November 1997

BY JIM CLIFTON AND ERIC NEWMAN

Collaborative computing is a phrase that conveys little to the average user. In simple terms, collaboration means working together. In the context of networks and remote workers, it can mean almost anything from e-mail to groupware to real-time conferencing. In end-user terms, networkbased collaboration means people working together from remote sites as if they were in the same room, regardless of method.

Collaboration, in this final sense, includes a wide range of solutions — with video, audio, and data as the basic building blocks. We can understand the collaborative landscape as a tradeoff between effectiveness and cost for these three components. For successful virtual meetings, users need a solution that closely approximates the real thing while keeping both real and soft costs, such as training, to a minimum. For example, full motion video to the desktop with no latency would effectively approximate face-to-face meetings, but technical and cost restraints make this approach difficult today.

One of the big questions with collaborative computing today is how to handle the audio aspects of a conference. Interactive audio and video conferencing technologies that utilize the Internet and corporate Intranets are available today, but they are limited by bandwidth, latency, and scalability constraints. The key to solving the network issue in the short term is to hit the sweet spot of collaborative computing — using packet network-based data collaboration in conjunction with POTS audio. When these two networks are brought seamlessly together in a conference, collaboration becomes an effective alternative for face-toface interaction.

AUDIO/DATA SWEET SPOT
Generally, effective work gets done with people conversing about documents (slides, proposals, contracts, spreadsheets, and so on). While video has its place, the need to see someone throughout the work process is less important than the ability to interact with that person verbally and review supporting documents in real-time. For instance, when someone makes a suggestion regarding a contract or product proposal, all meeting participants need to hear and view the proposal as it was intended. Actually seeing the person making the proposal is of much less value to evaluating the suggestion and reaching a decision quickly. For most business meeting scenarios, the real value of remote collaboration lies in audio and data interaction.

Role Of Audio
In remote collaborative sessions, audio is vital for effective communication. Audio conference calls are the simplest and most common substitutes for face-to-face meetings. Today, most consumers and businesses rely on highquality, full-duplex audio over the POTS network, because it is a tool that they can take for granted. Nearly everyone has a phone on their desk which allows them to take advantage of highquality conference bridges, PBX multipoint conferences or point-topoint calls. Until the quality of IP audio improves, POTS will likely remain the audio network of choice.

When using audio only, however, meeting participants lose a great deal because they cannot easily interact as a group with the agendas, documents, and presentations around which traditional meetings usually revolve. A speaker can fax, mail, or e-mail this data ahead of time, but has no way to control the information that meeting participants are viewing, introduce new visual content, or capture interactive exchanges.

Role Of Data
Participating in a data conference solves this problem. Data includes the information, documents, and images — the meeting materials — with which we get work done. To approximate face-toface meetings successfully, users must be able to share visual information and collaborate as a group in realtime. Fortunately, along with standard telephones, most business users now have access to desktop computers equipped with Web browsers. With the advent of the Internet, engaging in virtual data conferences on an enterprisewide scale has become as simple as surfing the Web. Now, everyday users are collaborating remotely in ways that previously seemed difficult to conceive, much less accomplish.

THE INTEGRATED SOLUTION
The integrated audio/data solution is promising in that it takes advantage of ways people are already doing work. For widespread acceptance of such a solution, however, it will have to be easy to use, useful for real-world applications, and universally accessible to all meeting participants. CTI provides the most effective means of bringing audio and data networks together because it meets all of these requirements.

Users who want the convenience of managing both audio and data from their desktops can get quality connections by using a PC for data and a handset for audio. Using CTI techniques, desktop telephones perform the same reliable service that they have in the past, but with a twist. The same server that bridges data collaboration over the Internet can control the POTS audio bridge and manage conversations in the context of the meeting. Real technology exists today to connect IP audio to the POTS gateway through the desktop as a server. By leveraging the Web browser as a user interface, users can easily tie real-time data collaboration with CTI-driven audio bridging.

Such an approach allows users to take full advantage of the desktop tools and resources with which they are already comfortable. They can integrate the virtual conference planning with existing scheduling and organizational software. They can also use Internet resources for locating others, such as e-mail and LDAP-based directory services, to invite conference participants by double clicking on a name in the conference listing. Alternatively, in a “meet-me” paradigm, everyone scheduled to attend the meeting can go to a Web address and join the same conference. All of these processes are a natural extension of everyday business practices. Also, by combining the data and audio call setup into a single Web interface, the technology meets the ease-of-use requirements necessary for widespread deployment.

