4
Facilitating Distributed Work

IDENTIFYING BARRIERS TO EXPANSION OF DISTRIBUTED WORK

In addition to the organizational and social processes that constrain adoption of distributed work, there are many technical barriers to its wider acceptance. Good database technology is not in place yet; applications in current use are expressed as massive, monolithic code structures that cannot easily be rewritten; and new applications operate in ways that are not as familiar as the message-passing routines of electronic mail and will therefore meet resistance from customers. Although many of the technical and design issues can be addressed in research laboratories, the issue of user acceptance must be considered at other levels as well: until software producers see a sufficient market for these products, for example, they may be reluctant to make the large investment in research and software development required to overcome the significant barriers.

Complexity-of-Use Barriers

Today, to effectively engage in distributed work or telecommuting, one must absorb a large amount of technical detail and have access to an array of equipment. Before they can become ubiquitous, the systems and equipment must be made easier to understand, and certainly



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 46
Research Recommendations: To Facilitate Distributed Work 4 Facilitating Distributed Work IDENTIFYING BARRIERS TO EXPANSION OF DISTRIBUTED WORK In addition to the organizational and social processes that constrain adoption of distributed work, there are many technical barriers to its wider acceptance. Good database technology is not in place yet; applications in current use are expressed as massive, monolithic code structures that cannot easily be rewritten; and new applications operate in ways that are not as familiar as the message-passing routines of electronic mail and will therefore meet resistance from customers. Although many of the technical and design issues can be addressed in research laboratories, the issue of user acceptance must be considered at other levels as well: until software producers see a sufficient market for these products, for example, they may be reluctant to make the large investment in research and software development required to overcome the significant barriers. Complexity-of-Use Barriers Today, to effectively engage in distributed work or telecommuting, one must absorb a large amount of technical detail and have access to an array of equipment. Before they can become ubiquitous, the systems and equipment must be made easier to understand, and certainly

OCR for page 46
Research Recommendations: To Facilitate Distributed Work easier to use. For example, a sales person who needs to access a customer accounts database while at home or traveling must have, at a practical minimum, a V.22bis-standard, 2,400-bps modem connected to, or internal to, a portable computer; communications software; and an analog telephone line with an accessible RJ11, or similar, standard jack. Additionally, most communications software packages require that the user be at least minimally aware of communications parameters such as baud rate, parity, data bits, stop bits, and local echo. All of these requirements may be fairly easy to manage at home. However, if a hotel or customer site lacks telephone jacks, or has only digital telephone lines, accessing the customer database can be impossible. In fact, inveterate mobile workers can sometimes be identified by the contents of their well-stocked portable computer cases: screwdrivers and pliers to disassemble telephones, various patch cords with alternate telephone plugs or alligator clips, handset adapters, audio-coupled modems, and digital line adapters. Users who need to have full, mobile access to their company's computer networks must master additional levels of technical sophistication ranging from integrated services digital network (ISDN) telephone line standards to the technicalities of Internet addressing and packet transport protocols. Inconsistent and inadequate user interfaces also increase complexity-of-use barriers and interfere with the adoption of new computing and communications technology. This is particularly true for distributed workers, who may find themselves using a variety of communications devices and services but in ways only incidental to their primary work tasks or to the primary use of the communications device. For example, it is often a matter of trial and error to determine which of the many services possibly used by a mobile worker, such as call waiting or cellular telephone services, interfere (and in exactly what way) with particular computer communications and how this interference can be eliminated. Often, the ability of end users to take full advantage of flexible and programmable services is limited by cryptic user interfaces based on touch-tone telephone keypads. Consistency of interfaces and connections across a range of environments are equally important to the mobile worker. Today, for example, a cellular telephone user often needs to make several attempts to dial a long-distance call while roaming in another cellular provider's system. This will certainly become exacerbated when users more routinely wander from areas with high-bandwidth, multimedia connections, to localities with low-bandwidth connections, to remote locations providing high bandwidth from a distant network

