Printer Friendly

LANs: manufacturing's prodigal child.

LANs: Manufacturing's prodigal child

The future factory will be a giant network with a lot of smaller systems connected to it--programmable controllers, machine tools, materials-handling equipment, even other computers. If you are striving for the factory of the future, you must think factory communications --machine tools talking to process controllers, and so on. Even when considering just-in-time manufacturing, data communications is critical.

Unfortunately, no vendor fully addresses the problems of integrating factory systems. Along with islands of automation, there are islands of information. For example, current CAD/CAM packages can't create bill-of-material files that can be used as input to MRP or inventory control systems.

Integration is needed, both at the supervisory and shop-floor levels. Manufacturers will have to hook-up their computer systems to inventory, control, and database management systems. All this eventually will make up the environment called CIM.

Three deep

I don't believe factories will be installing a single omnibus data communications system. A more sensible approach will be to have three communications levels--all accomplished by local area networks (LANs).

Level I--This will be a wide-area LAN fully integrated with all computer-based devices and systems within the corporate and manufacturing environment.

Level IIA--This will consist of an intrafacility or campus-wide LAN bridging factory and business environments.

Level IIB--This will be an interdepartmental LAN linking specific manufacturing cells and operations.

Level III--This can be a departmental or manufacturing cell-level LAN for linking factory floor machine tools, materials-handling equipment, controllers, and other computer-based devices.

Corporate-wide networks

A Level I LAN typically has roots in the management information system (MIS) department. It links different facilities together, i.e., office, plant floor, warehouse. Typically, Level I LAN communication is between computer and computer, or gateway and gateway. Major communications and computer vendors, such as AT&T and IBM, will actively participate in Level I LANs because they already have those kinds of systems installed, including IBM's System Network Architecture (SNA), DEC-net, and Data General's Zodiac.

Level I networks use fairly low speed, modem-type connections operating at 4800 bps (the digital communication line from AT&T can communicate at 56 kbps). Sophisticated users are communicating at faster rates by using their network architecture on a dynamic basis through a T-1 link.

High-speed communications at Level I, however, isn't necessary at present. Level I usually is moving large chunks of information from one plant or facility to another. The facilities--information islands --don't need to be in constant touch with each other. Once or twice a day bulk data will be moved between facilities for further processing.

Consequently, a Level I LAN should provide the following features for manufacturing: timely, reliable message delivery; reliable message acknowledgement; and a priority system for traffic during peak periods. There's just one problem: no coherent network exists today that provides these types of features. Instead, you must buy a network from Network System Corp or other vendors that supply T-1 links, such as SBS.

Another aspect of Level I LANs is the notion of a factory service network (sometimes called a remote service network). Design of these networks will especially impact smaller factories that can't afford a large MIS function, but instead have one or a few minicomputers sitting around.

These factories will want to connect their minis to service bureaus, such as McAuto or GEISCO. Then, once a month or so, these factories can connect their small CAD/CAM system, for example, to the service bureau's larger computer system to do CAE or other higher-level operations.

Intrafacility networks

Level II LANs integrate other networks and intelligent devices within a facility. These bridges between the MIS department and factory level manufacturing systems usually are a generalized form, such as Ethernet or an IBM network. Level IIB LANs are interdepartmental factory system networks; usually, machine tasks fit here.

Level II LANs are general purpose. They act as plant information conduits and ideally provide the ability to attach equipment at any location without special procedures or undue complexity. Further, they overlap other operating LANs. Consequently, some information exchange, such as data collection and data processing, is necessary.

Level II LANs should have the following characteristics: high reliability, high message integrity, guaranteed acknowledgement of each message, reasonable handling of peak loads, wiring suitable to the plant floor, and insensitivity to electronic noise.

Floor-level networks

Level III LANs are the most interesting to follow because that's where current development and implementation is happening. Gould, Allen-Bradley, Modcomp, GE, and others are keen on integrating their equipment with their own and the competition's equipment. It's a necessity for devices to talk to one another within a manufacturing cell.

Communications requirements for these LANs are application dependent; however, low cost is most important because today's shop floor is fluid, and companies won't invest in an expensive system if only to tear it down every two years.

Stumbling standards

Vendors can no longer wait for standards committees to develop a solution to LAN protocols, and hardware and software interfaces. Ironically, all these committees have done is come up with their own proprietary LANs.

Data communication standards are a lot more important in the factory than in the office. Obviously, office equipment needs to be able to communicate, but lack of standards in the factory is a real obstacle to automation. Users are forced to buy piecemeal solutions. Thus, they tend to solve limited problems, one at a time. Integration at a later date is difficult, inefficient, sometimes impossible, and always costly.

Therefore, users are tentative about making commitments to automation when they don't know what will be standard two or three years down the line. This is why I believe GM's MAP program is such a flagwaver (T&P, "A MAP to the future?,' Nov '84, pg 86). Whether it's the best solution or not is irrelevant; at least it's something for users to hang their hat on and for vendors to build towards.

The critical standards involve network management; that is, error correction, movement of information to devices, how devices get information, who talks first, and who has control of the session. So far, network architecture that manages information transmission between different devices is missing in LANs from most vendors. While it's easy to put an RS-232 interface on devices for exchanging information, it's extremely difficult to place network architecture on top of that interface to actually manage information flow and correct it automatically.

For Level I LANs, we are seeing some standards development revolving around the interconnect architecture, which is being pushed by a number of organizations, particularly the International Standards Organization (ISO). They've been actively pursuing Open Systems Interconnect (OSI) for years.

OSI, however, is basically the reference model that will enable LAN vendors to segment layers of communications and function within the same segment. Thus, for example, DECnet will be able to talk to an SNA net, and the SNA net will be able to talk to other networks.

OSI is at least a year and a half away from full specification. What's available today are LANs from vendors who say they're going to be OSI. In reality, what you're buying are piecemeal solutions because the upper level LAN protocols are not specified.

Question: If you buy a proprietary LAN today (SNA, DECnet, or Zodiac), will this limit you when OSI finally is standardized?

Answer: Probably not. Why? Because OSI and the LAN vendors are looking at defacto standards and actually agreeing upon them. They are, in effect, considering modifying the ideal solution, in an academic sense, and changing it to what the real world is doing now.


COPYRIGHT 1985 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1985 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:local area networks
Author:McGarry, Susan L.
Publication:Tooling & Production
Date:Jul 1, 1985
Previous Article:Industry report - information technology.
Next Article:Rx for downtime: machine diagnostics.

Related Articles
IBM and AT&T Join Xerox in the Proliferation of Local Networks.
Standards Chart Direction of Local-Area Network Advances.
On PCs and LANs: tying the factory together.
'Ping' is the key to managing 3M's growing network.
What's in a LAN?
Vendor, network thyself.
Longtime local comes back to play Eleanor Roosevelt.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters