Skip navigation

STANFORD UNIVERSITY

INFORMATION TECHNOLOGY SERVICES

Voice Strategic Vision

Written by Christine Moe

This document covers the strategic vision for Stanfords voice services infrastructure that provides telephony services for the campus (both faculty and administration), all residential students, Stanford Hospital, and Packard Childrens Hospital. While a successful voice service must take into consideration business needs of the wide variety of clients, this document focuses on the technical aspects of the voice infrastructure. This document does not cover specific voice based applications that may meet the business needs of the above mentioned clients.

Principles

The goal of the voice services infrastructure is to provide voice based services that are robust, reliable, always-on, scalable and cost effective for the University using both Time Division Multiplexing and Voice over IP technologies. Additionally, wherever possible, the voice infrastructure will use IT Services adopted system standards such as the Linux operating system and centralized authentication.

Distinctions between voice and data services on the network are evaporating at an astonishing rate. The technologies have converged onto one media type, the IP network. Whether or not the services share a wire is a matter of logistics and cost, not technology. With speed advances of todays silicon, networks, servers and desktops can now support latent sensitive voice applications. Network hardware and processing engines are all becoming commodities. Historically strong voice system vendors are moving out of the specialized hardware market and into software as their primary voice related product that runs on these engines. New vendors are offering voice services on vanilla commodity hardware. Voice is now an application on the network, one which requires preference over large data packets. Additionally, Voice Over IP requires that we as service providers address support criteria that have been taken for granted in the traditional telephony world, items such as power and security. However, there is no compelling business reason for immediate replacement of all legacy voice services with IP based services.

Technologies

Stable core technologies:

  • Time Division Multiplexing (TDM):
    TDM services as a technology are well established and well understood. They are reliable, secure, and treated as a utility. Stanfords TDM based services matches or exceeds the grade of service provided by the telephony industry at large. Stanford will use the existing TDM equipment for its full useful life where possible while we deploy the core VoIP infrastructure and research the emergence of wireless technologies that could cause a fundamental shift in the direction of communication technologies and services. The TDM equipment currently meets the voice needs of much of the campus.

  • Voice Over IP:
    Stanford is positioning itself to take advantage of technology changes by a careful migration of voice services to VoIP technologies. The always-on expectation for voice services itself requires the supporting infrastructure (network, closets, and power) be engineered for non-stop, reliable operations. Duplicated equipment and network paths are required to provide the required level of reliable service. This infrastructure upgrade of the approximately 1,470 communications closets on campus is extremely costly but can be done over time. Since there is no compelling business reason for immediate replacement of all legacy TDM voice services with VoIP services, Stanford can deploy VoIP based services when and where it makes business sense such as in new or remodeled buildings, as well as off campus locations. A buildings infrastructure can be enhanced to support VoIP while other construction is being done. VoIP and TDM voice services can co-exist depending on the clients needs.

    Once a building is VoIP enabled, new set installations will be simplified. Relocation of telephone instruments can be performed by the client. The VoIP telephone has becoming another network device and voice another network application. As an application, the potential for expanded voice services using VoIP is huge. Services such as presence, a concept that allows clients to control their own contact parameters, will allow clients to ability to decide when, how and with whom they communicate.

  • Encryption with IP-SEC:
    Both the signaling channel and the media stream must be encrypted. The choice of encryption is industry standard IP-SEC.

  • Connectivity with SUNet:
    The voice IP infrastructure must connect both SUNet or the VoIP backbone with secure, protected paths.

  • Automatic Call Distribution (ACD):
    ACD provides even call distribution to all members of any help desk, call center, or hospital clinic. ACDs provide call statistics such as total number of calls answered, number of calls abandoned, maximum and average call holding times, and call wait times. Metrics such as these are critical to effective management of answering services. Stanfords hospitals and clinics use these metrics to aid in accreditation. Any new ACD must support:
    • Call routing by calling line ID, called party information, agent skill set, Integrated Voice Response input
    • Script controlled call flow
    • Traffic/event based call flow
    • Detailed management reporting
    • Real-time monitoring
    • SIP

Technologies new to IT Services:

  • Unified Messaging (UM):
    Unified Messaging is the next generation of Voice Mail services. Voice messages can be delivered to the e-mail desktop. E-mail messages can be enunciated over a voice path. Controlling commands can be either with a telephone keypad (Telephone User Interface, TUI) or via speech recognition.

Emerging technologies:

  • VoIP over wireless:
    Just as hardwired computer connections have their place in the network fabric, fixed or wired VoIP will remain cost effective in the voice arena. However, wireless VoIP is poised to supplant many wired VoIP implementations. With the demands of mobile clients and a mobile society growing daily, telephone set emulation on a laptop with wireless connectivity will become the preferred method of communication for many, especially students.
  • Presence:
    Presence is the ability for a client to choose who, how and when voice services are delivered.

  • Session Initiation Protocol (SIP):
    SIP is becoming the signaling method of choice in the VoIP world. It is poised to replace the legacy SS7 used by the Bell Operating companies for the previous 30 years. Because SIP is IP based, advanced services that require sophisticated signaling can now be accomplished without massive capital investments.

    • Informational applications on phones
    • Interactive applications on phones

Deprecated technologies:

  • TDM only services
  • Voice Mail without unified messaging
  • ACD without skills based and event driven routing

Projects

First:

  • Deployment of the CS2100 into the core infrastructure. This was completed in May 2005.
  • Encrypt both signaling and media streams with IP-SEC.

Next:

  • Add gateways that allow the IP sets to connect to the Public Switched Telephone Network (PSTN).

Later:

  • Add gateways that serve as packet front ends for the TDM equipment in the large Electronic Communication Hubs (ECHs).
  • Replace older VoIP controllers with current hardware.

Research

  • Understand the impact of wireless VoIP on the existing voice and data infrastructures. Items of concern in this area are security, capacity, and wireless coverage.
  • Determine which SIP engine and services meets the needs of the Stanford community.
  • Determine which presence engine meets the needs of the Stanford community.
  • Provide method for third party applications to utilize IP portion of VoIP sets.
Last modified Wednesday, 27-Jun-2007 03:23:12 PM

Stanford University Home Page