Abstract

The Bureau of Economic Geology has successfully integrated DEC minicomputers, Apple Macintosh and IBM-compatible PC's, IBM and CDC mainframes, and a CRAY supercomputer in a flexible network that allows geologists as well as support staff to use a wide variety of software and output devices.

Introduction

The Bureau of Economic Geology (BEG) is a research entity at The University of Texas at Austin (UT) that also serves as the state geological survey. It is located at Balcones Research Center, which is 10 mi (16 km) north of the main UT campus. In 1987 the Computer Resources section at the BEG supported almost 200 users with a single VAX 11/780 (running the VMS operating system) and a few IBM-compatible personal-computer-based workstations (PC's) (Figure 1). Users also had access to DEC, IBM, and CRAY mainframes throughout the UT campus via a Micom system. However, many Bureau scientists began to buy home computers, in particular the Apple Macintosh. On the Macintosh, scientists were able for the first time to easily generate high-quality geologic graphics such as stratigraphic columns and facies diagrams. The BEG also purchased a Macintosh SE and a Laserwriter Plus to allow these scientists to generate laser-quality output and to allow other staff members to experiment with the Macintosh.

The impetus behind the Macintosh proliferation that ensued came when Bureau administrators saw what the Macintosh could do for them. This was critical to the successful planning and implementation of this sophisticated networked system. No matter how many lower-level staff use PC's or Macintoshes, without high-level managerial interest and support it is nearly impossible to acquire the additional components necessary for a completely integrated networked system.

As the BEG rapidly acquired more Macintoshes for administrators, scientists, graphic artists, and publications staff, the Computer Resources staff quickly began researching and planning networking strategies. Our goals included (1) planning for a large network of Macintosh computers, (2) sharing Macintosh files and peripherals, (3) facilitating software management and data security by using the existing VAX as a file server for the Macintoshes, and (4) improving networking to outside computers by connecting the BEG's local baseband Ethernet to the UT campus broadband Ethernet. With the acquisition of 25 Macintoshes, 8 Laser writers, 2 Apple Scanners, and a color Postscript printer, all within 1 year, many of our "long-term" Postscript printer, all within 1 year, many of our "long-term" plans became short-term realities. plans became short-term realities. By using innovative applications of local- and wide-area networks, we facilitated user-friendly data transfer between vastly different operating systems. We cost effectively enhanced the computer capabilities of the BEG, and we implemented the network system with a minimum of user training and with no disruption of research or support activities. We accomplished these goals within 1 year.

PLANNING PLANNING Planning for network implementation requires two important Planning for network implementation requires two important preliminary steps: (1) deciding on a network topology (wiring preliminary steps:

  1. deciding on a network topology (wiring configuration) for all phases of development, and

  2. using a "systems" approach to network development of short-term, mid-range, and long-term goals.

To begin, it is necessary to know and understand the different types of network topologies, the limitations of each topology, and the appropriate environment for each topology. A network topology should be selected to fit current and future facilities, levels of staffing, and computing needs. Failure to understand network topology, ignoring its importance, or "forcing" an inappropriate topology into the workplace will lead to certain disaster. Such disasters might include complete rewiring of buildings, inability to perform anticipated network tasks, and poor network performance. These types of problems can be poor network performance. These types of problems can be avoided by planning the network topology early.

Possible network topologies include (1) daisy chain, (2) trunk, Possible network topologies include (1) daisy chain, (2) trunk, (3) passive star, and (4) active star (Figure 2). The daisy chain is suitable for locations that house all of the devices in one room. Because we anticipated rapid Macintosh proliferation throughout the building, this alternative was not viable.

P. 189

This content is only available via PDF.
You can access this article if you purchase or spend a download.