Content area
The need for an integrated framework for selecting enterprise-wide hardware is addressed by presenting a comprehensive methodology for a computer family selection. The proposed methodology is incorporated into a decision support system that aims to help organizations select a family of computers from a manufacturer's product line, rather than choose a specific computer model for a given purpose.
Full text
ABSTRACT
The growing complexity of corporate information systems has prompted many organizations to seek applicable solutions for integrating their information systems. While the current enterprise information infrastructure in most corporations consists of a hodgepodge of hardware and software, requiring disparate supporting infrastructures and offering little in the way of integration, many organizations are trying to develop their core client/server competency around a unified hardware environment. Consequently, the marked tendency toward selecting scaleable, flexible, platforms rather than specific hardware for a given purpose, makes the decision to acquire new hardware increasingly complex.
This paper attempts to address the need for an integrated framework for selecting enterprise-wide hardware by presenting a comprehensive methodology for a computer family selection. The proposed methodology is incorporated into a decision support system that aims to help organizations select a family of computers from a manufacturer's product line, rather than choose a specific computer model for a given purpose. A case study follows the outline of the proposed decision support system and demonstrates its applicability and use in a real-life scenario.
Keywords: Systems integration, Interoperability, Downsizing, Computer family, Rightsizing, Computer selection, Analytic Hierarchy Process (AHP), Decision Support Systems (DSS).
INTRODUCTION
The rapid and ongoing advent of computers and data communications technology has brought about major changes in organizations' computerization processes but created a virtual Tower of Babel that information technology (IT) managers are struggling to decipher. In most corporations, the enterprise information infrastructure consists of a hodgepodge of hardware and software, requiring disparate supporting infrastructures and offering little in the way of integration (3).
In the race toward increased organizational effectiveness and enhancing an organization's ability to compete, the goal is to integrate all business data needed for critical competitive decisions (26, 27). While mismatched hardware and software impedes system integration, the benefits of system-wide compatibility include an ability to migrate from one system to another and transfer application software between systems, using a common operating system. Failure to do so can mean lost productivity, inability to seize business opportunities and ultimately, financial loss. Coping with this goal, however, clarifies why the hardware infrastructure should be a carefully planned integrated entity, rather than a simple sum of its parts (14, 17). Consequently, determining the computer acquisition strategy that best fits the organization's needs and ensures a scaleable, flexible, infrastructure, is one of the most important decisions that IT executives must make (4).
In the drive to create an efficient and robust computing infrastructure, IT managers are faced with the issue of hardware management and must identify methods for selecting strategic vendors to unify their infrastructure. Their challenge in the next years is to develop an adaptive architecture that includes selecting key vendors that will provide system-wide adaptability, scaleability and flexibility.
Even the most talented IT architects have been unable to project the evolution of hardware and software platforms nor have they been able to predict server/network application requirements (10). IT groups must adopt platforms that scale over at least a five-year window to avoid costly conversions, migrations and disruptions to business. Moreover, since applications forecast and longevity often exceed plans, the real need is to protect enterprise technologies by providing a growth path that will support the normal five to ten years life for applications (1).
This paper advocates the idea of addressing systems integration by creating system-wide compatibility through the early stage of computer selection and evaluation. The basic elements of a conceptual framework for computer family selection are explored and a decision support system (DSS), that can be used by managers to effectively direct the selection process, is developed. A case study follows the outline of the proposed DSS methodology and demonstrates its applicability in a real-life scenario.
COMPUTER SELECTION AND SYSTEM INTEGRATION
To maintain a maximum level of integration and allow for scaleability and flexibility of computer platforms, organizations are often faced with the problem of selecting a group of systems that will work in harmony rather than selecting a specific machine for known and identified needs. Such a group of systems is called a computer family and is defined as:
Computers of the same type, consisting of several models from the same manufacturer's product line, ranging from microcomputer to mainframe, with full compatibility in the operating system and the system's software, to enable transfer of application software from one family member to another without change (6).
The following are two examples of state-of-the-art computer families. The specific models within each family are classified into four categories: PC/workstations, mid-range servers, high-end/departmental servers, and corporate/ enterprise servers.
