Future Actions to Promote EI Collaborations Minutes of the Special Interest Group: "Future Actions for EI Promotion" held at the ICEIMT Conference, Hilton Head, South Carolina, June 8-12, 1992 as recorded by Robert Kwikkers The following people are working group members: Sandy Cavalli MCC Austin, Texas Robert Ducoffre Cap Gemini Innovation Brussels, Belgium Jim Fulton Boeing Computer Services Seattle, Washington Wim Gielingh TNO Construction Research Delft, Netherlands Ted Goranson SAIC Virginia Beach, Virginia Alexander Hars Institut fuer Wirstchaftsinformatik Saarbrueken, Germany Bruce Jorgenson Boeing Computer Services Seattle, Washington Kurt Kosanke ESPRIT Consortium AMICE Brussels, Belgium Rob Kwikkers TNO Industrial Research Eindhoven, Netherlands Brigitt Lutz-Kunisch Inst. Industrial Tech. Bremen, Germany Brian Seitz IBM Santa Monica, California Dirk Solte FAW an der Universitaet Ulm Ulm, Germany Bruce Speyer MCC Austin, Texas Joan Tyler NIST Gaithersburg, Maryland Jack White Industrial Technology Institute Ann Arbor, Michigan Section 1. Introduction This document recommends actions to promote further US/EC collaboration on Enterprise Integration Modelling. As a basis, it represents the opinion of the SIG members. It was presented in a plenary conference session on Friday, June 12, 1992. Remarks that were made there have been incorporated as well as possible. The Special Interest Group is of the opinion that an important task in the near future is to link the enterprise integration effort to the interests of the business community. All members have noticed that it becomes increasingly hard to obtain funds for research and development in EI, from funding agencies as well as individual enterprises. One of the reasons for this is that the EI community has failed sofar to convince businesses of the benefits of enterprise integration. EI has been seen by the developers as a goal in itself. As a result, hardly any deliberate effort has been made to build a solid business case for EI. In addition, the EI community is divided in itself about the ways and means for obtaining enterprise integration, each group promoting its own solutions against other suggestions. In so doing, no united front is presented to the outside world. This hurts the common interest. We feel that the business community will only invest in EI research and development if a convincing case can be made for the business benefit of EI, and if the EI community presents itself a united group which is committed to make EI work. The business community is reluctant to fund yet another development which promises largely the same benefits as Management Information Systems, Computer Integrated Manufacturing, and even automation itself have done in the past. The commonality of all these developments is that each of them failed to deliver the promised improvements in business results. The business community is also reluctant to fund isolated efforts in EI, fearing they will bet on the wrong horse and are investing in a technology that will turn out not to become the world standard after all. Experience with various operating systems, hardware platforms, software tools, and communication systems have been hard lessons for businesses. Especially in the economic downturn we are currently experiencing, business have a cash-flow saving wait-and-see attitude, hoping that others will invest in the new ideas. Business hang back, hoping to reap the benefits through adopting the technology once it has been proven. The Future Actions SIG concludes that the ICEIMT conference can be the starting point of a joint effort for all EI researchers, with the aim: o to promote EI to industry, by combining efforts to build a convincing business case for EI o to support integration of various EI frameworks by advancing the mutual understanding of EI modelling techniques Section 2. Recommendations To achieve the above goals, the working group "Future Actions to Promote EI Collaboration" suggests a number of actions to be taken. These actions are indicated in the recommendations below. Ted Goranson (US) and Kurt Kosanke (EC) are designated as Action Officers. They will (initially) lead and coordinate the implementation of future actions. 2.1 Development of Enterprise Integration Technology There should be continued international coordination of the investigations and developments of Enterprise Integration Technology Actions for the Action Officers include: o coordination of a well documented, open public debate on EI, identifying strengths and weaknesses of various approaches o dissemination of technical papers for this purpose. o maintenance of mailing list and information services (i.e., EINet at MCC) The Future Actions working group will act as an advisory board for the action officers who will work on: o further work on integrating actions o motivating the uniqueness of the EI effort o identifying pilots and business cases for EI o identifying liaisons with professional societies o establishing a forum for scientific review 2.2 Integrating the Integrating Structures We need to work on integration of the Integrating Structures. Actions to pursue this goal are: o cooperation among conference participants to achieve understanding o promotion of EI unification ideas to our companies We recommend use the following approaches: o use of a pragmatic approach for unification, trying to avoid incompatibility with other Integrating Structures. o alliance with other, ongoing efforts, given the fact that we have no separate funding. This step must include a commitment to bring relevant information from this conference into these other forums. o objective management of the effort by some disinterested party 2.3 EI Business Case and Promotion Enterprise Integration must be brought to Commercial Industry. A deliberate effort must be made to commit people in enterprises that have the decision power over architectural issues. Actions to achieve this include: o