Advertisement

Choosing the Link Product

This is the third and final article In the series on microcomputer to mainframe links. The first article was "Linking Micros to Mainframes" (The Scientist, October 20, 1986, p. 14); the second was "How to Develop Link Networks" (The Scientist, December 15, 1986, p. 14). How do you develop an implementation plan for connecting micro-computers to mainframes? What are the important considerations? How do you select appropriate products? If you have only six personal computers in your organization,

By | January 26, 1987

This is the third and final article In the series on microcomputer to mainframe links. The first article was "Linking Micros to Mainframes" (The Scientist, October 20, 1986, p. 14); the second was "How to Develop Link Networks" (The Scientist, December 15, 1986, p. 14).

How do you develop an implementation plan for connecting micro-computers to mainframes? What are the important considerations? How do you select appropriate products? If you have only six personal computers in your organization, you probably don't need a formal implementation plan, but if you end up with 60 personal computers in six months, you probably do. If you have 8,000 personal computers scattered throughout multiple company divisions, only a plan can avoid utter chaos.

Developing the Plan

To develop an implementation plan you need to know your objectives. How do you want the project to turn out? What are the stages along the way?

First, you are dealing with communications products, hardware and software. Micro-mainframe links, local area networks and file servers are products that require hardware and/or software installed in the personal computer to connect directly or remotely to a host computer. These products incorporate into your system by bringing external data bases and applications to an employee's desk top.

Below are six areas you need to assess in developing and implementing plans to connect personal computers to your mainframe. Some of them (such as education and training) involve added costs and must be assessed carefully. Misreading or ignoring these additional needs can gobble up time and money.

In planning, you must:

  • Treat link products like any other major communications equipment that interfaces with your host. Networks take many forms, and products installed in personal computers must be chosen as carefully as other system networking equipment.
  • Know how link devices connect to the host and the required prerequisites for linking them.
  • Know what diagnostics are available. What does the vendor supply to locate problems and identify system failures?
  • Know how link products work in other systems. What experiences have other users had with them?
  • Know how you will provide training and education for users.
  • Ensure that the users are involved in planning and implementation and kept informed of progress.

Implementing the Connection

A good implementation plan starts with the top level of design. It should begin with an examination of what software is available at the host and the personal computer, and what software still is needed. One good idea is to make transparencies of the configurations for your personal computers and your host computer and place one on top of the other so you can see the situation at a glance.

A good plan is also divided into phases so that individual areas are connected to the mainframe one at a time. The installation should progress in a gradual wave rather than as an abrupt changeover. This process allows you to monitor the installation in steps and note any problems. You can compare your results with your expectations, and your users will gain confidence in the system because they have observed its progress.

In addition, if you find that the drain on your host computer is more than you anticipated, you can temporarily halt any further link installation until the problem is solved. It might be necessary to modify your approach to connecting the personal computers, place restrictions on the size of data file transfers, stipulate certain off-peak periods for large transfers, or increase the host's capabilities.

Micro-mainframe links have costs, both tangible and intangible. In an IBM world, the tangibles include a hardware link card and its cable, and allocation of a 3274/6 terminal controller port. It is difficult to put a price on a link because user and host conditions vary so much, because the price is affected by the amount of support needed during implementation, and because the price increases as the complexity of the linkage increases. It is best to develop a list of link products and their prices.

The intangible costs involve the resources of the host computer. Once linked to the personal computers, the host must have enough processor size to drive the links and to support them with controller and communications resources. These costs include the controller memory storage available, the possible need for additional controllers, the need for sufficient telephone lines to support the links, and the capacity of the system to perform volume data transfers, if this is a requirement.

If you have only one or two links, the additional tangible and intangible costs are minimal. As the system grows, however, costs increase rapidly.

Which Product?

Once you've planned the process, it's time to select the link products. Links are determined by your need to interface with specific applications and to move data. If you know your short and long-term requirements, you can look through the literature, talk with vendors and other users, and choose the link that will do the job most cost-effectively.

You first need to define what you absolutely need and what can be sacrificed. Put these items into a checklist you can use in evaluating all link products on the market. Generally, the more complicated the criteria, the more complex the solution. Compare how well each link meets the criteria you've developed and compare the options each offers. These criteria are a guideline to, but not an absolute measure of, the product's effectiveness.

Choose link products on the basis of how well they meet your future needs as well as your current requirements. For example, if you choose local area networks, are they to be installed only within a department or to a host computer? If the former, are "gateway" products available to link your network to the host in the future?

Data sheets on link products may not misrepresent information, but they don't always tell the whole truth. They tell you what the link is capable of doing, not how it is going to perform for you. In general, the simpler the link, the more you can rely on the data sheets. Still, it is up to you to understand the nuances and operational peculiarities of each link.

Talk to your vendor: he or she can help you decide how to find the product that best meets your needs. Talk to other users, too. And, if necessary, consult an independent organization that can assist you in choosing products and designing networks. Their unbiased view of your situation can provide you with other economical and reliable solutions.

Kopeck is the author of Micro to Mainframe Links (Osborne/ McGraw Hill, 1986) and president of EdgeTech Associates, Suite 602, 582 Market St., San Francisco, CA 94104.

Advertisement

Follow The Scientist

icon-facebook icon-linkedin icon-twitter icon-vimeo icon-youtube
Advertisement

Stay Connected with The Scientist

  • icon-facebook The Scientist Magazine
  • icon-facebook The Scientist Careers
  • icon-facebook Neuroscience Research Techniques
  • icon-facebook Genetic Research Techniques
  • icon-facebook Cell Culture Techniques
  • icon-facebook Microbiology and Immunology
  • icon-facebook Cancer Research and Technology
  • icon-facebook Stem Cell and Regenerative Science
Advertisement
Advertisement
Mettler Toledo
Mettler Toledo
Life Technologies