extent that web sites of unlicensed VoIP providers have been blocked. Violations may be punished with fines of 50,000 Omani Rial, a two year prison sentence or both. In 2009, police raided 121 Internet cafes throughout the country and arrested 212 people for using or providing VoIP services.

Saudi Arabia

In September 2017, Saudi Arabia lifted the ban on VoIPs, in an attempt to reduce operational costs and spur digital entrepreneurship.

United Arab Emirates

In the United Arab Emirates, it is illegal to provide or use unauthorized VoIP services, to the extent that web sites of unlicensed VoIP providers have been blocked. However, some VoIPs such as Skype were allowed. In January 2018, internet service providers in UAE blocked all VoIP apps, including Skype, but permitting only 2 "government-approved" VoIP apps for a fixed rate of Dh52.50 a month for use on mobile devices, and Dh105 a month to use over a computer connected." In opposition, a petition on Change.org garnered over 5000 signatures, in response to which the website was blocked in UAE.
On 24 March 2020, the United Arab Emirates loosened restriction on VoIP services earlier prohibited in the country, to ease communication during the COVID-19 pandemic. However, popular instant messaging applications like WhatsApp, Skype, and FaceTime remained blocked from being used for voice and video calls, constricting residents to use paid services from the country’s state-owned telecom providers.

India

In India, it is legal to use VoIP, but it is illegal to have VoIP gateways inside India. This effectively means that people who have PCs can use them to make a VoIP call to any number, but if the remote side is a normal phone, the gateway that converts the VoIP call to a POTS call is not permitted by law to be inside India. Foreign based VoIP server services are illegal to use in India.
In the interest of the Access Service Providers and International Long Distance Operators the Internet telephony was permitted to the ISP with restrictions. Internet Telephony is considered to be different service in its scope, nature and kind from real time voice as offered by other Access Service Providers and Long Distance Carriers. Hence the following type of Internet Telephony are permitted in India:

South Korea

In South Korea, only providers registered with the government are authorized to offer VoIP services. Unlike many VoIP providers, most of whom offer flat rates, Korean VoIP services are generally metered and charged at rates similar to terrestrial calling. Foreign VoIP providers encounter high barriers to government registration. This issue came to a head in 2006 when Internet service providers providing personal Internet services by contract to United States Forces Korea members residing on USFK bases threatened to block off access to VoIP services used by USFK members as an economical way to keep in contact with their families in the United States, on the grounds that the service members' VoIP providers were not registered. A compromise was reached between USFK and Korean telecommunications officials in January 2007, wherein USFK service members arriving in Korea before June 1, 2007, and subscribing to the ISP services provided on base may continue to use their US-based VoIP subscription, but later arrivals must use a Korean-based VoIP provider, which by contract will offer pricing similar to the flat rates offered by US VoIP providers.

United States

In the United States, the Federal Communications Commission requires all interconnected VoIP service providers to comply with requirements comparable to those for traditional telecommunications service providers. VoIP operators in the US are required to support local number portability; make service accessible to people with disabilities; pay regulatory fees, universal service contributions, and other mandated payments; and enable law enforcement authorities to conduct surveillance pursuant to the Communications Assistance for Law Enforcement Act.
Operators of "Interconnected" VoIP are mandated to provide Enhanced 911 service without special request, provide for customer location updates, clearly disclose any limitations on their E-911 functionality to their consumers, obtain affirmative acknowledgements of these disclosures from all consumers, and 'may not allow their customers to “opt-out” of 911 service.' VoIP operators also receive the benefit of certain US telecommunications regulations, including an entitlement to interconnection and exchange of traffic with incumbent local exchange carriers via wholesale carriers. Providers of "nomadic" VoIP service—those who are unable to determine the location of their users—are exempt from state telecommunications regulation.
Another legal issue that the US Congress is debating concerns changes to the Foreign Intelligence Surveillance Act. The issue in question is calls between Americans and foreigners. The National Security Agency is not authorized to tap Americans' conversations without a warrant—but the Internet, and specifically VoIP does not draw as clear a line to the location of a caller or a call's recipient as the traditional phone system does. As VoIP's low cost and flexibility convinces more and more organizations to adopt the technology, the surveillance for law enforcement agencies becomes more difficult. VoIP technology has also increased Federal security concerns because VoIP and similar technologies have made it more difficult for the government to determine where a target is physically located when communications are being intercepted, and that creates a whole set of new legal challenges.

History

The early developments of packet network designs by Paul Baran and other researchers were motivated by a desire for a higher degree of circuit redundancy and network availability in the face of infrastructure failures than was possible in the circuit-switched networks in telecommunications of the mid-twentieth century. Danny Cohen first demonstrated a form of packet voice in 1973 as part of a flight simulator application, which operated across the early ARPANET.
On the early ARPANET, real-time voice communication was not possible with uncompressed pulse-code modulation digital speech packets, which had a bit rate of 64kbps, much greater than the 2.4kbps bandwidth of early modems. The solution to this problem was linear predictive coding, a speech coding data compression algorithm that was first proposed by Fumitada Itakura of Nagoya University and Shuzo Saito of Nippon Telegraph and Telephone in 1966. LPC was capable of speech compression down to 2.4kbps, leading to the first successful real-time conversation over ARPANET in 1974, between Culler-Harrison Incorporated in Goleta, California, and MIT Lincoln Laboratory in Lexington, Massachusetts. LPC has since been the most widely used speech coding method. Code-excited linear prediction, a type of LPC algorithm, was developed by Manfred R. Schroeder and Bishnu S. Atal in 1985. LPC algorithms remain an audio coding standard in modern VoIP technology.
In the following time span of about two decades, various forms of packet telephony were developed and industry interest groups formed to support the new technologies. Following the termination of the ARPANET project, and expansion of the Internet for commercial traffic, IP telephony was tested and deemed infeasible for commercial use until the introduction of VocalChat in the early 1990s and then in Feb 1995 the official release of Internet Phone commercial software by VocalTec, based on the patent by Lior Haramaty and Alon Cohen, and followed by other VoIP infrastructure components such as telephony gateways and switching servers. Soon after it became an established area of interest in commercial labs of the major IT concerns. By the late 1990s, the first softswitches became available, and new protocols, such as H.323, MGCP and the Session Initiation Protocol gained widespread attention. In the early 2000s, the proliferation of high-bandwidth always-on Internet connections to residential dwellings and businesses, spawned an industry of Internet telephony service providers. The development of open-source telephony software, such as Asterisk PBX, fueled widespread interest and entrepreneurship in voice-over-IP services, applying new Internet technology paradigms, such as cloud services to telephony.
In 1999, a discrete cosine transform audio data compression algorithm called the modified discrete cosine transform was adopted for the Siren codec, used in the G.722.1 wideband audio coding standard. The same year, the MDCT was adapted into the LD-MDCT speech coding algorithm, used for the AAC-LD format and intended for significantly improved audio quality in VoIP applications. MDCT has since been widely used in VoIP applications, such as the G.729.1 wideband codec introduced in 2006, Apple's Facetime introduced in 2010, the CELT codec introduced in 2011, the Opus codec introduced in 2012, and WhatsApp's voice calling feature introduced in 2015.

Milestones