• Keine Ergebnisse gefunden

COMPUTER-TO-COMPUTER NETWORKS Parallel to the development of terminal-oriented

Im Dokument COMMUNICATIONS SOLUTIONS (Seite 33-38)

A Brief Historical Perspective of Computer Communications

COMPUTER-TO-COMPUTER NETWORKS Parallel to the development of terminal-oriented

sys-tems, efforts were under way to allow computers to communicate directly with other computers in real time. The first step was, of course, to place two identical computers in the same building and to connect a cable between them. (Many of the com-puters being built today can be regarded as sophis-ticated computer networks in themselves.) To assist in this difficult task, devices very much like front ends were developed to handle the communications functions and other chores needed. Naturally, the communication lines became longer, necessitating communication hardware at the ends of the line.

A result of this approach is star-like networks (Figure 8) with a store-and-forward central switch. A signi-ficant network in this category is the AUTODIN System.IS AUTODIN was built and is maintained and managed by Western Union for the U.S.

Government.

An extension of this type is the ring computer network in which a front-end type device (often called a net-work interface processor) connects the netnet-work lines and the computers. Data for a computer is addressed to that computer and sequentially sent, link by link, in a circular fashion. At each step around the circuit, the data is interrogated by the interface processor, and when it finally reaches the interface processor connected to the destination computer, it is removed from the ring. Naturally, if a network like this is not planned very well, data may eventually circulate

for-(a)

(b)

(c)

Figure 8. Evolutiun of cumputer-to-computer networks

ever. Thus, control devices to remove data which is

"too old" from the network must be placed in the network. In addition, as such a network grows, its reliability can become very low because all elements along the ring must operate for the network to operate.

Therefore, additional lines for redundancy and more flexible routing techniques must be added for effective operation.

A more ambitious type of system is called ARP ANET.16 17 The concept of this system was to provide high flexibility by allowing any kind of inter-connections and adaptive routing of information. In late 1969, the first four elements were installed on the West Coast. The network grew to about a 25 node sys-tem in 1971, to about a 40 node system in 1973, and today has well over 50 nodes (Figure 9). This network

© 1979 DATAPRO RESEARCH CORPORATION, DELRAN, NJ 08075 USA REPRODUCTION PROHIBITED

JUNE 1979

HAWAII AMES

A Brief Historical Perspective of Computer Communications

December 1969 December 1970

(a) (b)

September 1971 August 1972

(e) (d)

September 1974

~\ v

(e) Figure 9. Geographical expansion of the A RP A network

JUNE 1979 © 1979 DATAPRO RESEARCH CORPORATION. DELRAN. NJ 08075 USA REPRODUCTiON PROHiBiTED

tsaslc l;oncepts

A Brief Historical Perspective of Computer Communications is one of the first major applications of the new

tech-nology called "packet switching" in which data is broken up into blocks that are separately addressed and then allowed to make their way independently through the network from origin to destination. This type of network must handle the problem of control-ling flows using a "distributed" control scheme.

The ARPANET significantly differs from the cen-tralized system approach. In a cencen-tralized system such as NASDAQ, nearly all the controls reside in the central computer. If it cannot handle the flow, the computer will slow down the concentrators and do whatever else is necessary to prevent additional calls from being sent. In a distributed network, very sophis-ticated techniques of flow control and routing adap-tion in case of a line or node failure had to be developed. Packet switching is now viewed as a major addition to the technology of computer networking, and has already been applied to radio communica-tions.lx A number of other networks are now being built or designed based on the packet switching tech-nology of the ARPANET, not the least of which is IBM's System Network Architecture (SNA), and the future of the field appears quite bright.

PROPHECY

Clearly an important part of the computer com-munications revolution has been the proposal and development of an incredible array of digital services.

This includes new technical offerings and tariff struc-tures by the common carriers dominated bv AT&TI9 and Western Union.20 A further develo-pment of crucial interest to the computer industry is the growth of the specialized common carriers including MCI, DA TRAN, and a large number of regional carriers such as Western Tele-Communications. The picture is further enhanced by the addition of value added networks and satellite communication. These topics are ~overed in more detail throughout the body of this service.

Our mandate does not include prophecy-for evil or for good. But after all the only reason for knowing

"How we got there," is so we can extrapolate to

"Where we are going." Some things are certain. As Fano says "The 'Marriage' of computers and com-munication has been celebrated and consummated.

But now the honeymoon is over, and the two partners are beginning to face the realities of their inter-dependence. "21

Looking into the very near future, networks are planned that tend to combine the distributed network control concepts of ARPANET for computer-to-computer communications with the centralized NASDAQ-like approach for terminal-to-computer and terminal-to-terminal communications. These

net-works are an extension of the multidrop centralized net where now the terminal processor replaces the computer, and the backbone communications is then through a packet-oriented net like ARPANET. An example of this type of net is shown in Figure 10. This particular example is a sample design for a planned FAA Air Traffic Control Network. This network has 21 air traffic control computers at appropriate loca-tions. It has a backbone communication network which is a simple loop like network. Emanating from the nodes of this network is an extensive terminal communications network, which is itself a collection of networks.

