Relational model-base structures 29
Relational model-base structures:
underpinnings for decision-driven (vs. database-
dependent) management support systems
John W. Sutherland* ― Elizabeth Baker **
*Department of Information Systems
Virginia Commonwealth University
Richmond, Virginia USA
**Department of Economics
Virginia Military Institute
Lexington, Virginia USA
______
ABSTRACT
As things now stand, administrative system designers are likely to see their options as being restricted to one or another of the products now being produced and promoted under the Enterprise Resource Planning banner. This dominance is troublesome because ERP packages are not a universally appropriate technical choice. Rather, because of their dependency on relational database conventions (which, after all, were conceived with routine commercial data processing requirements foremost in mind), ERP-based management systems are not well fit, in either form or function, to answer for anything much beyond the relatively pedestrian problems that punctuate ordinary business firms. They cannot be expected to efficiently or fully accommodate the more demanding analytical and operational challenges confronting the extraordinarily large and complex enterprises that are increasingly characteristic of the contemporary commercial sector, much less the quandaries that regularly arise in the modern military, civil services and public administration arenas. Hence the primary purposes of these pages: To introduce relational model-base structures as an alternative to conventional relational database structures, and thereafter go on to show why/how it is that management systems resting on relational model-base structures should be able to provide administrative authorities with a higher order of technical support than relational database-dependent ERP constructs.
KEYWORDS: Decision system, Enterprise Systems, Management Support Systems
______
1. Introduction
Relational model-base (RMB) structures have their conceptual origin in one of the grounding propositions of the management science movement: the contention that enterprises can be meaningfully comprehended as collections of decision points rather merely as collections of people. This puts the spotlight on the analytical aspects of administration in contrast to the socio-behavioral thrust of mainstream management theory, or the morphological matters that concern organization theorists. Hence the principal employment for a relational model-base structure: to capture and encode, mainly as computer-executable models of certain common kinds, the assemblage of interrelated administrative decision requirements entailed in some enterprise (where, as the term is used here, an enterprise may refer either to an activity undertaken entirely within the confines of a singular organization, or an inter-organizational endeavor of some kind).
The portrait of an enterprise as conveyed by a relational model-base structure would then show which decisions are linked with which others, in what ways, at some particular point in time. This suggests an instructive rationale for relational model-base structures. They offer executives an opportunity to get to know their enterprises in a new way. What they can provide, particularly, is a systematically-drawn portrait of an enterprise’s administrative process, including a detailed description of the distribution of actual administrative authority (as distinct, for example, from the depictions of administrative structures and distributions of nominal decision responsibilities available from organization charts). Equipped with such knowledge, enterprise executives might then institute an equally systematic search for ways in which an administrative process might be “reengineered” to make it more streamlined or operationally elegant.
Making enterprises’ administrative processes more apparent to their principals is not, however, the core rationale for relational model-base structures. Their primary purpose is to provide the technical wherewithal for the development of a new class of administrative artifices, decision-driven —or decision-model driven, more pointedly— management support systems. These would then stand as practical expressions of the management science emphasis on analytical administrative requirements. But they could also stand as alternatives to the relational database-dependent ERP-based designs that reflect the mainstream management information system community’s continuing preoccupation with the informational aspects of administration (business administration, in the main).
It’s as an alternative to ERP products (of which SAP, Baan, Microsoft Dynamics AX and Oracle/Peoplesoft are notable examples) that relational model-base structures are of most immediate practical significance. So, as a prelude to any detailed discussions of the configurational and functional features of decision-driven management support systems, it might be useful to give some quick consideration to how they would differ from their ERP-centered counterparts.
2. Contrasting ERP and RMB centered administrative constructs
Posing relational model-base structures as alternatives to ERP schema is not the same thing as casting them as competitors. For one thing, ERP products are instances of information technology, while relational model-base structures are comprehensible as conveyances of decision technology. Thus, for relational model-base structures, managerial decision requirements sit at the center of attention, while they remain something of an afterthought for ERP authors. ERP and RMB constructs are, or ought to be, aimed at different constituencies; they should not come into contact, much less conflict, so long as each remains in the functional application contexts for which it’s instrumentally best fit.
