ON OPTIMIZING SUBSYSTEM ADMINISTRATIVE INTERFACES IN TELEMETERING AND SUPERVISORY CONTROL SYSTEMS
By: F. F. McClatchie
REFERENCE: McClatchie, F. F. : ON OPTIMIZING SUBSYSTEM ADMINISTRATIVE INTERFACES IN TELEMETERING AND SUPERVISORY CONTROL SYSTEMS, Electronics Development Corporation, Newport Beach Calif. formerly Pacific Telephone and Telegraph Co., Los Angeles, Calif. Rec’d 6/28/67; revised 1/29/68. Paper 68TP9-COM, approved by the IEEE Telemetering Committee for publication after unsponsored presentation at the 1967 National Telemetering Conference, San Francisco, Calif. IEEE TRANS. ON COMMUNICATIONS TECHNOLOGY, 16-3, June 1968, pp. 475-478.
ABSTRACT: Complex telemetering, control, and supervisory systems ordinarily comprize a multitude of subsystems. Frequently, the entire system is controlled and/or maintained by two or more administrative entities. Thus, the interfaces between the various subsystems may involve administrative as well as electrical and mechanical considerations. It is evident that optimum functioning of the system requires that all subsystems must be compatible with and even complement with one another.
It is proposed that the key concept involved is that the administrative interfaces should be located at electrical or mechanical interfaces that can be defined by the least number of easily measured non-critical parameters.
Introduction
MOST ENGINEERS agree that system engineering is necessary to make any complex information transmission system viable. In relatively few situations, such as space probe information retrieval, this concept is rigorously applied and the results show it. However, the average telemetering or supervisory control system is more likely to be composed of standard units made by various manufactures, which are then assembled in the hope that the system will measure up to the composite specifications of the individual components of the system. In practice, however, the performance is sometimes so much worse than expected that excessive “debugging time” is spent trying to identify the offending portion of the system. Rather frequently, it develops that all the parts meet specification individually, but the system still fails to perform to overall specifications. Furthermore, while the system may work satisfactorily for a whole after initial acceptance test, it may later require excessive maintenance effort to keep it within specifications. While these types of difficulties can be induced by incomplete mission analysis or inappropriate system synthesis, in this paper the author will limit the investigation to the influence of multiple administrative entities, the validity of subsystem interface conditions, and the relationship between the two.
Types of Interfaces
Two basic types of interfaces between subsystems are considered to exist. One is the electrical and / or mechanical interface between subsystems, and the other is the administrative interface between adjacent administrative entities. All interfaces comprise electrical and / or mechanical parameters, but some of the interfaces may also be the site for administrative jurisdictional boundaries.
In either case, an interface is considered to be” a common boundary between two or more devices or items of equipment”. [1] Thus, the interface is not a device or common boundary between two or more devices or components but is dimensionless and can only be described in terms of its location and the physical, electrical, and administrative conditions that must be met on each side of the interface.
Therefore, this paper presumes the following conditions:
1) that the system is deemed necessary;
2) that the system is technically possible;
3) that all electrical and mechanical interfaces are technically coordinated;
4) that the system will be owned and / or maintained by two or more separate administrative entities;
5) that these administrative entities must be sufficiently flexible to place the administrative interfaces at the optimum locations, without prejudice from non technical consideration.
Multiple Administrative Jurisdictions
a Benefit or Curse?
One way to solve the problems engendered by administrative interfaces in complex data communication systems is obvious, if not always practical to have only on administrative entity.
While this solution works well in the small or less complex systems, it is rarely a practical solution in the large scale complex systems under consideration in this paper. In these more complex systems, a number of basically differing disciplines operate. Thus, since its frequently a great economic and operational advantage to obtain the services of experts in their respective disciplines, the problem of multiple administrative jurisdictions and thus administrative interfaces enters the picture.
So it seems that multiple jurisdictions are not in themselves a source of difficulty, but may develop into a benefit or a curse by virtue of other factors. Thus, the specific objective of this paper is to explore the impact on system viability created by the interdependence between the electromechanical requirements of an interface on the one hand and the requirement of the administrative interface on the other.
The Key Questions
Given that a particular system will operate under multiple administrative entities, the key questions then are as follows.
1) At just which particular location in the system shall the administrative interfaces be placed?
2) What factors influence the choice?
3) Can these problems yield, at least in part, to a general solution?
4) What general parameters are the measure of the optimum interface?
5) Will existing interfaces suffice as locations for administrative interface or will additional equipment be required to drive viable administrative interfaces?
Administrative Interface Criteria
A number of criteria appear to exert a major influence on a viability of an administrative interface. Since all of these criteria are closely interrelated and the relative importance of each one will vary according to circumstance, they are listed in random sequence.