The first example of a computer family is Digital Equipment Corporation's (DEC) Alpha systems, based on 64bit RISC technology, consisting of AlphaStations 250 and 600 at the PC/workstations level; workgroup (mid-range) server models 400, 1000 and 2000; departmental/high-end servers such as the 2100 and 8200 models; and Alpha 8400 at the corporate server level.
Another example is the Hewlett Packard HP9000 family, consisting of models 712 and 715 at the entry level; Graphic Super Deskside SMP workstations J and model 800/E as midrange servers; model 800/K high-end servers; and model 800/T corporate servers.
Making a decision about a computer family to serve an entire organization, up and down the hierarchy and across functions and geography, is inherently strategic (4). In a mature, large organization, a computer family offers a dataand application-sharing channel from corporate mainframes down to departmental microcomputers. A growing organization can buy into a family at the low-end workstation level, and grow with it rather than face the step-function trauma of outgrowing its early computing solutions.
The computer selection issue has been discussed in the literature since the early days of computing (7, 22, 25). However, because that era of computing was attributed by centralized, mainframe-based, infrastructure, the various methods for computer selection that have been discussed in the literature (Figure 1) addressed the selection of a specific computer to meet particular requirements.
As the shift of information technology toward distributed systems accelerated and the client/server craze swept the information systems industry, the need to build integrated architectures to ensure a uniform computing environment throughout an organization has been recognized (9, 11). IS managers realized that a well-planned hardware infrastructure provides a simple and elegant framework within which systems can be easily integrated and into which new systems can be inserted. Consequently, the issue of selecting a computer family was introduced (6).
In selecting a computer family, an organization cannot compare competing products on a one-to-one basis (e.g., the HP9000-712 machine versus Alpha Station 250), but must rather compare groups of computers with similar characteristics (e.g., HP versus DEC workstations).
Borovits and Zviran (6) proposed a generic method for the selection of a computer family. Their method was further elaborated by Zviran (29) who incorporated Saaty's Analytical Hierarchy Process (18, 19, 20). The revised method consists of ten steps, as depicted in Figure 2. This method covers the entire process of selecting a computer family, from the identification of possible vendors to the selection of a computer family that best suits an organization's needs. Saaty's AHP is incorporated into steps 5 and 7. It provides an objective weighting technique for setting the weighting-scale for the selection criteria (step 5). It also eliminates any influence by subjective considerations during the evaluation of competing computer families in step 7. Finally, it supports an examination of the evaluator's consistency and contributes to the overall effectiveness of the selection process.
Commitment to a computer family is an attractive alternative to a suboptimized patchwork of stand-alone systems that can be the price of tactical computer buying. Yet while the advantages of selecting a computer family are too great to be ignored, choosing a computer family is a methodical, tedious, time-consuming and counter-intuitive process (13, 29). Thus, an automated tool to aid a decision maker in the selection process is much in need.
THE ANALYTIC HIERARCHY PROCESS
The Analytic Hierarchy Process (AHP) was introduced by Saaty (18) as a method for assessing the importance of a large number of interacting factors, developing priorities among the factors and choosing a best alternative in an objective manner (18, 19, 20).
The method is based on pairwise comparisons between all relevant factors. In each pairwise comparison, a decision maker evaluates two factors and answers the question: "Which of the following two factors dominates the other, and by how much?" The first part of the question is clearly an ordinal question, while the second part is a cardinal one, requiring a numerical input. The answer is based on a nine-point numerical scale, as defined by Saaty (18) and presented in Figure 3.
The answers to these evaluations comprise the input data for a comparison matrix. The size of this matrix for n factors is n*n. Each cell in the matrix represents a pairwise comparison between two factors, showing the relative contribution (to the subject of comparison) of the ith element as compared to the jth element. The matrix has positive entries everywhere and satisfies the reciprocal property, i.e., a^sub ji^=1/a^sub j,i^.. Therefore, when the ijth element of the matrix is specified, the jth position is automatically determined to be its reciprocal value. Thus, the number of pairwise evaluations required for n factors is 1/2(n^sup 2^n). Figure 4 depicts an example of a comparison matrix of six factors.