2.1 Bounding the ERP Domain
ERP packages are the most current creations of the mainstream management information systems community. They are to be distinguished from their MIS predecessors by their far more expansive focus. While the attention of the typical MIS is fixed on the managerial requirements pertinent to a singular organizational unit, the domain of ERP packages extends to organizations in their entirety. So, whereas an MIS might have been designed to support, say, a department manager, ERP-type systems are aimed at an organization’s highest-order (central) administrative authorities.
Of the challenges that fall to central management operatives, one of the most obviously pressing is that which holds them responsible for realizing adequate, if not maximal, levels of lateral integration. Lateral integration is concerned with the character and intensity of the interconnections (couplings) among the components of which a system is comprised (Sutherland, 1998). Higher levels of lateral integration are generally to be desired because more “tightly coupled” systems can be forcefully asserted to have higher aggregate efficiency potentialities than their more loosely-coupled counterparts, albeit at the price of increased fragility (Perrow, 1984). For an organization, higher levels of lateral integration can then mean more in the way of constructive complementarities among the activities undertaken by its units (departments, offices, divisions, programs, agencies, etc.), and hence better performance in terms of input-output ratios (leverages), yield-type measures (e.g., profitability) or any of a variety of other return-related criteria.
In inter-organizational contexts, a higher level of lateral integration implies more intensely positive interactions among the otherwise autonomous members of some standing alliance or ad hoc association. A key practical test of the worth of any ERP-inspired management support system is then what it provides in the way of instruments that can assist an organization’s administrative authorities in obtaining/sustaining sufficiently high levels of lateral integration. This is a test that ERP products can’t pass all that well. Their limitations in this area stem from a constructional oddity. Despite their far more advanced ambitions, ERP systems ultimately rest on the same referential underpinnings as their more continent MIS cousins: relational database structures.
Actually, what might better be said of ERP packages is that they are essentially just relational database management systems writ large. Relational database management systems are reifications of relational database theory, which is made operational through the offices of certain quasi-axiomatic tools (e.g., Boolean algebra and its correlatives like the Pierce algebra; the calculus of sets, particularly for operations involving binary mappings that can then be used to develop query-language or SQL operators). Variations on these themes have been used by computer scientists to define various classes of canonical data models: Object-Relational models, Associative models, Network models, Entity-Attribute models, etc. Ultimately, however, even the most elaborate of such data models reduce to arrays of two-dimensional data tables housing elementary record-level entries. And the sorts of relations (per relvars) that simple data tables can support are themselves necessarily simple, restricted mainly to rudimentary row-wise connections among discrete data items.
It’s their essential simplicity that made relational databases so attractive to the designers of the archetypal MIS of fifty years ago or so; they were not really all that far removed from the familiar filing-cabinet or shoe-box based media that predated the computer era. But, as information structures now go, relational databases and database-derived models tend to sit towards the more primitive end of the spectrum. Their relevance is threatened by the emergence and spread of more sophisticated information modeling methods and more powerful presentational media (trajectory techniques for capturing the impact of real-time data and charting the course of asymptotic functions, Markov-type stochastic-state situation projections, graphic techniques for portraying probability or density distribution data, the dynamic, multilayer maps underlying Geographic Information Systems, etc.).
As for the sorts of integrative assistance that ERP packages can provide, their contribution will consist mainly in the establishment of extensive data sets (data warehouses) that can allow information-sharing and limited interoperability among some set of application programs. ERP products do not, in and of themselves, have any facilities for directly interconnecting application programs. Other than some arithmetic tools incorporated in Spreadsheet modules, ERP packages may contain nothing at all in the way of embedded, analytical instruments. So, whatever decision-related facilities are needed must be imported from outside vendors as plug-in programs (which more often than not turn out to be rather pedestrian statistical-inference instruments imperiously restyled as Business Intelligence Modules), and then grafted somehow onto the underlying database structure. Because these application programs are appended rather than embedded, any functional interconnections (re: interoperability) will be indirect, mediated by middleware running along the spinal database. The upshot is that, in practice, ERP constructs might not always turn out to be the seamless and stalwart systems promised by their promoters ( Davenport, 2000; Kumar and Hillegersberg, 2000; Markus, 2001; Hong and Kim, 2002; Wu and Wang, 2006).