development of business metrics for EI o development of a model-based and case-based demonstration of the business benefits of enterprise integration o improvements to operations and strategic position' problems it will help to avoid or help to overcome: - reduction of cost - definition of a forum where businesses that want to federate can define what they need - bringing the EI results for some sanctioning body 2.4 EI Pilots Pilots must be used to demonstrate the applicability of EI: o practical use of EI must be demonstrated to achieve commitment o business benefits must be emphasized to raise interest o appealing results must be the focus A prime candidate should be an international pilot, for example, a distributed manufacturer The Action Officers will work on finding pilots for EI: o talk to pilot subjects and pilot sponsors o find out which technical solutions are candidates 2.5 Establishing an EI Community We need to establish an identity for Enterprise Integration We must find the aspects that make the domain of Enterprise Integration unique: o international collaboration - multi-disciplinary approach It is desirable to establish an ongoing professional society group on this subject (for example, IEEE) 2.6 Metrics Development Metrics for Model Quality and Integration Quality must be developed (since you want to avoid integration with low quality models): o work should start on adoption of general quality concepts specific for EI NIST will raise interest and coordinate actions. 2.7 Intended Actions Connections will be sought with ongoing coordinating efforts on EC and US Ideas have been raised for: o CIMOSA to use some of the approaches presented in the conference and apply them in pilots, together with others o FAW to continue work towards integrating the various Execution Environments (AMBAS, DAA, DCE) o Wim Gielingh to introduce the EI issue at the advisory board for EC project coordination o Randy Aranguren to establish a connection with TC184 SC5/WG1: develop a standard for a framework for EI). The dialog may be supported in that forum. o involvement of TC184 SC21 for the technical issues o establishment of an ISO TAG for EI issues Future Actions - Position Statements Individual Position Statements submitted to the Special Interest Group "Future Actions for EI Promotion". Prior to the ICEIMT plenary in Hilton Head, South Carolina, an invitation was issued for position papers for the Future Actions SIG. Two such papers were solicited and are included herein. Section 3. Some Ideas for Actions / Collaborations Ted Goranson These ideas were generated during ICEIMT workshops, working groups, and subsequent associated discussions to synthesize results. Next actions are of two, interrelated types: research and pilots. 3.1 Research Topics: 3.1.1 METRICS: The need for metrics clearly resolves into technical issues. At root, the need is for models which span the enterprise, which capture quantitative information, and which support diverse user views. In this topic are included all of the difficult issues identified in ICEIMT. 3.1.2 FEDERATION TECHNOLOGIES: The need for "federation" is clear as a component of EI. Consensus has not been reached either on the technical options available or on the optimum approaches. Some hard requirements have been fleshed out by ICEIMT, however, and candidate approaches suggested. All of the most difficult EI problems are covered in this topic. Topics will be diverse, and include constraint and planning issues. 3.1.3 "HUMAN/SOCIETAL FACTORS": It is clear that many of the requirements for EI are determined by nontechnical factors. Some of these are related to specific "corporate" cultures, human/user requirements, group dynamics, "external" (i.e., legal) forces, and the laws (including the economics) governing change. One topic in this category which recurred in ICEIMT concerned the dynamics of collaboration and precompetitive activity of government and industry agents. ICEIMT collected some high level views in these areas. Better, scientific, understanding of the principles at work are required. 3.1.4 TAXONOMY/ID CLASSES: ICEIMT quickly identified the need for commonly shared, standard principles for models. Some, near-term "ontological repository" components can be generated under the pilots, for specific EI sectors. But the more global, common principles involved are not understood. Even their requirements and behavior have not attracted consensus. The research need for this topic extends beyond computer science to philosophical and mathematical core issues. In all of these areas, some immediate default positions are available for pilot action. Research would focus on a three-year and beyond result horizon. In the US, these topics are probably best addressed by DARPA, NSF, and the National Labs. The type of US/EC collaboration best suited will probably be project-to-project collaboration using continuing ICEIMT-like joint forum. 3.2 PILOTS: ICEIMT and prior studies concluded that action needs to be immediately taken in industry; that the lessons learned in the "real world" will drive the research agenda; and that enough is known now that benefit to industry can result. "Pilots" can be interpreted as experiments or prototypes, but here is meant in the sense of assisting enterprises reduce their risk on "real" activity. Therefore, the pilots must be opportunity-driven, and not synthetic. Different "dimensions" of pilots have been identified: 3.2.1 Three models of "Collaboration," each addressing different issues of commonality and portability: a. Defense-to-Commercial, specifically addressing EI infrastructure and associated processes; b. US-to-EC, specifically addressing portability of products, "open" markets for suppliers, and international standard acceleration; and, c. Sector-to-Sector, specifically addressing commonality of infrastructure from one sector to another (for example, aerospace to automotive). 