A. The Number of Parameters Requiring Control
One of the criteria exerting a major influence on successful system performance is the sheer number of electrical or mechanical parameters requiring control at an administrative interface. In general, the larger the number of parameters requiring control, the greater the administrative problem. This general rule is valid only as long as the other criteria relating to each parameter are roughly equivalent. Thus, a single parameter delineating the signal will be preferable to a signal requiring control of a multitude of parameters unless that signal parameter is so critical in one of the other criteria as to seriously threaten system viability.
An example of this principle exists at the interface of subsystems in which binary signals of the dc on-off type are converted to some form of modulated ac waveform. On the “dc side” of the modulator, the signal can be defined with a single parameter, i.e., the circuit is open or closed or the voltage is on or off, etc. On the “ac side” of the modulator, the signal may require the control of a number of parameters including bandwidth, envelope delay, signals-to-noise ratio, level control, etc. Thus, it would seem that in this example an administrative interface located on the “dc side” would be easier to administer than if it were located on the “ac side” of the modulator.
Conversely, a single analog parameter (such as a variable dc voltage) may be more difficult to exchange at an administrative interface than a pulse-coded transform of it, even though more parameters must be controlled to pass the pulses. This is because it is easier to establish that the correct code combination was exchanged than to agree on the exact value of an analog quantity.
B. Parameter Criticality
This refers to the relative sensitivity a given signal structure exhibits to variations in the parameters that define its environment. If the signal at a given interface is critically sensitive to minute variations in the value of some parameter defining that interface, and if this parameter tends to be hard to control or maintain, then this interface location would be a poor choice indeed for an administrative interface. If for some reason an administrative interface must exist at such a location, additional transducing devices should be introduced in one or both sides of the interface to translate the signal into a form inherently less sensitive to variations in parameter value.
An example of this might be an analog signal whose absolute value is to be transmitted. If the signal is first quantized into a suitable number of increments and coded prior to reaching the administrative interface, the probability of suitable system performance could be greatly enhanced.
Some interfaces can impair system operation with very small deviations from optimum parameter value, and hence can induce errors and system failure, while other interfaces can sustain substantial deviations from the norm without affecting the system operation at all. Thus all other things being equal, the less critical the value of each parameter at the administrative interface, the greater the system viability.
C. Parameter Measurability
This criterion demands that each significant variable parameter be easily measurable with test equipment available to both administrative entities. It also requires that a clear statement exist with which both administrative entities agree and which defines these limits so that when they are met, proper functioning at the interface will be ensured.
D. Interaction Between Parameters
Interaction between the significant variable parameters is intrinsic to some types of electrical interfaces. An example of this criterion would be where a change in voltage at the interface induces a phase shift as well (differential phase response).
While this can be considered a disadvantage at any interface, the maintenance difficulties engendered by this interaction are multiplied when this point is also an administrative interface. When the degree of interaction between parameters is influenced by trans-interface conditions, criteria D and E can combine to create a most unfortunate combination of circumstances.
E. Interdependence Between the Two Sides of an Interface
Every interface requires electrical and mechanical coordination, thus each side is dependent on the presence of known conditions on the other side. The point being made here is that in some types of electrical interfaces, the parameters are particularly sensitive to cross-boundary condition and thus are poor choices as locations for an administrative interface. An example of such a critical interface could be the junction of a transmission line and a negative impedance amplifier where the gain-frequency response of the amplifier is a function of the driving point impedance of the line. The system’s sensitivity to this interface criterion is dependent not only on the intrinsic interaction between the two sides of an interface, but also on the sensitivity of the transmission signal to the influence of these interactions. Thus, an interface that exhibits considerable interaction may be satisfactory if the signal to be sent is somehow insensitive to this interaction at the point in the system.
F. Accessibility
It is evident that both administrative entities must have reasonable access to the administrative interface. This problem can take two forms: “geographic” and “political.” Obviously, a snowed-in mountaintop or a dangerous environment provides a poor location for an administrative interface, but even an otherwise convenient physical location may be many miles from the nearest source of maintenance skills for one of the two administrative entities. Political difficulties can include restricted entry such as secret clearance requirements to gain access to the interface location, etc.
G. Interface Standards
The very existence of an industry-wide accepted standard for a given type of interface devolves some benefit beyond the purely technical probability that the standard interface would also be the optimum interface in a given application. Once a standard is widely recognized and endorsed, suitable test instruments and testing procedures tend to become codified with the result that the several administrative entities can more readily agree whether a given interface is in fact satisfying all operational criteria. A good case in point is the Electronics Industry Association (EIA) Standard RS-232-B which has done much for standardizing the interface between data processing terminal equipment and data communication equipment. A similar standard for telemetering and supervisory control systems could well do as much good as the data processing standard has already done in that field.