After a comparison matrix is filled, its eigenvector corresponding to the largest eigenvalue is calculated and normalized so that its elements sum to unity. Saaty (18) provides a set of mathematical proofs that the values of this normalized Eigenvector (the right-most column in Figure 4) constitute the factors' relative weights.
Another matter of concern is the quality of the answers provided in the comparison matrix and in particular, the problem of consistency. An inconsistency occurs if A is preferred over B, B is preferred over C, and A is not preferred over C. The consistency ratio is a measure of the reliability of the subjective judgments used in the creation of the comparison matrix. This is assessed by considering whether a^sub ij^=(a^sub i,k^ *(a^sub kj^) holds for all triplets. The consistency ratio (CR) is calculated for the maximum eigenvalue according to a mathematically proved formula suggested by Saaty, and is required to be less than 0.10 (on a scale of 0-1) for acceptable consistency.
The application of the AHP to the computer family selection methodology makes the resolution of ranking and weighting alternatives less arbitrary. In step 5, the AHP allows a decision maker to objectively create a prioritized and weighted list of criteria. At each level of the hierarchy, every criterion can then be compared to all the others in its group, on a one-to-one basis. Step 7 consists of the process of receiving, comparing and analyzing bids. This is the second opportunity to incorporate the AHP into the selection process. After assigning each relevant computer model from each proposed computer family to a category (PC/workstation, mid-range server, high-end/departmental server, or corporate/enterprise server), each category is evaluated using the criteria scheme from step 5. Again, the advantage in applying the AHP to this step is achieving greater objectivity as categories of computers from different manufacturers' product lines are evaluated.
DESIGN OBJECTIVES FOR SELECT
Figure 2 portrays the complexity that is inherent in the process of selecting a computer family, particularly when Saaty's AHP is used (29). Thus, a computer-based decision support system (DSS) to aid a decision maker in the selection process would be of considerable utility.
The principal design objectives for a DSS to support the computer family selection process include:
Comprehensiveness - the system should cover the entire selection process and provide support to all stages of this process.
Prescriptiveness - it should guide a user through the various stages in the selection process and should clearly explain the activities involved in each stage.
Ease of use - since typical users of this DSS are likely to be senior managers rather than computer experts, it should be easy to use, producing output that is reliable and comprehensive.
Responsiveness - many organizations will avoid using multi-criteria evaluation schemes (e.g., Saaty's AHP) because they are lengthy and complicated. A DSS that can respond quickly will enhance the chances that organizations will actually undertake this planning activity.
Reliability - a DSS must employ well-accepted and validated models.
SYSTEM ORGANIZATION AND FUNCTION
SELECT is an interactive DSS that was developed to facilitate the use of Saaty's AHP in computer family selection by assisting managers in carrying out the selection process. It is designed for a single user but can be used, through a facilitator, to support a group of decision makers to select a computer family.
SELECT follows the ten steps of the computer family selection procedure depicted in Figure 2. Its main menu, presented in Figure 5, allows a user to access any of the major steps in the selection process. Many menu options lead to a hierarchy of sub-menus which direct a user to the desired functions. When a specific function, such as Analyze Bids, is requested, a user is prompted to provide the required inputs. The prompts are presented with either a question-and-answer or query-by-form format. The following is a step-by-step description of the workflow with SELECT.
The General Information entry in the main menu describes the selection process and the functionality of SELECT. It provides a brief description of the computer family selection methodology, its rationale, and the major steps involved. It then describes the features of SELECT as a computer-based system to support the decisions involved in the selection process. On completing this item, the user returns to the main menu
The first step in selecting a computer family is to create an initial vendor list. This step requires the inputting of all vendors whose product-lines might suit the organization's needs, in accordance with the definition of a computer family. The output of this step is an initial list of vendors whose product line may suit the organization's needs. Selecting Prepare Vendor List from the main menu allows a user to initiate a vendor list and manipulate it. Inclusion of a vendor on this initial list does not signify that the vendor is able to satisfy the organization's requirements. Figure 6 presents the form used to set up a vendor record.