(!) DPwIa MID ...alQ11111, OL COMPVTlIII""0 .... 'c"HilGc'''' •••

-• COIC""""'OIVAill-ouT.

'A"'UC COtIlAOL Chll'Q .1UIIIt ... ~ -VOtCIGilADIL_

·----Tlu""" ...

Figure 10 . .4 computer I)/ofled OU!pllt/()r a data communicatiol1 fU:'!II'ork ol5(){) iocation.\'

The growth of computer communication networks has clearly left the linear part of its presumed exponential growth. In-house systems or inter-corporation facilities abound not only on paper but in actual implementation. In addition many more facilities are on the horizon. For example:

• In Canada, the Datapac Network is a nation-wide, packet switched, shared, data network, which has been designed to become the basic Canadian network for data communications. There are four network nodes: Toronto, Montreal, Ottawa, and Calgary. These four nodes, or networks switching centers, \vill initially serve the entire country. By 1980, at least fourteen Canadian cities will have network nodes. After 1980, the network will con-tinue to expand to meet Canada's data require-ments.

• Also in Canada plans are being developed for CANUNET, Canadian Universities Computer Net-work, a packet switched network sponsored by the Ministry of Communications to link some 20 universities.22

• An international effort is planned by the Organiza-tion for Economic CooperaOrganiza-tion and Development.

The result is to be a European data

communica-© 1979 DATAPRO RESEARCH CORPORATION, DELRAN, NJ 08075 USA REPRODUCTION PROHIBITED

JUNE 1979

A Brief Historical Perspective of Computer Communications tion network between certain universities and

re-search centers. This network, which will work on the "packet switching" principle, in reminiscent of the ARPA network. Secondary networks can be connected to nodal centers. Nodal centers will exist in Italy, France, Switzerland, the United Kingdom, and within the OECD administration.

Norway, Sweden, Portugal and Yugoslavia have also joined the project.23

Beyond extrapolation we indeed enter the realm of prophecy. We can only list a few achievements we all know are here or on immediate horizon, make an obvious observation, and relate a personal expenence.

First the list. The following developments are here:

• Minicomputers

• Programmable calculators

• Hand calculators

• Microprocessors

• Hand held radio transmitters

• Cable TV system for data transmission

Second, the obvious observation. Even without look-ing into the far future of hand held minicomputers on a chip or optical fibers, it is clear that computer networks will soon look nothing like they look now.

Mobile users with hand held terminals dialing into vast networks of minicomputers and maxicomputers, with little difference between front ends and processors, is clearly possible.

Finally, a personal experience; as usual, one of us did his Christmas shopping on Christmas eve. He was at the counter at Macy's trying all the calculators, using one calculator to calculate the cost per feature on all the other calculators at the latest bargain price of overstocked Japanese calculators with Italian names. A woman standing next to him, silent for many minutes, finally got up the courage to ask the sales-man what memory was used for on a calculator. He tried to explain several times and failed. Finally, he showed her how it was used to store an intermediate answer. A glow of discovery appeared on her face.

For the first time after years of propaganda, adver-tising, and intimidation about computer memory banks she understood what memory was. A new American became intimate with the computer. This element of citizen acceptance of the computer when combined with the technical elements make a new revolution both inevitable and unpredictable.

Many others are, of course, actively speculating on the effect of the computer communications revolution on society. Some of this speculation is didactic. Says Peter Goldmark,24 "What I propose is that the advances of telecommunications technology-satellites, cable TV, broadband circuits and similar devices-make it possible to attract future generations into the smaller towns of America beyond the com-muting dependency range of the big city and suburbs and thus cut down on the excessive use of power."

Some of the speculation is more ruminative. Says Paul Baran,25 "The key man in the new power elite will. be the one who can best program a computer, that is, the person who makes the best use of the available information and the computer's skills in formulating a problem. In a world where knowledge is power, and where communications mean access to power, he who can most effectively utilize this access will be in the driver's seat. Some persons (primarily computer programmers) claim that the richest man in the world in the year 2000 will be a computer pro-grammer. This may sound outlandish, but few really good programmers laugh when they consider this assertion."

But the best appraisal is by Steward Brand,26 humanist author of "The Whole Earth Catalog." In his essay,

"Fanatic Life and Symbolic Death Among the Com-puter Bums," he sums it all up, "Ready or not, com-puters are coming to the people."

REFERENCES

I Asimov, 1., "The Ugly Little Boy," Nine Tomorrows, Doubleday &

Co., Inc., 1959.

2Toynbee, A. J., Change and Hahit: The Challenge of Our Time, Oxford University Press, 1966.

-'Thorndike, L., A History of Magic and E'(perimental Science:

Volume III and IV, Fourteenth and Fifteenth Century, Columbia University Press, Second Printing, 1953.

4Boorstin, Daniel L The Americans: The Democratic Experience, Random House, 1973.

5McLuhan, M. and Q. Fiore, War and Peace in the Glohal Village, McGraw-HilL 1968.

hFrank, H., 1. T. Frisch and R. Van Slyke, "Testing the NASDAQ System-Traffic and Response Time," Proceedings of the Symposium on Computer-Communications Networks and Teletraffic, Poly-technic Institute of Brooklyn, 1972, pp. 577-586.

7Sac kman H . ., COnlj'Juter S.rstern Science and Evolving Society., John Wiley, 1966.

XEnticknap, R. G. and E. F. Schuster, "SAGE Data System Con-siderations," Transactions of the American Institute of Electrical Engineers, January, 1959, pp. 824-832.

9Taplin, J., A History of the Remote Access Computer Industrr in the United States 1964-1972, Master's Thesis, University of Pennsyl-vania, August II, 1972.

IOKnight, J. R., "A Case Study: Airline Reservations Systems,"

Proceedings of the IEEE, Vol. 60, No. II, November, 1972. pp. 1423-1430.

II Hargraves, Jr., Robert F., "The Dartmouth Time Sharing Network."

Computer Communication Networks, N. Abramson and F. F. Kuo (Editors), Prentice Hall, 1974.

12Gaines, G. and J. Taplin. Time Sharing Todar, Vol. 3. Nos. I and 2.

May, 1972.

JUNE 1979 © 1979 DATAPRO RESEARCH CORPORATION, DELRAN, NJ 08075 USA - - - -

-CS10-120-110 Basic Concepts

A Brief Historical Perspective of Computer Communications

11Mauceri, L. L "Control of an Expanding Network An Opera-tional Nightmare," Netll'orks, Vol. 4, No.4, 1974, pp. 2'67-297.

14Bcere, M. P. and N. C Sullivan, "TYMNET A Serendipitous Evolution," IEEE Transactions on Communication.\', Vol. COM-20.

No.3, June, 1972, pp. 511-515.

I'iCOX, J. E., "Western Union Digital Services," Proceedings or Ihe IEEE, Vol. 60, No. I L November. 1972. pp. 1350-1356.

16Frank, H., I. T. Frisch and W. Chou, "Topological Considerations in the Design of the A RPA Computer Network," S.lCC May. 1970, pp. 5'61-587.

17Roberts, L. G. and B. Wessler. "Computer Network Development to Achieve Resource Sharing," S.lCC 1970, pp. 543-549.

'XAbramson, N .. "The ALOHA S~stem:; Computer Communication Nellt'orks, N. Abramson and F. F. Kuo (Editors). Prentice Hall.

1974.

IYJames, R. J. and P. E. Muench, "AT&T Facilities ano Services:' Proceedings or the IEEE. Vol. 60, \10. I L November, 1972. pp. 1342-1349.

20Gerla, M .. "Moving Bits by Air. Land and Sea." in this session.

21 Fano. R. M .. "On the Social Role of Computer Communications."

Proceedings or the IEEE. Vol. 60. No. II. November. 1972. pp. 1249-1253.

~2deMercado. J .. R. Guindon. J. DaSilva and M. Madoch. 'The

Canadian Universities Computer \letwork: Topological Consioera-tions." Proceedings or the First Internalional COIl/i.'renee on COli/-pliler Communication, S. Winkler (Editor). Octoher. 1972. pp. 220-225.

c.1Larsson, T.. "Data Communication in Sweoen ano Some Aspects of the Situation in Europe." Proccedings or the First Imernatiollal Conrerence on Computer COll1l11unicatiof1.\. S. Winkler (Editor).

October. 1972. pp. 17-25.

24Goldmark. P. C. "A Rural Approach to Saving Energ\." The Nell' York Times, Sunda\. November II. 1973.

c'iBaran. P .. "On the Impact of the New Communications Media Upon Social Values." Law and Contemporarr Prohlems. Vol. 34, No.

2. Spring. 1969. pp. 244-253.

chBrand. S .. /I C"hernetic Frontiers. Random House. 1974.

27Gaines. E. C. '''Specialized Common Carriers Competition ano Alternative:' Telecommunications. Septemher. 1973. pp. 17-26.

2XLuther. W. J., "Conceptual Bases of CYBERNET." COllllJUter Nefll'orks. R. Rustin (Editor). Prentice Hall. 1972.

2YSchwartl, M .. R. R. Boorst~n and R. L Pickholtl. "Terminal-Oriented Computer Networks." Proceedings or the IEEL Vol. 60.

No. II. November. 1972, pp. 1408-1422.

1llWorley. A. R .. "The Datran System." Proceedings orthe 11:'1:·/-. ... Vol.

60. No. I I. November, 1972. pp. 1357-1368.0

© 1979 DATAPRO RESEARCH CORPORATION, DELRAN, NJ 08075 USA REPRODUCTION PROHIBITED

JUNE 1979

Teleprocessing-The Modern Marriage of

Im Dokument COMMUNICATIONS SOLUTIONS (Seite 33-38)