Abstract

Introduction

The dependency of modern man upon the computer is getting increasingly irrevocable. Computer technology, both hardware and software, is expanding and improving faster than most of us can keep pace. Much of our work is done by computers, from keeping track of bank accounts, process control in industry to adjusting the fuel mixture in automobile carburetors. Some major industries, such as banking, process such vast volumes of data that their task of information management would be impossible without computers. On university campuses, the realization that numerical tasks can not be completed without computers is deeply rooted. The control of library facilities, maintenance of student record and management of entrance examinations are but a few of the potential computer applications in these environments.

This paper focuses on one particular facet of computer technology - that of computer networking between university environments. The term "computer network" is used to include any hardware/software entity that may serve to link computer systems in an orderly manner so as to establish inter-computer communications. The network only concern the exchange of information between systems, and not the internal functioning of each individual system. This information transfer is performed by means of physical media for systems interconnection.

Design Philosophy

UNINET is intended to provide educators and researchers in the computer field with resource sharing and communication capabilities. Administrators will benefit from the remote data access feature which supports integrated processing activities among various organizations.

The design of the network is guided by three main principles

1. UNINET is to be available first to any state university/ institute and the Ministry of Education and Culture. To promoted the cooperation in computer-oriented education and research, and administrative data processing.
2. Second, it is also to be available to other interested post-secondary educational institution active in the computer field, and to any industrial and government organization in Indonesia which is engaged in computer-related education and research, and special-purpose databases.
3. Third, it should be capable of utilizing the various existing (and planned) physical communication facilities in order to attain the best (i.e. the most economical) network configuration.

Network Architecture

By serving the user community with network facilities and access, UNINET will enable institution to participate trough a network interface. It is a logical network utilizing the services of various physical networks. Figure I shows UNINET's logical architecture. Participating nodes will consist of local-area-networks(LAN's) and single-host systems. This implies that UNINET will not accommodate direct terminal hookups. Instead, nodes are required to possess a certain level of intelligence, which can be implemented through "gateways", if the node is itself a network or, in the case of single-host nodes, through a network interface. Practically this gateways are similar to the network interfaces, in that they both must interface the local communication characteristics to the attributes of the physical network (for example: X.25 ) used by UNINET. They can thus both be based on the same microcomputer, but configured (and programmed) differently.

Figure 1: Logical Network Architecture

A Layered architecture is proposed to achieve systems modularity and improve compatibility. The proliferation of machines and communication protocols have made it necessary to maintain and architecture versatile enough interface properly with various systems. The open systems Interconnection scheme [ISO 79] lends itself nicely into this requirement. Although there my be some disadvantages to be gained far outweigh the problems.

The LAN's will conceivably be a collection of Ethernet systems [MeBo76], high-speed fiber-optics systems such as described for example in [CRKL81], and other LAN architecture.

Institutional Framework

Network Functions

A user interfacing with a computer network may want to perform any combination of the usual network functions, namely:

• Send/receive electronic mail
• Transfer files Submit a batch-job to a remote computer Interactively logon a remote system
• Perform distributed processing
• Access a distributed database

In developed countries, electronic mail is widely use for communication among individuals in a given organization and more recently, among peers belonging to different organizations. In the past few years a number of such mail systems became operational, among others MAILNET [HeLL82] it uses a star topology with dial-up polling during reduced-rate periods (I .e night time). A user can create, send, and receive messages and documents on his/her own local system even if the correspondent is a user of a totally different system. A computer located at the hub of the network performs all of the functions a post-office clerk, viz. Dealing with different envelope formats, sorting of letters, making decisions about how to deliver the letter, and how respond to a variety of errors. In addition to this, MAILNET' s electronic clerk makes the necessary copies of messages sent to multiple recipients.

Implementation Plans

References

• [CRKL81] W.Y. Cheng, S. Ray, R.Kolstad, J. Luhukay, R.Campbell and J.W-S. Liu, "ILLINET = A 32 Mbits/sec Local - Area Network", Proc. National Computer Conf. 1981, pp 209-214.
• [HeLL82] P. Heller, "MALLNET: A Convenient Inter-Campus Electronic Mail Service", ACM-SIGUCC Newsletter 12/4, Winter '82, pp. 12-16.
• [ISO79] International Standards Org., "Reference model for Open System Architecture", ISO/TC97/SC16/N227 revised, 1979.
• [MeBo76] R.M. Metcalfe and D.R. Boggs, "ETHERNET: Distributed Packet Switching for Local Computer Networks", Communications ACM 19/7, July'76, pp. 395 -404.

EndNote