The second step in the computer family selection process is to determine mandatory requirements. These requirements define the basic features required of a computer family. They are derived from the definition of a computer family as well as from the organization's information systems (IS) policy. The output of this step is a set of requirements which are considered as prerequisites for a vendor's candidacy (e.g., full compatibility of system's software, ability to upgrade each model to a higher one without change in software and operating procedures, etc.). The Set Mandatory Criteria option in the main menu allows a user to create the mandatory criteria through a hierarchy of menus and on-line forms.
The third step aims at evaluating vendors' compliance with mandatory requirements. Information regarding each vendor's compliance with these requirements is obtained from all potential vendors and is examined by the selection team. The list of vendors is screened and those suppliers that do not comply with the mandatory requirements are winnowed out. Through the Mandatory Criteria sub-menu of SELECT, a user selects one vendor at a time and is then presented with the list of mandatory criteria. For each vendor, the user highlights the mandatory criteria with which the vendor complies. That information is then stored in a database. When all vendors have been evaluated for mandatory criteria compliance, the system automatically searches for vendors who have not complied with all of the mandatory criteria and presents a list of candidates to be eliminated. These are dropped from further consideration only after the user's approval. The vendors remaining after this elimination procedure constitute the mailing list for the Request For Proposals (RFP).
Next, an evaluation scheme must be created. This step establishes a framework within which all bids will be evaluated. In order to select a computer family that best fills its requirements, an organization must designate the features that will be used to compare the computer families. These characteristics are its evaluation criteria.
All criteria used in the evaluation process can be slotted into the hierarchical scheme, as illustrated in Figure 7. The top of this hierarchy, denoted as Total in Figure 7, stands for the total score achieved by a computer family. The next level down sorts the criteria into qualitative and quantitative. The next level under quantitative criteria defines the categories of computers. Subsequent levels define specific attributes by which competing families will be evaluated. The lowest level consists of atomistic elements which describe specific characteristics by which the particular computer models are to be evaluated.
The Formulate Evaluation Scheme option in the main menu covers all seven sub-steps involved in formulating the evaluation scheme (step 5 in Figure 2). It compels a user to prioritize the overall importance of qualitative versus quantitative criteria; set qualitative criteria (such as manufacturer and vendor support, vendor reputation, etc.); select applicable computer (PC/workstation, mid-range server, high-end/departmental server, or corporate/enterprise server); set quantitative criteria (e.g., memory, disks, printers); and select sub-criteria for each criterion (e.g., size, speed, cache and expandability of the main memory), down to the lowest level.
After criteria lists are generated, they must be weighted according to their relative importance. Research evidence indicates that the consideration of all criteria simultaneously to determine their relative ranking is difficult (19). The AHP approach avoids this problem by using a pairwise comparison technique. All criteria within a given level are compared to each other to determine their relative importance. Every element being compared is rated against all other elements on the same level, on a one-to-one basis. For example, a typical question might be: "In comparing Memory Capacity to Memory Speed, which is more important to the organization and to what extent?" Rather than asking a decision maker to determine a quantitative value, the AHP technique only requires choosing between nine comparison levels, ranging from "both criteria are equally important" to "one criterion has absolute preference over the other" (Figure 3). Based on the scheme identified in Saaty (18), a numerical value is obtained and entered into the system. All pairwise questions are automatically generated by SELECT and presented sequentially to the user. The input values are stored in comparison matrices. When the user has made all comparisons, the relative scores and consistency ratios are calculated and presented to the user. The relative scores represent the weights assigned to each criterion, according to its relative importance. The normalized principle Eigenvector of the comparison matrix at a given level represents the priority of the criteria for that level (18, 19). Figure 8 depicts a comparison matrix with its pairwise comparisons and relative weights calculated for memory characteristics.