FUTURE
In the future, the market for realtime collaboration will take on many innovative forms. For the most part, IPbased data collaboration will continue to be an effective tool when tied to POTS audio conferencing, and a number of vendors are developing products to meet this need. In addition, an outpouring of IPbased audio and video conferencing products will become available for the end-user. This genre of collaboration technologies will become an important means of communication for many businesses as they look to better utilize the ever-increasing bandwidth and power of their Intranets and the global Internet.

The resolution of these two types of audio and video collaboration technologies will be an important issue. Not every user will have IP-based audio/video conferencing systems at their desktop and, in some cases, users may no longer have traditional phone handsets either. For example, a traveler whose hotel room has only one phone line can participate in a POTS audio conference, but loses the ability to engage in data collaboration. An alternative in such a situation will be IPbased audio in conjunction with IP-based data collaboration technology, sharing the same line. For this to work, we will need gateways and other infrastructure devices to translate between the world of IPaudio and those who use POTS audio. As the popularity of IP-based audio and video products grows in the marketplace, the need for new CTI technologies will become more important than ever.

Jim Clifton and Eric Newman are product managers at DataBeam Corporation, a supplier of standards-based collaborative software and developers’ tools for the Internet, corporate Intranets, and telecommunications networks. The company’s H.323 and T.120 toolkits enable third-party communications, networking, telephony, and conferencing products with real-time collaboration capabilities. DataBeam’s awardwinning applications and servers are used by Fortune 500 companies and major organizations worldwide. For more information, visit the company’s Web site at www.databeam.com

To demonstrate the feasibility of integrating an audio bridge with Internet data collaboration, DataBeam developed a logical way to tie a Dialogic board-based POTS audio bridge with DataBeam’s own Internet/Intranet-based data collaboration server, the neT.120 Conference Server. In this scenario, conference participants use a regular phone to call the audio bridge and communicate with other people in a neT.120 data conference. There is a one-to-one relationship between the data conference and the audio conference hosted by the audio bridge. Hardware Requirements

• Pentium 133 with 48MB of RAM.
• Windows NT 4.0 Server.
• 16-channel Dialogic D/160SCLSI.
• 64-channel Dialogic DCB/640.
• neT.120 Conference Server.

Software Requirements
In order to incorporate audio bridge features with the neT.120 Conference Server, we modified the neT.120 Web-based user interface slightly to include additional HTML pages and an additional CGI script that performed the necessary database operations. The audio bridge itself consisted of a stand-alone software application that processed incoming calls, validated audio conference IDs, and managed the audio conferences. We also wrote the necessary integration code using the Dialogic NT API for use with the DCB/640 to allow flexible integration with the neT.120 Conference Server.

The Dialogic API has several options for programming both synchronous and asynchronous models which are the basic nature of call processing. The real challenge was expanding the API provided for the DCB/640 hardware. Conceptually, we needed a more abstract and object oriented model that could manage the dynamic nature of conferencing. To do this, we created a primitive C++ framework of classes that abstracted the idea of a central conference manager, conferences, conference resources (each DSP on the DCB/640 is a conference resource), and finally the conferees (the callers themselves).

Integration
The complexity involved in integrating the audio bridge with the neT.120 Conference Server varies directly with how tightly bound developers want the audio and data to be. In our case, our goal was simple, to create a unique association between a neT.120 conference taking place on the server and another bridged audio conference on the DCB/640 hardware.

Thus, our primary change was to ensure that each neT.120 conference would have only one conference ID. Because neT.120 conference names are alphanumeric and can be as many as 255 characters long, we created a CGI script that generates a unique fourdigit number that callers can use to identify the conference they are participating in. This identification number provided enough information to create a unique audio conference where conferees could speak only with others in the same neT.120 data conference. On our initial audio bridge server, we chose only to allow inbound calls to join conferences. Subsequent development efforts did involve conferees joining the audio portion of the conference via a Java-based applet launched from the neT.120 conference in which they were participating.

Call Flow
In this configuration, the user creates a neT.120 conference via the neT.120 Conference Center at which point a unique four-digit audio conference ID is generated. Callers use this unique audio conference ID to join the audio portion of the neT.120 conference. While participating in the neT.120 conference, users can choose to view the audio bridge instructions that will display the audio bridge phone number and the neT.120 unique audio conference ID.

The users then call the audio bridge, at which point they will be prompted for the audio conference ID for the neT.120 conference they are attending. If a neT.120 conference with the matching audio conference ID is found, the caller is placed in a bridged audio conference with other conferees. Bridged calls to the conference continue until everyone leaves. At that point, the bridged audio conference terminates.