Volume  1 of 2

Electronic Data Interchange
Implementation and Transition
Road Map

for the

OSD CALS IWSDB Project

An MVP Joint Venture

March 21, 1995

Submitted by
ManTech International Corporation
Technology Applications Operations Center
1313 Locust Avenue
Fairmont, West Virginia

In support of
Contract DAAB10-89-D-0503
And in compliance with
CDRL Sequence Number A040

SOW Numbers 3.9, 3.9.1.2

EXECUTIVE SUMMARY

Introduction
Today, we have two (2) major standards supporting Electronic Data Interchange (EDI). Here in the United States, we have the American National Standards Institute (ANSI) Accredited Standards Committee known as X12. This set of standards is very robust and has matured since its founding in 1979 "to develop uniform standards for electronic interchange of business transactions."¹ Internationally, primarily in Europe, we have the United Nations Electronic Data Interchange for Administration, Commerce, and Transport (UN/EDIFACT). There has been considerable debate in the United States on when, and in some industries if ever, the U.S. should adopt the UN/EDIFACT standard. In a report for the Department of Defense (DOD) by the Logistics Management Institute (LMI) on the future DOD use of UN/EDIFACT, LMI presented an analysis of the two standards and identified a key element in resolving the issue of incompatibility. The report states that "we need to determine where dictionary commonalty [sic] is necessary." The report also stated that "until a detailed study is performed to determine what data can be carried in those messages, DOD will not know how well current and developmental UN/EDIFACT standards can support its business requirements and how many changes to those standards will have to be made."²

Based on LMI's initial findings, DOD initiated this assessment to:

· Analyze data elements supporting X12 and UN/EDIFACT transactions and develop a data dictionary structure and mapping to account for inconsistencies in definition and usage.
· Determine translation feasibility and test leading commercial translation capabilities for accuracy in translation to and from user environments into X12 and UN/EDIFACT.
· Develop a model for determining a transition plan to UN/EDIFACT in coordination with ongoing LMI and DISA activities.
· Assess Impact of UN/EDIFACT transition on product data and "interactive EDI" for DOD procurement processes.
· Develop measurement criteria for managing and assessing the transition to UN/EDIFACT including metrics for determining return on investment within procurement functions.
· Evaluate INTERNET for Integrated Weapon System Database (IWSDB) support of EC/EDI in the procurement process.
· Assess commercial Electronic Commerce (EC)/EDI computer software and value-added networks (VANs) that could support transition to UN/EDIFACT.
· Analyze security capabilities/plans of UN/EDIFACT.
· Develop a road map for transitioning to UN/EDIFACT, including organizations, processes, objectives and schedules.

The purpose of the study was to determine the technical feasibility of translation and transition from X12 to UN/EDIFACT, and to evaluate support for an EC/EDI data dictionary that is key to translation and transition activities. The analysis focused on the data elements supporting these two standards for purchase order messages.

Given a perfect world, there should be only one EDI standard, and more importantly, a consistent data dictionary supporting that standard. Currently there is neither. The data element analyses of the X12 850 Purchase Order and the UN/EDIFACT ORDERS messages indicate a significant degree (45%) of consistency between the two standards' data definitions. The percent of data elements that do not match is also significant at 39%. The remaining 16% of data elements have some partial degree of consistency. This finding leads to the conclusion that not only are X12 and UN/EDIFACT different at the message comparison level as the LMI report details, but they are also different at the data definition level. Thus, when implementation conventions (ICs) required to use the standard are developed, choices made at the data definition level will compound the difficulty for enterprises to implement a cost effective international EDI solution, or integrate it with their business infrastructures. The proliferation of ICs and resulting inconsistencies in messages and data definitions can undermine the standards themselves and create a costly EDI environment for U.S. Government and industry. Direct translation between the standards is not reasonable to implement without greater consistency between the data definitions.