Next, the reliability of the comparison matrix, which reflects the subjective judgments of the evaluator, is measured. The consistency ratio is a built-in mechanism to ensure the consistency of all responses and weights. The calculated ratio is on the scale of 0-1 and in cases where the consistency ratio is greater than 0.10, the user is advised to go through another iteration of pairwise comparisons at that level before proceeding to the next level. The process is repeated for every level, down to the atomistic items at the bottom of the hierarchy. After all relative scores have been calculated, the absolute weight for a criterion is computed by multiplying its relative weight by the relative weight of each of its predecessors in the hierarchy, or by the absolute weight of its immediate predecessor. The end-product of the weighting process is a weighted evaluation scheme by which all bids will be evaluated.
Once an evaluation scheme has been established, the next step in the selection process is to send requests for proposals (RFP) to vendors which meet all the mandatory criteria. The RFP consists of a summary list of specific requirements according to which vendors will be asked to write their proposals. The first part of an RFP focuses on quantitative criteria relating to each computer model within a proposed family. The second part addresses qualitative criteria. It concentrates upon issues such as uniformity and transferability of systems software, conversion of present applications software to the new computer family, environmental considerations, etc. The Prepare Requests for Proposals option in the main menu helps a user to formulate the RFP and mail it to vendors that meet the mandatory criteria.
In response to the RFP, bids for proposed computer families will be received. These bids will be evaluated as a basis for selecting a computer family using the Analyze Bids option from SELECTs main menu. The user is prompted to indicate which vendors returned a bid, and how many separate computer families were included in each bid. Vendors which did not return a bid are eliminated from further analysis.
Next, each relevant model from each proposed computer family is assigned to a computer category. The determination of the category into which a computer will be placed is based on, but not limited to, such factors as CPU performance, memory size, external storage capacity, number of disk drives, cost, etc. The decision as to which factors will indicate placement into a particular category will have been specified when the categories were determined.
After the bid information has been entered, the next step is to evaluate the bids using the evaluation scheme. This step incorporates Saaty's AHP to evaluate competing computer families for each selection criterion. During this stage, each computer model is evaluated within the category to which it is assigned. If a vendor proposes more than one model in a given category, all proposed models are evaluated (e.g., if DEC's proposal listed the 400, 1000 and 2000 Mid-range Servers, then each of these models would be evaluated for that category). All computer models in a given category are compared pairwise according to each criterion. SELECT automatically generates questions such as: "In comparing model A and model B, on the basis of criterion X, which model is preferred?" The numeric scale for the responses is based on Saaty's 1-9 scheme (Figure 3) (18). Pairwise comparisons of all competing models in each given category and for all criteria are performed through the set of questions presented by SELECT and the response value is entered into the comparison matrices. After the values for a criterion are obtained, SELECT calculates the relative scores and consistency ratio, and displays them. When all bids have been evaluated for both qualitative and quantitative criteria, the user can display the resultant scores. The Display Scores option totals and displays scores for each computer model as well as for an entire family. A user can also display the members which scored the highest within a family.
Following the evaluation of the bids, a final vendor list is determined. On the basis of the final scores obtained by each computer family, irrelevant candidates can be disqualified, leaving vendors most likely to meet an organization's needs. These computer families are then further evaluated through benchmark tests to ensure they have the proper capabilities and characteristics.
Once a final vendor list has been determined, a user can choose a computer family or continue the evaluation process with benchmark tests. A benchmark, in the context of this discussion, is a set of live tests designed to examine the actual performance of proposed systems (hardware and software). At this stage, SELECT will passively support a user in monitoring a benchmark process and recording its results. A user can select Enter Benchmark Test Results from the main menu to create, edit and display the benchmark test list and results. The benchmark test results will be manually evaluated and recorded for all remaining candidates.
The final step, Determine Final Ranking in the main menu, displays the evaluation and benchmark scores for all remaining computer families and ranks them as a basis for final selection. SELECT suggests the highest ranked computer family, but the user makes the selection decision.
The last entry in the main menu, Display Process History, allows a user to display the information that was recorded throughout the selection process. It can also produce a hard copy of the history or export an ASCII file to be used by any word processor and incorporated into other documents as well. While not an integral part of the actual computer family selection process, this option provides an organization with an audit trail of the selection procedure.
SELECT was used by a large construction company to assist in selecting its computer family. The appendix of this paper demonstrates the actual step-by-step use of the system in the computer family selection process performed by this company.
