Configuration Management and Data Management

OVERVIEW

In this age of rapidly developing information technology, data management and particularly the management of digital data constitute an essential prerequisite to the performance of configuration management. Digital data is information prepared by electronic means and made available to users by electronic data access, interchange, transfer, or on electronic/magnetic media. There is virtually no data today, short of handwritten notes, that does not fall into this category. Configuration management (CM) of data is therefore part of data management activity; and management of the configuration of a product configuration cannot be accomplished without it.

Figure 10.1 is an activity model for configuration management of data. All the activities shown apply to configuration documentation. Most of the activities apply to all data. The model illustrates that the process is driven by business rules established based on the concept of operations for the processing of digital data, and specific data requirements.

Figure 10.1: CM-Related Data Management Activity Model

When the data process is initiated to create or revise an item of data, or to perform any of the actions necessary to bring it from one status level to the next , the various rule sets illustrated in Figure 10.1 are triggered to facilitate the workflow. The result is a data product with:

• Appropriate document, document representation, and data file identification
• Version control
• Clear and unambiguous relationships to the product configuration with which it is associated, and to the changes that delineate each configuration of the product

In addition, the data is available for access in accordance with contractually agreed-to rules for submittal, transmission, or online access (as appropriate), in the prescribed format (document representation) that can be used by the application software available to the authorized user .

CM RELATED DATA MANAGEMENT CONCEPTS AND PRINCIPLES

Configuration management principles ensure the integrity of digital representations of product information and other data, and enhance good data management practice. The concepts are described, as follows , based on elements and principles expressed in EIA Standard 649 (EIA-649):

• Document identification
• Data status level management
• Data and product configuration relationships
• Data version control and management of review, comment, annotation, and disposition
• Digital data transmittal
• Data access control

Document Identification

Each document reflecting performance, functional, or physical requirements or other product-related information must be given a unique identifier so that it can be:

• Correctly associated with the applicable configuration (product identifier and revision) of the associated item
• Referred to precisely
• Retrieved when necessary

Document identifier formats include all or most of the following parameters:

• Date
• Assigned numeric or alphanumeric identifier unique to the document
• Revision indicator
• Type of document
• Title or subject
• Originator/organization

A document is digitally represented by one or more electronic data files. Each document representation is the complete set of all the individual digital data files (e.g., word processor, CAD/CAM, graphics, database, spreadsheet, software) constituting one document.

As shown in Figure 10.2, the same document can have several different, equally valid representations, such as different word processing or standard neutral formats (IGES, ASCII, SGML-tagged ASCII). Any individual file (such as a raster graphics file, an ASCII file, or a spreadsheet file) may be part of several document representations of the same document/same revision; same document/different revision; or different document. The business rules relating documents, documentation representations, and files are as follows:

1. Each document iteration exists as one or more document representations, identified by:
   1. Document identifier
   2. Document representation identifier
   3. Document representation revision identifier
2. Each document representation is comprised of zero or more files. To facilitate the proper relationships, apply the following digital data identification rules to maintain document, document representation, and file version relationships:
   1. Assign a unique identifier to each file
   2. Assign a unique identifier to each document representation
   3. Assign a version identifier to each file
   4. Maintain, in a database, the relationship between:
      • Document identifier and its revision level
      • Associated document representation(s)
      • File identifiers and versions
      • Retain multiple versions of files as necessary to recreate prior document revisions and provide a traceable history of each document
3. Identify the tool and version of the tool (e.g., MSWord 2000) used to generate the document when the document is not in neutral format

Figure 10.2: Illustration of Document Representation Concepts

Data Status Level Management

Document status level is important as a foundation for the business rules defining access, change management, and archiving of digital data documents. It is the basis for establishing data workflow management and enhances data integrity. The standard data life-cycle model shows the data status levels (also referred to as states) that a specific document or document revision is processed through in its life cycle.

• Working is the status used to identify data (document representations or document revisions) that is in preparation ” a work in progress that is subject to unilateral change by the originator. Each design activity can define any number of subordinate states within the working category, to define the unique processes that different document types go through before release in their organization.
• Released is the status of document representations, and revisions thereto, that have been reviewed and authorized for use (such as for manufacture, or for submittal to, or access by a customer or supplier). Released data is under originating organization (for example, a contractor) change management rules, which prohibit a new revision of the document representation from replacing a released revision of a document representation until it has also been reviewed and authorized by the appropriate authority. The content of a document representation revision is fixed once it is in the released state. It is only changed by release of a superseding document representation revision. Once a document (or document revision) is in the approved state, changes are made only by release of a new document representation related to the next document revision.
• Submitted data is a proposed or approved document revision in the form of a released document representation that has been made available for customer review. This status applies only to data that requires submittal to or access by a customer (usually the government). If a submitted document revision that has not been approved is commented to or disapproved, a new working revision of the related document representation can be started and eventually submitted to replace the original document representation without affecting the identifier proposed for the new document revision. If a submitted document revision that has been approved is commented to, or disapproved by the customer, a new working representation of the next document revision can be started and eventually replace the original document revision.
• Approved is the status of documents and document revisions signifying that the data (document revision) has been approved by the Current Document Change Authority (CDCA) of the document. The content of a document revision is fixed once it is in the approved state. It is only changed by approval of a superseding document revision.
• Some tools include Archived as a data status for document representations and documents. This status is independent of the approval status (released, submitted, and approved) and merely means that the data has been removed from an active access storage mode.

No changes are allowed in the document representations that progress to the released state, or in document revisions that progress to the approved state. If there are changes to be made, they are accomplished by the generation and release or approval of a new revision. Documents must have at least one released document representation in order to be approved by the CDCA or submitted to a non-CDCA customer for review and adoption. Some data will exist only at the working level.