Driving Forces
U.S. industry is modernizing its information technology infrastructure to more effectively use technologies such as database management systems, CASE tools, and workflow management. The need to have uniform transaction messages for commerce activities is widely accepted, and in some industries such as transportation, banking, insurance and most recently medicine, the need is immediate. The federal government has also recognized the business need for electronic commerce through the White House commissioned Federal Electronic Commerce Acquisition Team (ECAT) and the Streamlining Procurement Act signed into law in October 1994. These initiatives call for government-wide use of electronic commerce before 1996 and the establishment of a Federal Acquisition Network (FACNET) within five years. Clearly, the business and government communities in the U.S. are not going to wait for international standards alignment before proceeding to modernize. But there still exist two quite different standards. The ASC X12 standards community has wrestled with the issue of migrating to UN/EDIFACT for some time and recently drafted an alignment plan that calls for the establishment of a single standard after 1997. The draft ECAT report states, "The two (standards) must coexist in every system during this transition."³ The dilemma now posed for the U.S. is that if EDI, FACNET, and critical industrial dependencies are built upon X12 before 1997, how will the international UN/EDIFACT requirement be supported? Will the U.S. have become so dependent upon X12 that UN/EDIFACT can only be supported separately? Is it cost effective to support both standards given that the consistency of data translation is less than fifty percent?

Findings
Our main conclusion is that both standards can be supported if the underlying data definitions are consistent. That is, if there is a significant effort to align the data directories of the two, then the impact of different message syntax and functionality can be minimized. Any translation between the two, as is done today, requires a force-fitting of the data definitions. Thus, most translations are customized and therefore unique to those trading partners. These are then implemented through mapping tables that map the EDI data elements to the enterprise data elements. While the government has an effort underway to align the ICs for the government, industry is faced with a proliferation of data definitions and the daunting task of mapping these into the enterprise and maintaining the maps for each IC and potentially each trading partner. For global enterprises, this can present a significant inhibitor to such concepts as the integrated enterprise, virtual enterprise, and global electronic commerce. For small businesses, the impact and costs associated with this proliferation cannot, and should not, be underestimated. When the technical challenges are coupled with the legal and competitive issues in a global electronic commerce environment, a simplified approach to EDI becomes highly desirable. Figure E-1, International EDI Today, depicts the EDI environment today and the manner in which X12 and UN/EDIFACT are managed. The translation function basically maps the implementation conventions for the two messages together. Most software products analyzed as part of the study provide independent support for ASC X12 and UN/EDIFACT, but not always translation between the two.

Figure E-1  International Electronic Data Interchange Today

A Way Forward
The question then becomes, how does the U.S. implement consistent EC/EDI data definitions and align them with the international communities? The analyses conducted for this study clearly suggest that the problem is not insurmountable. There are a finite number of data elements, and there already exists a significant degree of consistency. Implementation of a uniform data dictionary supporting electronic trading would provide a foundation for the types of commerce activities envisioned by ECAT and the National Information Infrastructure (NII) such as Interactive EDI, Electronic Catalogues, and Automated Trading. Supporting differences in EDI standards become less problematic because the underlying data can be made unambiguous. Message support becomes a matter of parsing the required data into the proper syntax. Our survey of existing software and services supporting EDI found a number of object and data mapping tools. These could easily support such an approach and could simplify software purchasing decisions by providing a standardized set of data definitions. Few vendors, however, provided "open" access to their internal data representations via Standardized Query Language (SQL) or application programming interfaces. With a standard Electronic Commerce Data Dictionary (ECDD), low-cost SQL parsers could be made available to parse the data into appropriate message syntax. The ECDD would simplify the ICs and the number of mapping tables required to support international EDI trading. The impact on the current EDI support tools and services would be minimal. The chart (Figure E-2), International EDI Tomorrow, depicts such an environment.

Figure E-2  International Electronic Data Interchange Tomorrow

Window of Opportunity
In the U.S. there exists a window of opportunity to implement a dictionary-driven electronic commerce environment. While this study analyzed two similar messages, they were chosen for covering a wide range of data elements also found in other messages and considered representative of the broader issue. The considerable overlap in definitions was not unexpected. As the U.S. is defining ICs at a national level, every effort should be made to align these at the data definition level with those of the international EDI community. There needs to be a directed effort in both the national and international business and standards EDI communities to recognize the need for this consistency and act to ensure it occurs. The U.S. should make every effort to achieve this objective and avoid the risk of a national infrastructure that is not compatible with international standards. This can be avoided by first aligning the U.S. data definitions with the international ones, and second by taking an active role in the UN/EDIFACT community to work through any discrepancies and promote a continuing course of alignment.

