The office discussed in this paper, of which I am a staff member, is part of the Global Enterprise Services (GES) organization of the Multimedia Communications Services (MCS) division of Northern Telecom Inc. (Nortel). The Ann Arbor office has experienced many layoffs, reorganizations, and relocations over the past 20 years; and as recently as 7 years ago, employed nearly 300 people. Today, the office staff consists of thirty employees: 2 directors, 2 senior managers, 3 managers, 2 administrative assistants, and 21 technical support personnel, whose primary mission is supporting the corporation's internal computing infrastructure. The organization has a decentralized management structure (local managers with remote staff and local staff with remote managers) and all projects involve collaboration with other Nortel locations. The majority of activity is with Ottawa, Toronto, Nashville, Raleigh, Dallas, Marlton, NJ, and several locations within the UK.
There has been a concerted effort over the past six months, initiated by executive management, to drive down the travel expenses incurred by the MCS division, which is made up of approximately 10,000 employees. Prior to a travel freeze imposed in September, these expenses were estimated at $1 million a week. When possible, travel is to be replaced with videoconferenced meetings. A simplified breakdown of travel expenses would include the following categories:
The Ann Arbor office has had a PictureTel 4000 series videoconferencing room system since August, 1992. It consists of twin 37" monitors, a separate document camera, far-end camera controls, and speaker phone. The system has been rarely utilized over the years due to the liberal travel policies and the preference of person-to-person over video meetings. Additionally, the system is configured at 112 kilobits per second (Kbps) with 15 frames per second (fps). At that speed, the resolution is less than desirable and considered by most a distracting, off-sync, talking heads situation. Also, since it is an installed, stationary system, the cameras cannot be moved to view group activities in the technical services' lab, nor is it tied into any desktop PCs for data conferencing collaborations such as file transfers, applications sharing, or shared whiteboard activities.
Simply put, the office requires a better quality videoconferencing system from which to communicate and conduct company business. What equipment would it take to implement a reliable, cost-effective, TV-quality videoconferencing unit? Or what could be inexpensively deployed that would give the office good quality desktop units with the collaborative functionality to permit personal conferencing? I volunteered to research our options and their associated implementation costs. This has evolved into a 60-day commitment to install and evaluate certain equipment, on-going discussions with several vendors, a growing list of internal contacts involved in similar projects, and scheduled tests to determine the total bandwidth utilization of the Ann Arbor office. Since this is a work in progress, for the purpose of discussing the project within the scope of this paper, I will endeavor to describe: What elements of video conferencing are viewed as critical; What are some of the trade-offs; What equipment is being evaluated now; What will be evaluated in the near future. Although purchase price will be a significant factor in the eventual equipment selection, the prices I have received are under non-disclosure agreements and therefore, specific details cannot be revealed.
I have spoken with several equipment vendors (PictureTel, CLI, VTEL, BT, and Teleos) and have compiled various specifications and prices; this has raised my awareness of the proprietary modes, and the current and emerging standards in the marketplace. This technology has a great many vendor specific algorithms -- proprietary features that cannot work unless the receiving systems has the same setup. However, now that a number of international protocol standards have been agreed upon, there is a higher degree of interoperability between systems from different manufacturers. Below is an overview of some of the frequently used industry terms and standards:
At this point, there were several options to examine. As PictureTel is the market leader in this arena and since their equipment was already installed on-site, I first pursued the option of upgrading the existing equipment. The present equipment is running the oldest release of software and hardware possible but is upgradable to 384 Kbps (in H.320 standard mode) or a maximum speed of 768 Kbps (in SG3 proprietary mode). This is a limiting factor of the installed system; a standards based network interface greater than 384 Kbps cannot be achieved without a complete system replacement. Therefore, this upgrade path is viewed as a dead-end since it lacks support for H.320 at higher bandwidths -- considered a critical factor in any purchase of a room unit.
It was necessary to find out how videoconferencing requirements were being handled in some of Nortel's central locations. I discovered that a group in Ottawa was looking at many of these issues. From discussions with them, I learned that Nortel has a fairly large installed base of 120 videoconferencing rooms worldwide, which logged a total of 63,000 room hours in 1994. This group was in the initial stages of a soliciting vendor pricing for room units. They had likewise, dismissed any upgrade of equipment that would limit the H.320 compliant network interface to 384 Kbps. Through discussions that ensued, I gained a better perspective of some long-term, future directions relevant to videoconferencing for Nortel, such as broadcast video, ATM multiplexing, and support for the MPEG-2 codecs.
However, while this information was useful, it was still apparent that some work could be done concerning evaluation of strategic purchases that would improve user satisfaction in the immediate future, outside of a room unit. I was particularly interested in how the level of satisfaction of the Ann Arbor employees could be increased through the purchase of several individual desktop videoconferencing systems (DVS). The office environment is well suited for implementing DVS: Everyone in the office has a Mac or PC, most with 17" monitors or better; the need for collaboration through screen and application sharing is strong; and the digital transport infrastructure is there and can be expanded if necessary. The usefulness of the desktop features and the convenience of having them in a less formal setting could outweigh some of the detractions -- such as a much smaller video window and slower speed. I was aware that Nortel manufactures a desktop system named VISIT Video (which has seen some success in the market) and evaluation units could be acquired.
The initial information turned up concerning the VISIT product was not promising. The current release of VISIT (2.0) only utilized one channel of a ISDN BRI line (56/64 Kb) for video and is not H.320 or T.120 compliant. A new release of the product, due in January, is going to be H.320 compliant, support the full 128 Kb bandwidth of an ISDN BRI line, do 20-25 fps in QCIF mode (10-15 in FCIF mode) as well as support screen and applications sharing. I expect this version of the product will be a useful addition to the office and have added my name to the list for a beta version, anticipated to be released sometime in November, for installation in my Compaq P5. In the mean time, I have received a copy of the current product and installed it in a spare Mac IIci, so I will have something to compare the new release against.
I had recently received information regarding a video router which integrated the IMUX and MCU features. It was at this point that my discussions with Teleos Communications Inc., began. Teleos was quick to see an opportunity and supplied my office with a video router for the purpose of evaluation.
A new ISDN PRI circuit was installed between the in-house PBX and the Teleos in the video room. This allowed me to configure the MCU capabilities of the Teleos and test the IMUX features at various bandwidths. The Teleos is interfaced to the existing PictureTel system and I have successfully tested the unit at 336 Kbps (56 Kbpsx6) with a PictureTel unit in Toronto. According to Teleos, their system is H.320 compliant and the next release of the video router software will support conference speeds up to 768 Kbps, T.120 data conferencing, and a dial-in broadcast server for up to 900 sites. The unit also has 8 ISDN BRI ports which would permit the connection of future desktop videoconferencing units. The video group in Ottawa is now also interested in an evaluation unit, and I am involved in obtaining a second unit so our locations can evaluate performance between the two.
Videoconferencing technology involves a great number of components. Fortunately, manufacturers are migrating to the open systems approach -- adopting standards in the way that audio, video, and data is transmitted -- so there will be greater interoperability in the near future. This technology is migrating down to the desktop where collaborative groupware applications are opening up a new range of communications activities. Organizations will soon be able to move beyond textual and graphical information to video itself as an information form. Video communications, accelerated by hardware and software developments, and enabled by advances in network infrastructure, will soon make a fundamental impact on the workplace.