Technology Positioning Statement Report

3.1.5 Audio Teleconferencing, Digital Voice Communications

Description: Communications hardware and software to provide telephone-like voice access over Internet and other digital networks.

Category: 3 - Communication Technologies   Subcategory: 1 - General Purpose Communication Technologies


RealPlayer Plus
Windows Media Player


Industry UsageSC Usage

Performance Metrics

Audio quality; network quality/efficiency product; codec and transmission standards compliance; total cost impact.

Usage and Dependencies

Industry Usage:†Regardless of the products used, telephony traffic over a connectionless network demands high quality guarantees. It has quantifiable requirements and its value depends on these requirements being strictly met.† One way to provide high quality guarantees is to significantly over-provision a network. This can lead to large inefficiencies and needless costs. A more cost-effective approach is to implement QoS (Quality of Service) mechanisms such as RSVP signaling, which sends service requirement metadata along with the voice data.

Voice over Internet Protocol (VoIP) standards are still evolving, and interoperability remains an issue. The major standards that preside over VoIP technology are H.323, SIP, MGCP and Metaco/H.248.

H.323 developed by the ITU is considered the mature and stable protocol. However, it has major limitations due to its lack of scalability and complexity. Its leadership status is being challenged by newer protocols, such as H.248/Megaco, MGCP, SIP and SIP+.

Session-Initiated Protocol (SIP) is based on standards set by the Internet Engineering Task Force (IETF) and is used to initiate and terminate interactive sessions between Internet users. SIP is simpler, more scalable and goes beyond H.323 in creating voice applications. Many consider it similar to HTML and find it is easier for programmers to develop new applications. Its open design supports interoperability with H.323 and offers a migration path for service providers to deliver enhanced services. Many vendors are considering SIP+ but regard it as relatively new and untested.

Media Gateway Control Protocol (MGCP) is proposed by the IETF as the standard to convert audio signals carried on the public switched telephone network (PSTN) to Internet data packets. It has several variations based on different market needs that have resulted in compatibility and interoperability issues with equipment manufacturers. It is seen as losing ground to the emerging Megaco/H.248 standard.

Megaco/H.248 offers a richer set of capabilities that also corresponds to more complexity. It adds peer-to-peer interoperability capabilities and delivers a control function for IP devices. The protocol allows low-cost gateways to interface with signaling systems found in circuit-switched networks, reducing expenditures. It is still considered a new and unproven technology.

Interoperability between standards remains a major issue with vendors. Implementing a standard in a product does not assure that this will interoperate with another vendorís product that uses the same standard. Only a few vendors, such as Cisco, have announced other products that interoperate with their products. Although many equipment manufacturers cite some interoperability based on H.323, its impact as a leading protocol is declining. Interoperability based on SIP+ and Megaco/H.248 is seen as increasing. Much more work by vendors and standards committees on interoperability is needed before this issue is resolved.

Voice Extensible Markup Language (VoiceXML) is a computer language that is quickly becoming the standard for creating Web content and services that are accessible by the telephone. Leading providers of voice equipment and speech engines, such as Lucent, Motorola, AT&T and IBM, have contributed their markup language to the development of the open VoiceXML standard. Prior to the VoiceXML Forum, application developers had to use proprietary languages that were developed by each vendor. With the standardization of VoiceXML, improvements in voice applications are simpler to develop, resulting in a shorter time to market.

Many consider phone access to the Internet the next big market. The rapid rise in cell phone use and ubiquity of telephones create a demand for new self-service applications that allow callers to interact with Web content for information and transactions. With the use of natural-language speech recognition and text-to-speech technology, a phone user can browse the Web simply by talking and listening to responses, using many of the same basic commands as browsing with a PC. While not a replacement for complex Web searches, voice access to the Web supports anytime, anywhere communications.

VoiceXML as defined by the VoiceXML Forum is designed for creating audio dialogs that feature synthesized speech, digitized audio, recognition of spoken and DTMF key input and recordings of spoken input. In May 2000, the VoiceXML Forum announced the World Wide Web Consortium (W3C) acknowledged the submission of Version 1.0 VoiceXML specifications. Acknowledgement by W3C helps accelerate the reach of the Internet though voice-enabled Web content, and the W3C Voice Browser Working Group will use VoiceXML as the basis for its dialog markup language.

The move toward using VoiceXML as the standard for dialog markup language is seen in the number of member companies supporting this language. VoiceXML Forum now has more than 150 member companies and includes major voice recognition vendors and voice communication companies. The need to simplify the creation of Web-based applications is apparent, and VoiceXML appears to be the agreed upon standard going forward.

SC Usage: †Audio (and video) telecommunications in meeting rooms is provided by PictureTel systems. Analog telephone (POTS) service is provided by Verizon and funded at the departmental level, so there is no financial incentive to move to IP-based voice communications at this time.

In the future, if higher-bandwidth audio and video communications are to be supported over TCP/IP, it is likely that new provisioning of the network will be needed. It is also recommended that network QoS (Quality of Service) mechanisms be established prior to heavy usage of Voice over IP to the desktops.

SC Application Impacts: None in near term.

Last Update: Valid Until:


Previous TPS Report
VoiceXML Gains Widespread Acceptance With Leading Vendors, Elizabeth Herrell, Giga, 7/5/2000.

List all Categories

Administer the Database