There has been considerable discussion and debate in the U.S. concerning the UN/EDIFACT organizational structure and processes. The standards development time, committee structure, rules and regionalized representation do not seem to be responsive to the needs of U.S. industry. The issue of binary support within EDI is representative of this debate. UN/EDIFACT for years would not acknowledge the business requirement, and once that was acknowledged, a technical debate ensued that seems endless. Business cannot afford to wait on these debates, and the U.S. has consistently and independently moved ahead. Both U.S. government and industry need to continue to pressure the UN/EDIFACT community to improve this situation. If UN/EDIFACT is to be truly international, the structure, processes, and rules may need to change.

In the United States, the federal government's implementation of EDI will have a great impact on EDI use nationally, and there is an opportunity to do so and include international alignment as a requirement. This should be done as a matter of high priority and urgency. Underlying the work being done at the National Institute of Standards and Technology (NIST) should be an Electronic Commerce Data Dictionary activity that supports the international alignment of data definitions and the on-going federal harmonization of ICs. This objective of alignment of data definitions can be an opportunity to promote and accelerate the adoption of UN/EDIFACT in the U.S.

The cost and risk associated with this approach are not nearly as significant as the cost and risk of not doing it. It is consistent with the activities and plans of the standards organizations, who would most likely welcome the activity and support it. It is consistent with the plans of ECAT and NIST, and will allow us to support multiple messaging standards as they are aligned. Most importantly, it will provide those key industries that have opposed the adoption of UN/EDIFACT in the U.S. a way to move forward without jeopardizing the significant investment they have made in X12 implementation.

Conclusions
In conclusion, the study found a high enough degree of consistency between the two standards in the data definition area to support a strategy of alignment. A national program to align the U.S. data definitions with the international UN/EDIFACT definitions will minimize the significance of message syntax and functionality difference, and should be undertaken as a priority activity. This strategy will allow the U.S. to support business today and build a foundation to rapidly move into the global trading arena. Without it we risk the costs and difficulties of translations and duplicative systems, one to trade in the U.S., and another to trade internationally. We recommend that the U.S. should not allow this to occur.

Document Overview
Table E-1 provides a road map to this document showing those parts of the document that respond to the requirements of the statement of work. The road map document consists of an executive summary, three parts, and six appendices. Part I includes an EC/EDI Implementation Road Map Summary. Part II contains a detailed discussion of the EC/EDI implementation issues for procurement functions, transition metrics, and the status of various standardization efforts within the International Organization for Standardization (ISO), the X12 and UN/EDIFACT committees, as well as the technologies and capabilities of the Internet to support EDI.

Part III contains a summary and the detailed analyses of the X12 850 Purchase Order, and the UN/EDIFACT ORDERS data elements. The appendices include the following supplementary information:

· Appendix A Glossary/Terms
· Appendix B Bibliography
· Appendix C EDI Commercial off-the-self (COTS) Products
· Appendix D Value-added Networks (VANs)
· Appendix E X12 Purchase Order and UN/EDIFACT ORDERS Segments
· Appendix F Implementation Milestone Chart
· Appendix G ANSI X12 Purchase Order (850) and EDIFACT ORDERS Data Elements
· Appendix H Translation Mapping Results for X12/EDIFACT

Table E-1  Road Map to the EC/EDI Road Map

1  "1994 Publications Catalog Incorporating an Introduction to EDI, "ASC X12S/94-172, Data Interchange Standards Association (DISA), January, 1994.

2  "Department of Defense Use of UN/EDIFACT Standards: A Readiness Study and Migration Strategy," Logistics Management Institute, May 1994, p.2-19.

3  "Streamlining Procurement Through Electronic Commerce," Draft, Federal Electronic Commerce Acquisition Team, April 29, 1994, p. H-8.