Year 2000 Compliance as Repository Justification
There is little doubt that the millennium change will bring a software crisis. The question is whether the looming catastrophe will be completely out of control or to a large extent manageable. The
more details that are known about the impact ahead of time, the greater the likelihood that the catastrophe can be managed.
The good news for data administrators–who have been repository advocates for many years–is that along with the coming millennium change, the time for repository has arrived. Repository can
support Y2K compliance efforts; Y2K can support repository. There has never been a better time to convince your organization to invest in a repository.
However, this synergy is not widely recognized. A search of the Year 2000 Web site FAQ page yielded no hits on “repository”.
Year 2000 compliance is a given in all organizations (this is more or less true in most businesses) and the amount of time remaining to become Y2K compliant is getting shorter as the big date
creeps closer! The major steps toward Y2K compliance can be summed up as:
- Impact assessment
- Change strategy selection
- Coding of changes
- Testing
- Production Implementation
Repositories can play a major role in impact assessment, testing, and implementation. For example; Y2K necessitates comprehensive system inventories. While compiling that inventory, it makes sense
to save the results into a centralized resource (i.e., repository) that can yield benefits not only throughout the Y2K effort, but in data administration and system support well into the new
century.
Tools abound for parsing and analyzing code and operating system control language, searching for potentially-affected components. The sequential files which are the common output from
scanning tools need to be recreated en masse if they are to be kept in synch with changes in the production environment. A repository maintenance process can be put in place to recognize and apply
incremental changes to persistent meta data.
If your Y2K inventory effort is making use of parsing tools, the task of identifying meta data sources-JCL libraries, proc libraries, copybook libraries, etc.-may already underway as well. This is
work the repository administrator can build on.
One way to mesh Y2K and repository projects is to reformat the output from the Y2K scanners and load the output into the repository making use of “batch load” utilities provided by repository
vendors. This is a “quick-payback” introduction to repository which may provide significant value (until a long-term repository solution hooked into production change control is completed). This
type of simple database can get Y2K analysts and developers “hooked on repository” through the analysis of meta data. In addition, queries against the meta data tables can be used to validate the
contents of the repository as it is loaded.
Year 2000 Impact Assessment
After source code and JCL code inventories are completely loaded into the repository (easier said than done), repository functions can be used to identify system components that may be at risk for
Y2K non-compliance. Repository software such as Platinum’s Scan COBOL can support this effort. Once code and data element changes are identified, the impact of changes to programs and dataset’s
extends to batch jobs and transactions. Y2K impact begins at the data element level (where the century digits were omitted). However, the impact of changing that date may extend up to one or more
entire applications. Repository cross-reference functions can be used to trace the impact of a change from bottom to top (element to table to program to job) to identify all components directly or
indirectly impacted by a change.
Year 2000 Change Management
As part of the Y2K compliance process, components (such as programs and files) that are shared among applications need to be identified. The processes which use these components must be changed,
tested and moved into production in a controlled, synchronized manner. A comprehensive system inventory stored in a repository can enable the identification of cross-application dependencies.
Repository cross-reference functions can be used to develop comprehensive “work units”, i.e., sets of components which interact and whose changes must be tested and implemented in synch.
Project Tracking
Several levels of management are likely to require regular reports on the progress of the Y2K compliance effort. A repository that is populated with a comprehensive inventory of system components
and their
relationship to programmers, business units, software packages, etc., can support status reporting from the most detailed level (example, the number of date data elements identified and/or changed)
to the highest level (e.g., applications which have or have not begun Y2K efforts). The data model below shows how meta data about system component types and management structures are related.
Vendor Strategies
At this time, there are two primary players in the large enterprise repository market: Platinum and R&O.; Platinum intends to focus it’s Year 2000 support strategy on their newly-acquired
Year-2000-specific products, including Adpac’s SystemVision. It is the author’s opinion that a lack of information meshing Y2K and repository projects has made it difficult for Platinum to
articulate a clear strategy for repository use for Y2K support. Details about the interface between Adpac’s SystemVision scanners and the Platinum Repository are difficult to obtain. The Year 2000
page at the Platinum Web site contains no mention about repository; the Platinum Repository page mentions nothing about Year 2000. In their favor, Platinum does have a white paper describing how
their repository can support Year 2000 activity. This paper is very good but it is only available by request from the vendor (i.e. it is not available through Platinum’s Repository or Y2K
web-sites).
R&O;’s web-site, upon most recent review last updated in June of 1996, states that they had a major project underway to release a Year 2000 product for the Rochade repository in Q3 1996.
Evidently, R&O;’s acquisition in December of 1996 by Viasoft (who has their own Y2K strategy) has caused some change in this direction. Their current plan is to announce a Y2K/repository
product called Enterprise Portfolio Manager (EPM) at their User Conference June 1.
To summarize, vendor acquisitions and mergers in the repository marketplace have resulted in delays and confusion for customers and potential customers interested in repository Y2K support.
Summary
There is a crucial need for persistent meta data to support the management of Y2K compliance efforts. An unprecedented window of opportunity exists for data and repository administrators to offer
an effective solution for this need. The alternative is a proliferation of incompatible, unconnected and redundant MS Access databases and Excel spreadsheets. Expect spotty support from vendors.
Personal initiative and ingenuity is called for-and could well be rewarded.
William J. Lewis has spent twenty years in the Information Technology field, the last thirteen managing data, meta data and data models. His work has appeared in previous releases of The Data
Administration Newsletter, in DM Direct, Database Programming and Design, and IDUG Solutions Journal.