H. Number of Subsystems
While in theory a system may be more reliable as the number of subsystems and total components is reduced, this does not necessarily turn out that way in a practical system, particularly where administrative interfaces are required. Thus, a system may operate at a higher level of effectiveness and with less downtime even when extra subsystems are included for the sole purpose of making the administrative interface more manageable. Obviously, this too can be overdone. Somewhere between the two extremes lies an optimum balance that would yield the greatest system viability.
Commentary
While the foregoing list of criteria makes no claim to contain all of the possible technical considerations that may influence the choice of location for the administrative interface, it is useful as a check-off list when evaluating such a choice.
In the practical evaluation of a proposed administrative interface in a large-scale system, some interface plans can be summarily scrapped, some other are worthy of further evaluation, while rarely does one stand out clearly as the best plan. Thus, in most cases a decision must be made between several possible interface plans. Usually, these decisions must be made at the time the least information for their solution is available. That is to say, very early in the system planning stages, long before the exact nature of the signal at all interfaces is precisely known. This should be a familiar dilemma to most systems engineers. Yet, these interface decisions could well become the ultimate controlling factors, sometimes even determining whether the system “flies or flops.” Thus it is well to build a certain flexibility into these points in the early stages of planning by allowing for extra equipment as may later become evident, For instance, a systems may go to completion and final testing or even into service before it is discovered that a particular interface is administratively unmanageable. If this possibility had been postulated in the planning phases, and an alternate retrofit plan proposed, then minimum time would be lost and certainly less money expended than if the ultimate solution is left to the field” to correct. Especially in the case of the administrative interface point since at least two “fields” are then involved, which may not see eye-to-eye on the problem by this time. At the very least, there will be the problem of “who is going to fix it” as well as “how are they to fix it” if prior retrofit plans have not been formulated.
Summary
The administrative interface has been postulated as an entity deserving special considerations when planning large-scale systems. A number of criteria that can influence the operational viability of a system have been suggested, based on the manner in which they influence the conduct of an administrative interface as well as the way the presence of an administrative interface modified the effectiveness of the electrical/mechanical interface
Each of the criteria must be considered in the light of the special nature of the individual system under consideration and weighed accordingly. There is danger in assuming that since these interface conditions seem pedestrian and obvious that all participants will see them in the same way. That this is often not so is borne out by the “postmortems” that occur when a system fails to function properly because of an interface difficulty.
The EIA Standard RS-232-B has standardized the interface between data processing terminal equipment and data communication equipment in the United States while the CCITT Recommendation V.24 has extended basically the same standards internationally. The Inter-Range Instrumentation Group (IRIG) has established some standards in the telemetering field, but the scope of this standard does not extend to the bulk of the telemetering and supervisory-control systems operation in the United States today.
Much work still remains to be done to establish standards for the bulk of the telemetering and supervisory-control applications. Establishment of these standards will be difficult because of the very wide diversity of applications and data transmission techniques. Nevertheless, the benefits that would accrue to the users, vendors of equipment, and the communications industry in general should at least match those already derived from the standards established in the data processing field and thus will be well worth the effort.
References
[1] J.F. Holmes, Communications Dictionary.
[2] “Signaling speeds for data transmission,” Electronics Industry Association Standard RS 269.
[3] “Interface between data processing terminal equipment and data communications equipment,” Electronics Industry Association Standard RS-232-B.
[4] “Signal quality at interface between data processing terminal equipment and synchronous data communications equipment for serial data transmission,” Electronics Industry Association Standard RS 334.
[5] CCITT Recommendation V.24 interface specification.
[6] MIL-STD-188B interface specification.
[7] Telecommunications Working Group of the Inter-Range Instrumentation Group (IRIG) Standard Time Formats Document 104-60
F.F. McClatchie (M’54) was born in Stuttgart, Germany, in 1923. During World War II, he served in the United States Navy at the Naval Research Laboratory in communications security, electronic countermeasures, and with the Combined Research Group. From 1954 to 1967 he was with the Pacific Telephone Company in charge of engineering of large-scale switching and transmission systems for voice, television, data, and telemetering purpose. While on loan to American Telephone and Telegraph company, he participated as Engineering Consultant in a nationwide survey on interface problems in telemetering and supervisory control systems. He is now Vice President of engineering for Electronics Development Corporation, Newport Beach, Calif. He has organized several technical symposia on a national scale as well as being the author of a number of technical papers on wide-band microwave transmission systems. He has also taught courses on power systems protective relaying and transmission parameters of electronic switching systems.
Mr. McClatchie has served as Chairman of the Electronics Division of the AIEE Los angels Section and on the Joint Council of Engineering Societies of Los Angeles.
Reprinted from IEEE TRANSACTIONS ON COMMUNICATIONS TECHNOLOGY
Vol. COM-16, No.3, June 1968
pp. 475-478
Copyright 1968, and reprinted by permission of the copyright owner
printed in the U.S.A.