Business rules related to document or data status apply to each document type by defining requirements such as the following:

• Is submittal to (or access by) customer(s) required?
• In which application software and data format is submittal or access required?
• Who will be granted access privileges to the data in each of the applicable states?
• What are the approval requirements (reviewers/approvers) and method of approval (e.g., electronic signature) to promote a document to the released state?; the approved state?
• What are the archiving rules for this document type (e.g., all released versions upon release of a superseding version, all released versions, 90 days after release of a superseding version, etc.)?

Data and Product Configuration Relationships

A product data management system must provide an effective system to maintain the key relationships between digital data, data requirements, and the related product configuration so that the correct revision of an item of data can be accessed or retrieved when needed. Data files are related to documents via document representations.

Each product document, with a specific source, document type, document identifier (title, name , and number), and document revision identifier, may have the following relationships:

• Program/project or contractual agreement
• Contract data item identifiers
• Document revision/change authorization
• Associated product (hardware or software) name
• Associated product (end item), part, or software identifying number and revision/version identifier, where applicable
• The effectivity in terms of end item serial numbers for the associated product, part, or software item
• Status (working, released, submitted, approved, archived) of the data
• Associated data (document name, document title, document revision number, and date)
• Associated correspondence: document number, subject, date, references

The business rules for document retrieval should use these key relationships within a database to assure the integrity of the data that users can extract. Thus, information concerning a given product or part is associated with the configuration and effectivity (serial number) of the end item that uses the part.

This capability is particularly significant during the operation and support phase, when data is needed to support maintenance activity and to determine the appropriate replacement parts for a specific end item.

Data Version Control

Disciplined version control of data files is the prerequisite to effective electronic management of digital documentation and must be encompassed within the product data management software. Version occurs whenever a file is changed. The simplest form of version management is the file save feature incorporated in application software, which advances the file date and time identification each time a file is saved.

However, to retain the superseded version, it must be renamed . True version control business rules require automatic version identifier advance whenever a file is revised and not when the file is saved without change. Furthermore, they require all versions to be retained, subject to archiving guidelines and special rules pertinent to specific document types.

Because a single document representation can consist of many files, a very disciplined process is necessary to manage a document review process electronically . Version control rules facilitate the establishment of an audit trail of comments and annotations by reviewers, and the disposition of each comment. Each version of each document representation provided to, or received from, each reviewer is uniquely identified and associated with the source of the comment. Essentially, this means that a reviewer's version of a set of files (document representation) constituting a document being reviewed is renamed to enable the annotated comment copy to be distinguished from the official current version of the document.

Digital Data Transmittal

Part of the obligation of the sender of any document, regardless of transmission method, is to make sure that the document is in a format (document representation) that can be read by the receiver and converted to human-readable form. Appropriate identification is affixed to physical media such as floppy disks or tapes to clearly identify its contents. If all the file identifications cannot be included on the label, a directory, a reference to an accompanying listing or to a read-me file is used.

EIA-STD-649 lists the following common-sense guidelines for information to be provided to the user (via such means as "read-me" files, reference to standard protocols, online help), where applicable:

• Identification of the files included in the transfer by file name, description, version, data status level, application/file type, and application version
• Applicable references to associate the data with the basis (requirement) for its transmittal, approval, and payment, where applicable
• If there are multiple files, such as separate text and graphics, how to assemble each included data item for reading, review, or annotation, as applicable
• The naming convention for file versions and data status level distinguishes altered (for example, annotated or red-line/strike-out) file versions from unaltered files
• If and how changes from previous versions are indicated
• How to acknowledge receipt of the data, provide comments, and/or indicate disposition of the data digitally
• Time constraints, if any, relating to review and disposition

Data Access Control

Access to digital data involves retrieving the appropriate files necessary to compile the correct version of each digital data document, view it, and perform the prescribed processing. Seeking digital data access should be as user-friendly as possible. Users should be provided with data/documents they are entitled to in the correct revision/version. Before this can be accomplished, there are a number of pertinent parameters concerning access privileges, security, and protection of data rights that must be set up.

Access privileges limit access to applicable users. Access privileges vary according to the individual's credentials (security clearance, need-to-know, organizational affiliation , etc.), data status level, the document type, program milestones, and user need. Users of accessed data must respect all contractual and legal requirements for data rights, security, licenses, copyrights, and other distribution restrictions that apply to the data. The applicable distribution code, which represents the type of distribution statement, must be affixed to a document or viewable file to indicate the authorized circulation or dissemination of the information contained in the item.

Typically, working data should be made available only to the originating individual, group , or team (such as an integrated product development team); or to other designated reviewers of the data.

EIA-STD-649 provides us with the following checklist of ground rules to be preestablished prior to initiating interactive access (i.e., predefined query and extraction of data):

• How data is to be accessed
• Request for access and logging of access for read-only or annotation
• Naming of temporary working version of the file(s) for purpose of annotation/mark-up
• Means of indicating whether a comment/annotation is essential/suggested
• Re-identification of marked -up versions, as required
• Method of indicating acceptance, approval, or rejection , as applicable
• Time constraints, if any, on data acceptance
• Tracking of disposition of required actions
• Re-identification of changed files

SUMMARY

Software engineering methodology requires developers to carefully plan data requirements for all systems. This usually entails the development of a data dictionary for the databases and file structures required by the system. In companies where formal data dictionaries and repositories are in use, data management migrates easily to configuration data management. In companies where no formal data management processes are in place, the control of the flow of data through the organization must be rigorously addressed. This chapter serves that purpose by providing a very detailed structural framework for managing organizational data.

REFERENCES

This chapter is based on the following report: MIL-HDBK-61A(SE), February 7, 2001 , Military Handbook: Configuration Management Guidance .