Click for next page ( 37

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 36
Personal Computer Networks Robert M. MetcaZfe * There are three kinds of people in the woricI: the technologists who believe that technology can, will, ant! should turn the world upside down and who are engaged principally in revolution; the users who believe that everything is moving too quickly, wish it were 1965 again, and whose principal activity is counterinsur- gency; and those who recognize that technology will reach useful application more slowly than most technologists would like ant! more quickly than most users wouIct like. The progress of this technology is measured not by any absolute timeframe, but on the basis of how successful we are in matching the superstructure of technology to the infrastructure of organizations. Local networking, for example, is one technology whose super- structure is now being matched, successfully, to the infrastruc- tures of various organizations. It is a technology bent on revolu- tion: in my view, the next clecade can be characterizes! from a computing standpoint as the clecade of the local networking of personal computers. This represents the third major phase in the history of computers and in the application of computing technol- ogy. Not Tong ago computing meant batch processing on main *Robert M. Metcalfe is founder, chairman, and vice-president of strategies and proj- ects, 3Com Corporation, Mountain View, California. 36

OCR for page 36
PERSONAL COMPUTER NETWORKS 37 frames the first phase. Then computing became timesharing on minicomputers the second phase. Now we are moving more and more into local networking of personal computers. Each phase broadens the market for computing ant! the number of applica- tions for computing and brings more ant! more computing to a greater number of people. Networking adds a sixth category to Mitchell Kapor's five ap- plications word processing, spreadsheet, clatabase, graphics, and communications. What are variously called local computer networks, local area networks, or more frequently just network- ing answer the need for a system to connect huncireds of com- puters on separate desks. There are approximately 200 varieties of local networks now being sold, of which Ethernet is only one. There are two approaches to this third stage of computer devel- opment: what I call the ethercentric view and the boxcentric view. It is easy to identify which is which. Ask people to draw their computer systems on the blackboard. If they begin by drawing a box they are boxcentric. If they begin by drawing a sweeping line across the board they are ethercentric and have adopted the revo- Jutionary view that communication is central. Computers are viewed as the array of resources around communication; they are seen less as arithmetic devices and more as communication tools. With the arrival of personal computers particularly the IBM PC in large enough numbers and with sufficient power to be use- fully connected at high speed in local networks, the pressures for the development of networking became overwhelming. Now that the IBM PC is here, many believe that all progress can stop. I do not subscribe to that point of view; I do believe that many other personal computers will continue the trend toward increased com- puter power on the desk. But the important crossover has already occurred! between the cost of providing multimegabit, high-speect local networking, as represented by Ethernet, and the density and power of personal computers. This means that it is now im- portant enough and cheap enough for a small portion of the indus- try to get involved in networking personal computers. So the ceil- ing tiles are coming down again and more cables are going in. For what are these local networks of personal computers being used? First there are the basics, such as peripheral sharing. With more and more personal computers economics becomes increas- ingly important. Sharing resources, principally printers and disks, is the first step. We start out by trying to do what we are