3.2.2 Three models of "Integration," each addressing a different need for functional interaction: a. "Vertical Partnerships," concerned with company-to-company teams, and focused on prime-to-subcontractor/supplier relationships; b. "Horizontal Partnerships," concerned with addressing integration of processes (ordinarily from infrastructure suppliers) into a working whole within a single firm; and, c. "Value Chain Linkages," concerned with electronic commerce among enterprises in a looser way than partnering. 3.2.3 Three EI "Missions." Each of these address a different philosophy of the value of integration: a. "Product" focus. This type of pilot will be measured on the success of the pilot product, for example the costs and time associated with introducing a new type aircraft; b. "Sector/supplier" focus. This type of pilot will be measured on how it improves the cost/time/quality/flexibility equation for a sector and its suppliers, across a class of products; and, c. "Infrastructure" focus. This type of pilot will be measured on the benefits associated with networks, standards, cooperation and policy, perhaps including new business types. 3.2.4 Three modes of sponsorship. Note that "sponsorship" in this context applies only to the portion of the project which makes it a pilot rather than an internal project. In some pilots, this may only be a small percentage of the total cost. a. Government-sponsored with the intent of indirectly supporting a sector of strategic (possibly defense) importance; b. Industry/Government sponsored with the intent of improving the viability of a sector or supplier base; and, c. Industry (i.e., consortia) sponsored with a direct profit motivation. 3.2.5 Sector type. Three industrial sectors are each key for EI (and the global industrial economy): a. Aerospace (including materials and engines); b. Electronics (including semiconductors); and, c. Automotive. This sector is deliberately defined "in the large" to include non-automotive machining and assembly. 3.2.6 Integration paradigm: a. "Product Data" (and other data models) is the central baseline against which process models are integrated; b. "Process Models" (functional processes) are the baseline against which other models are integrated; and c. "Knowledge-based". In this case, enterprise-wide knowledge is employed as a new paradigm. Some analysis has been done concerning pilots and these dimensions. It is suggested that the more "natural" combination of requirements would be three pilots (or pilot families): combining all "a's" together, and similarly "b's and c's". Type "a" might focus on commercial aircraft introduction. US/EC collaboration would be in supplying shared infrastructure for dual sites, one on each side of the Atlantic. Type "b" might focus on semiconductor manufacturing and involve a transatlantic partnership and related consortia. US/EC collaboration could be in jointly supplying technology to reduce risk. Type "c" might focus on vertical partnering infrastructure, the baseline in automotive suppliers; sector-to-sector transference might be to shipbuilding. US/EC collaboration might address common development of infrastructure methods and standards. In the US, pilot actions are best managed by ManTech. Section 4. WAIS as a Collaboration Mechanism Bruce Speyer WAIS (Wide-Area Information Services) was initiated by Thinking Machines, Inc. Dow Jones News Information Retrieval Service, Apple Computer and KPMG Peat Marwick. WAIS provides a distributed client/server architecture that is in the public domain (anonymous ftp from think.com). Since its placement on Internet late last year WAIS has grown to more than 300 publicly accessible information servers. WAIS uses the ANSI Z39.50-1988/V2D3 protocol for interchange of typed information. WAIS applications include regularly updated weather satellite images using GIF and MIME multimedia examples. Commercial, private and local WAIS servers exist too. The Dow Jones server resides on their own DowVision network providing for-pay customized information retrieval for the Wall Street Journal, Barons, and more than 450 magazines. MCC's EINet, under USAF WL/MTI sponsorship, is using WAIS and enhancing it for commercial viability for EI. We propose that the ICEIMT consider recommending WAIS as a distributed collaboration mechanism for development of EI. Collaborators would provide structured typed information as WAIS servers. Presently, WAIS supports transparent retrieval of information across multiple platforms and networks using its capability to construct sophisticated contextual searches. A huge advantage of using WAIS for EI collaboration is that is happening today and is readily accessible worldwide. Discussions are taking place between the WAIS developers on how to best create an WAIS API (application programming interface). Once this is accomplished it will be easy to integrate applications with WAIS and to create a variety of integration frameworks. For example, it is not difficult to imagine hypertext or various other information resource coordination front-ends becoming available including the ability to launch services or enter into another EI integration system or application. WAIS extensions such as authentication and X.500 directory services are being worked upon by MCC and others. Since WAIS supports a peer-to-peer control paradigm it is possible to integrate various types of systems and information resources. These can range all the way from unsecure experiments on Internet to commercial for-pay transactions on secure networks. It would be possible to access all but the private collaborations from secure environments (e.g., EINet) but not the other way around. MCC is developing EINet to facilitate the types of collaborations articulated in this message and is willing to supply the administration support (e.g., server of servers, etc.) to loosely coordinate and support the interactions of EI collaborators; whether or not the actual systems and information sources are on EINet itself.