DISCUSSION AND CONCLUSION
The need to select a computer family arises from the trend toward distributing computing resources. The distributed client/server model puts unprecedented power on the desktop by giving users access to applications and huge amounts of data stored on local servers. This trend, however, also brings unprecedented complexity, as most organizations end up with multi-vendor environments that make interoperability one of the biggest challenges facing technology managers today. Thus, organizations that are either considering a shift toward distributed systems or are restructuring their distributed infrastructure should be in a position to evaluate and select a computer family rather than selecting a specific computer model.
Selecting a computer family ensures a uniform computing environment for an entire organization. This environment provides hardware and systems software compatibility across platforms and minimizes the cost of systems integration. Moreover, the ability to transfer application software from one family member to another without change obviates duplication in software development and lays the foundation for coordinated development and implementation of a consistent organization-wide information system. Another advantage of selecting a computer family is the opportunity for a one-time evaluation and selection procedure.
Selecting computer families is a complex process. A formal methodology has been developed to aid decision makers in selecting computer families and reduce the uncertainty a decision maker faces in choosing the right system for an organization. However, this selection process requires the decision maker to collect and analyze a large amount of information. The administrative tasks associated with such a volume of information significantly increase the time a decision maker must spend on the evaluation and selection process. SELECT aids the decision maker to efficiently evaluate this information. It eliminates the time-consuming administrative tasks associated with handling the information, allowing a decision maker to concentrate on the evaluation and selection process.
Using Saaty's analytic hierarchy process, SELECT has made the computer family selection procedure usable for all decision makers. It is user-friendly and does not require detailed training. Through the use of this decision support system, decision makers can quickly apply the computer family selection process to support their organizational needs. SELECT offers four advantages to a decision maker contemplating the selection of a computer family.
1. Walk through. SELECT covers the entire process of selecting a computer family, from the identification of possible vendors to the selection of a computer family that best suits the organization's needs. It guides the decision maker through all the steps involved in the selection process. It prompts the user to provide all the necessary inputs in each stage, and it produces the relevant outputs.
2. Computations. The application of Saaty's AHP to the selection methodology makes the resolution of ranking and weighting alternatives less arbitrary. SELECT automatically performs all the calculations within the selection process and releases the user to focus on activities that are directly associated with the evaluation and selection process.
3. Consistency. A threat to the weighting procedure is inconsistency. By calculating the consistency ratio in each iteration of Saaty's AHP, SELECT monitors the weighting and scoring to ensure the consistency of all responses.
4. History. SELECT documents the entire selection process and provides a complete log of all its stages - from initiation to the final selection.
Future research should assess the usefulness of decision support systems and multiple criteria evaluation schemes (e.g. Saaty's AHP) for computer selection tasks. Another avenue will focus on transforming SELECT into a group decision support system (GDSS), preferably intranet-driven. Since decisions of this type would, most likely, be made collectively by a group of decision makers rather than by a single individual, a GDSS seems to be appropriate. Such a system will retain the capabilities of SELECT while supporting the collaborative work of the decision group through advanced features, including inter-organization communications among group members, progressive rounds of voting toward consensus building and methods for calculating and aggregating weights and scores.
Eigenvalues and eigenvectors are specific characteristics of a matrix. An eigenvalue, lambda, of a matrix A is deEmed as the root of its characteristic equation; The eigenvector is a nonzero vector (xi) such that (xi)A=(xi)lambda, where lambda is the eigenvalue of matrix A. Further discussions on the specific properties of these characteristics can be found in any text on linear algebra.
REFERENCES
1. Ahituv, N. and M. Igbaria. "A Model for Predicting and Evaluating Computer Resource Consumption," Communications of the ACM, 31:1, December 1988, pp. 1467-1473.
2. Ahituv, N., S. Neumann, and M. Zviran. "Factors Affecting the Policy for Distributing Computing Resources," MIS Quarterly, 13:4, December 1989, pp. 389-401.
Allen, B. "New Dimensions in the Data Center," Computerworld (White Paper Supplement), 31:42, October 1997, pp. 1-11.
