Database Integration: Criteria and Techniques
Tom Lonski
Senior Technical Consultant, UGC Consulting, 6200 S. Syracuse Way, Suite 222,Englewood, CO 80111
(303) 773-6166,(303) 773-6618 (fax)
Email: Tom. Lonski@ugc.com
Abstract
Database integration is a key challenge for AM/FM/GIS projects since almost all projects have existing
legacy systemswith data needed by the AM/FM/GIS. A large and confusing array of combinations,
approaches and products exist for accomplishing this integration. The paper wil I define six general
techniques that encompass al I of these combinations, address various integration issuesand analyze these
techniques using ten specific criteria that affect a project’s performance, capabilities and cost.
Introduction
Only a short time ago AM/FM/GIS systemswere almost solely used by the engineering or drafting
organizations within a utility. However, with the re-regulation and increasing competition in the utility
industry, uti Iities are under increasing pressure to gai n an advantage by operating more intelligent y as a
business. As a result, many utilities are faced with a common challenge on their AM/FM/GIS projects. To
gain that competitive edge, utilities must fully realize the benefits of their AM/FM/GIS project by making
the faciIity and geographic information available throughout the company to both internal organizations,
such as customer service, marketing, construction, operations, and system planning and, external
organizations, such as subcontractors or customers, who haven’t used or had access to the AM/FM/GIS
information in the past. Likewise, information from those other organizations is needed within the AM/FM/GIS.
The effect of this goal to ful Iy distribute the AM/FM/GIS information is that utilities now frequently face the
prospect of integrating their AM/FM/GIS system with other computer systems. In practical terms, “sharing
information” from the AM/FM/GIS system with other internal or external organizations at a utility means
integrating the database(s) in the AM/FM/GIS system with the database(s) in the computer systems used by
those other organizations. A functional Iisting of some of the computer systems at a uti Iity that may be
integrated with an AM/FM/GIS includes: Customer Information System, Materials Management System,
Planned Maintenance System, Computer-Aided Dispatch, One Call, Human Resources Information System,
Outage Management System, Work Management System, Property Accounting, Supervisory Control and
Data Acquisition, Engineering Analysis/System Planning, Accounts Payable, Non Utility Billing Internet or
Intranet Web Server, and Land Management Systems. Note that this is a functional listing and that the
various systems above may be combined or divided as actual Iy implemented at a particular utility.
The Integration Jungle
As utilities tackle these integration requirements, our experience has been that they often find themselves in
a quandary over how to identify the best integration solution and implement it. One reason the integration
problem can seem daunting is that many factors conspire to compl icate the integration problem. Some of
these complicating factors are:
- A large number of potential interfaces can be implemented with the AM/FM/GIS.
- The different computer systems are usualIy on different hardware platforms.
- The different hardware platforms may use different network protocols.
- The different computer systemsoften use different database management systems (DBMS) for managing
their data. Different DBMSes use different data models-some support the relational model, some use
an object model and some use a hierarchical model. Some of the computer systems may not use a
- modern DBMS, instead they use an indexed file system such as VSAM with a transaction monitor such
as CICS.
- There is no standard terminology that can be used to discuss and compare integration solutions.
- There area plethora of commercial products available not only from the DBMS vendors and the
hardware vendors, but also from third-party vendors.
- The same information is stored in multiple databases because existing databases are not integrated and
these copies do not match. A common example is mismatches between the set of customer service
addresses in a customer information system and the set of customer service addresses in an
engineering/operations database.
One approach to the database integration problem is to begin surveying the products available for
integration as the first step toward building a Iist of possible solutions. Once al I of the possible solutions
have been identified, they can then be compared and the best solution chosen for implementation.
However, our experience has been that this bottom-up approach does not usually work well. Typically
numerous products are avai Iable to integrate databases and this approach can turn into a prolonged
research effort unless arbitrary Iimits are placed on what vendors and products are considered. Another
drawback to this approach is that the integration products frequently use different database standards and
different network protocols. By focusing on the integration products, it is difficult to determine what
database standards and what network protocols will provide a coherent environment for integrating the
util ity’s computer systems. Whi Ie a solution designed by a bottom-up approach may initial Iy work, there
usually isn’t a high degree of confidence that a solution designed this way is a good one, i.e. that it meets
al I of the functional needs, that it minimizes the cost, that the integration productltechnology can be reused
and that the integration productitechnolo~ is flexible/general/expandable enough to support future
requirements.
These kinds of difficulties are common whenever a bottom-up approach is used for a problem in computer
science. Instead, the conventional wisdom in computer science advocates the use of a top-down approach.
In the course of wrestling with integration problems on a number of projects, we have defined ten general
integration criteria to help focus on the functionality required in the database integration and six general
integration techniques that can be used to classify the possible integration solutions. Using these criteria
and techniques provides a top-down approach to the integration problems. By focusing on the integration
functionality, planning and designing a solution is simplified. The integration criteria can be used to
identify the appropriate integration technique as wel I as evaluate specific integration products. Once an
integration technique is chosen, then the number of products that need to be examined in the search for
candidate solutions is reduced. Another way the integration problem is simpl ified is, once an integration
solution is identified, the network and communication part of the problem is typically simplified because
the products used for the integration often specify the required protocols and networking.