OCR for page 36
38 MANAGING MICROCOMPUTERS already doing more cheaply. We then try to do things we haven't done before, which begins the second phase of local networking. This second phase involves the use of local networks to give access to personal computers or to give personal computers ac- cess to information. Most of the information currently computer- ized is on mainframes. Thus, a high-priority item is to give per- sonal computers access to mainframes. That is why the IBM PC 3270, which allows PCs to act as terminals and get information that already exists on mainframes, is so significant. As less and less of our information exists on mainframes, however, the prior- ity will shift to personal computers communicating with each other. Now that personal computers exist, there is a trend to bring data into the local workgroup where they are generated and used. In the future, I believe we will see as much as 80 percent of the computerized data being kept not on a corporate mainframe but on a departmental or even a worEgroup-shared file system. The third and final use of local networks of personal computers is as tools for communications. This brings up the subject of elec- tronic mail. In 1970 the appeal of electronic mail was that you didn't have to move paper around anymore you could move electrons. And the transmission of electrons was much less expensive than the trans- mission of cellulose. It appeared that electronic mad! was con- cerned with transmission, moving information, and the econo- mies of moving electronic information. However, we quickly realized that although the cost of sending the information might be low we were spending 15 dollars to prepare the document we were sending. Electronic mail was synonymous with its prepara- tion, and it became very preparation-intensive. We also realized that we were spending a lot of time moving electronic mail from desk to desk manually after computers pre- pared and transmitted it. The fact is, most people don't want to send messages from one post office to another, they want to send them from one desk to another. Electronic mail then became a distribution problem-specifically the development and mainte- nance of distribution lists. We are now in the distribution phase of this trend in personal communication. The focus of current prog- ress is the creation of distribution mechanisms for electronic mail. And the computer industry is moving into the next phase. Because local networks of personal computers are so effective at generating and delivering electronic messages, electronic mail

OCR for page 36
PERSONAL COMPUTER NETWORKS 39 has become a filing and database problem. We now receive so many important messages we can't afford to throw away the 10 megabytes that can be quickly consumed by our electronic corre- spondence. In the future, electronic mail will involve even more data-intensive mocles- for example, voice integration in the mes- saging system, eliminating the telephone tag but maintaining voice, and eventually forms-base messaging in which the con- tents of the message is machine processable. Where does software development fit into the trench toward lo- cal networking of personal computers? I would put software into five broad categories in terms of its relation to networking. In some of these categories actual software floes not yet exist. The first category I would call "unnetworked" software, and there is little of it left. Unnetworked software runs only on a personal computer and cannot in any way be networked, either because there is no transparent networking available or because the soft- ware implementors have not used the standard operating system available. Very few of these software packages would interest ~ organizations. The second category can be called transparently networked software. This is software that uses the operating system cleanly, and networking facilities that have been developed for that oper- ating system can be used transparently by preexisting software. Most of the software that is available today, inclucling 1-2-3, can be and is transparently networked. The third! category I refer to as network-delivered software. The use of the network to deliver the software is a substitute for the floppy disks that most of us have come to think of as a delivery mechanism for software. Examples are Visicalc and Visiword that can now be distributed without diskettes through 3Com?s local network. These facilities can be bought for a large group of users. Supply users with the appropriate number of manuals, and they can get that software over the network much more quickly than they can off a floppy disk. And, they clo not have to worry about storing the floppy disk. Those of us who were involved in the struggle to eliminate punch carcis during the 1970s are now in the process of eliminating floppy disks. The fourth category of software I call multiaccess (or multiuser) networked software. The best example of this wouic! be a dBase II personal computer database package that would allow users on multiple PCs to be concurrently accessing ant! updating a shared

OCR for page 36
40 MANAGING MICROCOMPUTERS database through the network. This is not the same as transpar- ent networking using the current dBase, which is a single-user system that does not provide for shared access to the same data- base. With such a system, there is the danger of multiple accesses damaging each other. Thus, the obvious next step is making available multiaccess database software using local networks to bring a number of PC users to the same database for concurrent access and update. The fifth and final category ~ call network-integrated software. Just as in 1-2-3, which integrates database, word processing, and graphics into a uniform user interface, ant! just as in Vision, where the entire user interface has been integrated along with a variety of applications, we can think of integrating networking into other applications for volume management. The movement of data from one place to another is a part of the natural use of integrated applications. As far as ~ know, only the first three software categories now exist: unnetworked, transparently networked, and network-deliv- ered. But the other two are coming. In fact, the objective of the fifth category of software is to eliminate itself. In other words, when networking software exists, we won't have to talk about it as a separate category because it will be lost in the integration and become part of the applications that users actually need. No one actually needs networking, they need the applications it makes possible. These applications will be the focus of the coming decade.