Ameen, D. "Evaluating Alternative Computer Acquisition Strategies," Journal of Systems Management, 41:9, September 1990, pp. 15-20.
Borovits, I. Management of Computer Operations. Englewood Cliffs, NJ: Prentice-Hall, 1984. Borovits, I. and M. Zviran. "Computer Family Selection for Organizational Systems," Information and Management, 12:3, March 1987, pp. 107-115. Brandon, D.H. "Computer Acquisition Method Analysis," Datamation, 18:9, September 1972, pp. 44-52. Ein-Dor, P. "A Dynamic Approach to Selecting Computers," Datamation, 23:6, June 1977, pp. 103-108. 9. Gunton, T. Infrastructure - Building a Framework for Corporate Information Handling. Englewood Cliffs, NJ: Prentice-Hall, 1989.
10. Igbaria, M. "Managing and Forecasting Hardware Resource Consumption," Information Systems & Operational Research (INFOR), 28:3, 1990, pp. 412421.
11. Jackson, L. "Organizing for the Integration of Information Technologies." In Umbaugh, RE. (Editor), Handbook of MIS Management, Second Edition. New York: Auerbach, 1988, pp. 135-140.
12. Joslin, E.O. Computer Selection. Arlington, VA:
Technology Press, 1977.
13. Muralidhar, K., R. Santhanam, and R. Wilson. "Using the Analytic Hierarchy Process for Information System Project Selection," Information and Management, 18:2, February 1990, pp. 87-95.
14. Nilsson, E.G., E.K. Nordhagen, and G. Oftedal. "Aspects of Systems Integration," Proceedings of the 1st International Conference on Systems Integration, Morristown, NJ, April 1990, pp. 434-443. 15. Robertson, P. "Integrating Legacy Systems with Modem Corporate Applications," Communications of the ACM, 40:5, May 1997, pp. 39-46.
16. Roenfelt, R.L. and R.A. Fleck. "How Much Does a Computer Really Cost?" Computer Decisions, 7:5 November 1976, pp. 75-79.
17. Rossak, W. and P.G. Ng. "Some Thoughts on Systems Integration: A Conceptual Framework," Journal of Systems Integration, 1:1, 1991, pp. 97-114. 18. Saaty, T.L. "A Scaling Method for Priorities in Hierarchical Structures," Journal of Mathematical Psychology, 15:3, 1977, pp. 234-281.
19. Saaty, T.L. The Analytic Hierarchy Process. New York:
McGraw-Hill, 1981.
20. Saaty, T.L. The Logic of Priorities. Boston: KluwerNijhoff, 1982.
21. Seidmann, A. and A. Arbel. "Microcomputer Selection Process for Organizational Information Management," Information and Management, 7:5, December 1984, pp. 317-329.
22 Sharpe, W.F. The Economics of Computers. New York: Columbia University Press, 1969.
23. Shoval, P. and Y. Lugasi. "Computer Systems Selection: The Graphical Cost-Benefit Approach," Information and Management, 15:3, October 1988, pp. 163-172. 24. Shoval, P. and Y. Lugasi. "Models for Computer System
Evaluation and Selection," Information and Management, 12:5, March 1987, pp. 117-129. 25. Timmreck, E.M. "Computer Selection Methodology,"
Computing Surveys, 5:4, December 1973, pp. 199-222. 26. Trino, J. "The Competitive Advantage is Provided by Integrated Information Systems," Industrial Engineering, 22:2, February 1990, pp. 36-40. 27. Weber, Y. and N. Pliskin. "The Effect of Information Systems Integration and Organizational Culture on a
Firm's Effectiveness," Information and Management, 30:2, May 1996, pp. 81-90.
28. Wyse, J.E. and C.A. Higgins. "MIS Integration: A Framework for Management," Journal of Systems Management, 44:2, February 1993, pp. 32-37. 29. Zviran, M. "A Comprehensive Methodology for Computer Family Selection," Journal of Systems and Software, 22:2, July 1993, pp. 17-26.
Tel Aviv University
Tel Aviv 69978, Israel
Copyright International Association for Computer Information Systems Winter 1998/1999