OCR for page 46
Research Recommendations: To Facilitate Distributed Work host. Ubiquity of capabilities—the common denominator, which may be relatively low—will be a key constraint on the speed of distributed work applications. For the foreseeable future, learning to use networks effectively will require a great deal of human support. It has even been possible to make a business out of providing that support. For example, Omnet, the network company that provides electronic mail and bulletin board services to oceanographers, has a toll-free telephone number enabling an oceanographer to talk to a real person who will tell him/her how to set the parity bits or the local access phone number for accessing electronic mail from Bogota, Bulgaria, or Burlington. Omnet also manages electronic mail distribution lists for distributed projects, ensuring that everyone's electronic mail address is up-to-date and providing hard-copy postal or facsimile delivery to people without electronic mail. Other value-added network vendors offer similar services. More generally, the spread of sophisticated computer-based technologies from the sophisticated pioneering user base into the general population has increased the need for support services, which are provided by both employers and system vendors. Cost-of-Technology Barriers The most difficult challenge is how to deploy high-speed communications networks, connections, and equipment at a reasonable price and on a wide scale. As the National Information Infrastructure is advanced and the nation's economy becomes increasingly dependent on information, equitable and inexpensive access will be important to a very large segment of the population. Inclusive policies and deployment decisions will best serve both public and private interests (CSTB, 1994b) and permit wider utilization of distributed work practices. In fact, many modern information technologies depend on wide utilization for their effectiveness and efficiency. For example, it is instructive to examine the recent, rapid spread of facsimile machines. The basic principles of facsimile transmission were developed in 1842 and thus precede even the telegraph (McGraw-Hill Encyclopedia of Science & Technology, 1987). The first significant commercial use was by large news organizations that began sending ''wirephotos" in the 1930s. However, the high cost and low speed of the early equipment severely restrained adoption of the technology. Office use began to increase with the introduction of digital facsimile equipment and lower-cost long-distance telephone service in the late 1970s. In the early 1990s inexpensive integrated circuit chips became available, speeds increased, prices dropped drastically, and a "critical

OCR for page 46
Research Recommendations: To Facilitate Distributed Work mass" of installed machines was reached. The market almost literally exploded, and the least expensive of the machines became home consumer items. If telecommuting and other forms of distributed work are to increase, lower-cost communication devices and services will be needed. In 1993, high-speed modems generally cost between $100 and $350, and faster ISDN terminal adapters were about twice as expensive, still without offering the bandwidth needed for multimedia and video applications. Video codecs, devices necessary for serious two-way, real-time video work, cost 10 times the amount of current ISDN adapters. Additionally, inexpensive, high-quality displays are clearly needed. Currently a good-quality 17-inch computer monitor, which is generally considered necessary for serious desktop publishing efforts, costs $1,200 or more. This will need to decrease by a factor of at least three before it will be affordable to most users. While general market forces will probably continue to reduce unit prices, researchers and individuals making decisions about distributed work should remember the importance of reducing prices for both infrastructure items and services. DESIGNING COMMUNICATIONS INFRASTRUCTURE TO SUPPORT A RANGE OF CAPABILITIES FOR DISTRIBUTED WORK High-speed, broadband communications will be needed to handle the multimedia environment envisioned for both distributed work and the National Information Infrastructure. The next few years should see increasing amounts of bandwidth available due to the deployment of new network facilities and services such as asynchronous transfer mode (ATM) and other broadband network backbones running at speeds ranging from 45 megabits per second to a gigabit per second, or more. However, most of the research in this area has focused on point-to-point communications of a single medium; almost nothing has been done in the area of multiparty, multimedia technologies. Supporting Typical Work Routines Currently, the ability to engage in distributed work depends to a large extent on the availability of communications services to support typical activities of an individual or group working remotely, such as the following:

OCR for page 46
Research Recommendations: To Facilitate Distributed Work Contacting another individual, or a group of individuals, using regular telephones and circuits; Exchanging messages by electronic mail, voice mail, or facsimile transmission; Exchanging documents or images by electronic mail or facsimile transmission; Accessing and using computers or computer networks from a remote location, an activity often referred to as a remote log-on; Sharing computer files locally and remotely by using explicit schemes such as file transfer protocols, or by using indirect techniques such as distributed file environments; and Engaging in videoconferences with participants at two or more locations. These activities impose different requirements for data transmission rate, or bandwidth, on the supporting technologies. The data transmission rate is a measure of the speed with which a given technology can exchange data and is normally measured in bits per second (bps). The size of data objects (e.g., pages of text, still images, digitized audio clips, movie-like sequences of still images) is measured in bits, and dividing the size of the data object by the transmission rate yields the minimum time that it takes for an object to be transmitted, disregarding any "overhead" for processes like error checking and retransmission. In practice, about 10 percent of the bandwidth is lost to overhead in asynchronous communications. For example, one page of simple text with no fancy formatting or graphics typically comprises at least 24,000 bits, or 3,000 bytes, of information. At a nominal 2,400 bps, a slow but still common modem speed, it takes at least 11 seconds to transfer each page of text. Text that must be transmitted with specific fonts or formatting, or in word processor or spreadsheet format, requires a somewhat larger number of bits depending on the features and formats that are included. A 1-megabyte file takes more than 1 hour to transfer at 2,400 bps but only a few seconds over a very lightly loaded network running at 10 megabits per second (Mbps). Depending on the resolution and number of shades of gray required, a one-page, black-and-white graphic image comprises between 1 and 4 megabytes. Bandwidth is even more critical for media that must be transmitted in an ongoing data stream such as audio or video. One second of telephone-quality audio requires approximately 64,000 bits. Full-color, full-motion video conforming to regular U.S. television standards (as set forth by the National Television Systems Committee) requires a bandwidth of 180 Mbps. These numbers can be reduced by limiting

OCR for page 46
Research Recommendations: To Facilitate Distributed Work the resolution of the image, transmitting fewer frames per second, and utilizing various compression algorithms. However, bandwidths on the order of 1.5 Mbps to 8 Mbps are still required to achieve reasonable quality. From these considerations, it is easy to see that video applications are far more demanding than audio applications, which are in turn more demanding than still-image and text-only applications. Envisioning the Applications Enabled by Unlimited Bandwidth The introduction of microprocessors and their rapid penetration of the market have resulted in a drastic reduction in the cost of computer processing cycles. The National Information Infrastructure (NII) will have a similar effect on the cost of communications: it will make long-haul, high-bandwidth connections less expensive and more widely available. However, lower-bandwidth cellular and wireless technologies will also become widely available, and low-bandwidth, analog telephone lines will continue to be used in many homes and sparsely populated areas. Thus, distributed work applications utilizing the NII will need to be integrated well across different network technologies and bandwidths. Important research areas will include system infrastructure, user interfaces, and developing applications that will effectively use the varying bandwidths available. For example, the processor, the internal data communications bus, and the input/output systems of desktop computers are now very powerful, with further increases coming at a rapid rate. How could high network bandwidth and abundant, inexpensive desktop computing power be combined and integrated to enable, for example, useful and innovative distance learning applications to improve education? Some possibilities, with widely differing bandwidth requirements, might include the following: Dynamically creating a single video data stream with the "audience" incorporated into the image. This would allow, for example, each participant in a distributed industrial or professional training session to pan and zoom within the videoconference or presentation just as he or she might look back and forth between the presenter and the other participants in a small workshop or seminar; Dynamically analyzing the facial expressions of the audience to create different measures and displays summarizing audience response (e.g., 30 percent of the audience shows evidence of understanding what is being presented);

OCR for page 46
Research Recommendations: To Facilitate Distributed Work Capturing and organizing audio questions from audience members during the lecture that can be reviewed and/or responded to by the lecturer later. In other words, members of the audience can create private audio messages that are passed to the lecturer; Capturing and indexing the lecture so that members of the audience can more easily locate and review what the lecturer said; Providing audience members access to the source material on which the lecture is based so that they can browse through it during the lecture; and Broadcasting several video streams and allowing audience members to select among them or to watch several at the same time. Similar possibilities exist for other applications. The key idea is to determine how bandwidth could be used if it were effectively unconstrained, so as to better understand its usefulness for distributed work. Enabling Integration of Low-and High-Bandwidth Applications At the same time, bandwidth limitations will exist for traditional telephone service, wireless connections, and cellular services for a long time. Thus, applications and devices that can provide essential functions using limited bandwidth and yet be integrated well with devices and services provided over high-bandwidth networks also should be developed. One example might be to allow a person to cooperate in a videoconference using a personal digital assistant (PDA) with limited computational power. While the video portion would likely be a casualty of limited bandwidth, the audio channel could be played and a shared drawing browser could be used in a discussion between a field repair person and a technical support desk. The audio channel might be sent over a cellular network, and the shared drawing tool commands might be sent over the PDA wireless network. More generally, communications protocols that have little overhead and that can utilize processing power to substitute for bandwidth are fertile areas for further technical research. MEETING REQUIREMENTS FOR MULTIMEDIA COMMUNICATION The distributed work environments of the future will clearly involve the transmission and reception of multiple types of communications traffic, including voice, data images, video, and data files. Multimedia communication is particularly demanding because video