All this suggests that the prime candidates for ERP products will be organizations whose administrative requirements are clustered around clerical and accounting-related applications. This means the simpler sorts of organizations likely to be found in any sector, such as by-the-book bureaucratic agencies in the governmental arena and, even more obviously, what might be thought of as ordinary business firms (ordinary in the sense that they are configured as process-shops, are manufacturing products or offering services of an unexceptional nature, and are resident in markets characterized by monopolistic or constrained oligopolistic competition.). So, because of their great number and the ubiquity of clerical and accounting based managerial applications, ordinary business firms are the customers that ERP vendors are most anxious to attract.[1]
It’s thus entirely fair to criticize ERP products should they fail in some way to efficiently, effectively or economically serve the interests of any ordinary business firms installing them. And there is some empirical evidence of ill-service, c.f., Linthicum (1999), Scott & Vesssey (2000), Cotteleer (2002) and Sumner (1999). It’s not then to the discredit of the authors of ERP products that problems of the sort that beset firms operating in more extraordinary protean or parlous competitive contexts, or by critical-mission organizations in the public administration arena, would remain somewhere beyond the effective technical reach of the instruments with which they equipped them. On the other hand, ERP promoters would certainly be subject to censure were they to pretend that their products are equally appropriate for all comers. Overselling is unfortunate for all concerned, as many of the more embarrassing misadventures involving ERP packages may stem from their being sold to organizations for which they are technically unfit.
2.2 Roles and Rationales for Relational Model-Base Structures
If relational model-base structures are going to have any pragmatic appeal above and beyond that of ERP products, they must be able to do something to bring more demanding administrative decision situations more fully within the grasp of modern managerial technology. Of the ways in which administrative decision situations can become more demanding, the two most pertinent to relational model base structures are, firstly, an escalation of the magnitude or complexity of the attendant analytical operations and, secondly, any tightening of time constraints.
With respect to administrative matters, heavy computation burdens are most likely to affect organizations that see aggregate efficiency as the critical arbiter of their success or survival. This means, for example, firms engaged in industries/markets where competition is particularly intense, public agencies struggling with deep needs and shallow budgets, and any institutions ―military, medical, governmental or some admixture― whose missions mandate reasonably well-developed emergency or crisis management capabilities. As for time constraints, these will be of most consequence to organizations resident in volatile environments, where threats can materialize and mature quickly and where opportunities are fleeting.
The organizations for which relational model-base structures should be most relevant will then be those where administrative decision situations which tend to be long on analytical burdens and short on leisure. There is, however, an important pragmatic reservation that needs noting. As currently conceived, the only analytical burdens that relational model-base structures are meant to help carry will be those that are rooted in computational (rather than judgmental, conjectural or creative) requirements. To fix the boundaries even more finely, the domain of technical authority for relational model-base structures extends to cover only those administrative decision situations where a satisfactory solution can be arrived at via one or another of the two most common families of quantitative analysis instruments (i). Deterministic optimization techniques, and (ii). Statistical-inference based tools for addressing probabilistic problems.
The practical conclusion of all this is that relational model-base (RMB) structures are designed to serve two sets of users: (1). Enterprises whose onerous efficiency objectives put them in need of more capable integrative instruments than ERP packages can be relied on to provide, and (2). Organizations attempting transitions from traditional planning-based (anticipatory) managerial platforms towards adaptive/dynamic (real-time) regimes, this as a tactic for coping with increasingly erratic environments. The primary way in which relational model-base structures can assist such organizations is by providing the instrumental foundations for the construction of decision-driven management support systems.