OCR for page 46
Research Recommendations: To Facilitate Distributed Work and audio data must generally arrive in real time, although delays may be tolerated in the receipt of data files and images. The objective, in the general sense, is to enable people to engage in multimedia communications independent of location, distance, computing and communications environment, or the number of parties involved in a given collaborative session. User Requirements The type of distributed work session might range from that involving a single individual carrying a wireless-network computer interacting with a remote database to a large-scale multimedia teleconference or group work session involving many remotely located parties as well as specialized servers for video, image, and data. Individuals in such a multimedia, multiparty conference or work session must be free to come and go. Subconferences might be convened and later, the full session reconvened. Different types of media might be transmitted to all or subsets of the group at any given time. Participants in a distributed work environment do not want to be aware of the underlying communications infrastructure required to make their sessions or conferences possible. They want to turn on intelligent systems, indicate with whom they wish to communicate, and start the session within, almost literally, the blink of an eyelid. They want the chance of a session setup being denied due to lack of network resources to be as small as it is currently for voice telephone calls—about 1 percent or less. While users may recognize that video, voice, and images in a wireless environment may not be received with the same fidelity as in the wired environment, the quality of service clearly has to be acceptable no matter where users are located and how they move. Quality-of-service requirements vary greatly, depending on the type of data being transferred. For example, computer data files must be transferred without loss or corruption but can generally be delayed during transmission. Real-time voice, on the other hand, is recognizable in spite of some loss or corruption during transmission, but substantial loss or delay results in unacceptable degradation. The delay and loss requirements for real-time video over communication networks are not completely understood at this time and depend on the kind of compression techniques used in transmitting the video and on the quality of image that is required. A multimedia, multipoint network must ensure that the different media arrive in reasonable synchronism at any given receiver site and appear at nearly the same time at the different sites.

OCR for page 46
Research Recommendations: To Facilitate Distributed Work As an example, consider a three-party work session with users located in New York City, Boston, and Honolulu. The New York user sends a multimedia message involving video with audio, a number of images, and some accompanying data, with each medium to be displayed on a different window on the recipients' screens. Four different kinds of media are involved. Assume that the transmission to Hawaii encounters a network problem resulting in some loss of data. The lost data are repeated a number of times before finally arriving correctly at their destination. The video and its accompanying audio information are separated from each other and from the data, following different paths because of differing bandwidth requirements. How will the audio and video be appropriately resynchronized at the two receiving workstations, and how will the system and/or users handle latency problems resulting from the fact that the complete, correct multimedia data took longer to reach Hawaii than Boston? Issues such as these must be addressed in the context of research into multimedia, multipoint protocols. Technology Requirements At least three enabling technologies are required for effective multimedia communication among users distributed in time and space. A high-speed, broadband, wired communications infrastructure to handle diverse multimedia traffic; A wireless networking environment capable of handling multimedia traffic; and Multimedia communication protocols appropriate over wide areas and involving multiple parties in heterogeneous computing environments, i.e., multimedia, multiparty protocols. Multimedia, multiparty communications capability is already available to a limited extent over local area networks and is clearly a topic of great interest among computer, workstation, and communications vendors. Multimedia communication over wide areas is still in its infancy, with many research issues yet to be addressed. A limited amount of multimedia protocol design for the Internet has been carried out. Some work on point-to-point multimedia communication protocols, as contrasted to multiparty communications (group work sessions and teleconferences), has appeared in the technical literature. However, the bulk of the work carried out thus far on multimedia communications has focused on workstation design and issues such as the synchronization of different traffic types. In the terminology

OCR for page 46
Research Recommendations: To Facilitate Distributed Work of network communications architecture, research attention has been focused principally on the "higher-layer" protocols. Relatively little work has been carried out on the communications infrastructure, including the transport layer, required for long-distance multimedia communications over wide area networks. The recently published Realizing the Information Future (CSTB, 1994b) has called for an expanded program of research to support an Open Data Network architecture that would facilitate multimedia traffic.