Out of the basement: Moving from mapping to enterprise GIS
Shirley S. Sanger1 & Robert W. Finkle2 1Assistant Director, Information Systems and Services City of Fort Worth, 1000 Throckmorton Street, Fort Worth, Texas 76102 SangerS@ci.fort-worth.tx.us 2President, IT Nexus, Inc. 10129 W. Dartmouth, Suite 9304, Lakewood, Colorado 80227 rwfinkle@it-nexus.com
Enterprise GIS: An old promise
In the early 1990’s we started to refer to GIS as “Enterprise GIS.” While this term has begun to be a buzz-word used to refer to any method of implementing GIS, at its heart Enterprise GIS suggests the delivery of some very specific capabilities. In our view, critical capabilities associated with Enterprise GIS include:
Ironically, since the very early days of GIS the user public expected to receive this vision of Enterprise GIS, but after investing years implementing GIS found themselves with an automated mapping system that queued map plot requests. Why? Well, up until the last 6-8 years GIS technology simply did not possess the technical ability to deliver Enterprise GIS. For example, until the early 1990’s GIS software could not support a “logical view” of a database – the ability to access and report data from multiple database tables – it could only report from a single data table. Without this capability, it was technically impractical to use GIS as a spatial integrator of multiple information systems; it even made using GIS for anything more than a map-plotting tool difficult. The good news is that several GIS vendors now provide all the capabilities needed to implement Enterprise GIS – a GIS that provides desktop access to multiple departments and that can serve a pivotal role in integrating data across an organization’s disparate information systems. The bad news is that the gap that existed for almost 20 years between user expectations for GIS and what GIS could technically deliver has resulted in at least three obstacles to realizing the promise of Enterprise GIS:
The city of fort worth's experience Background By 1996 the City of Fort Worth, Texas had over 10 years experience with GIS, having implemented an automated mapping system in the early 1980’s. As is typical to GIS implemented in that time period, the system provided map maintenance and map plotting services to a wide range of City departments, but failed to meet the needs and expectations of many departments. As a result, while the City supported an “enterprise” automated mapping system, several departments implemented dissimilar GIS software that better addressed their data analysis needs. In 1996 the City reorganized its central data processing department and brought in a new generation of information technology (IT) managers. The new IT Director, Michael Di Paolo and his Assistant Director Shirley Sanger established what is now know as the City’s Information Systems and Services (ISS) Department. Their mandate was to improve support for the City’s information technology needs and operations. Di Paolo won early approval for dramatic moves, two of which were considered central to achieving a contemporary enterprise-computing environment:
All large IT projects, at a minimum, must have two essential “keys” if they are to achieve success: top management approval and an appropriate plan. This was particularly true for the City of Fort Worth’s Enterprise GIS project. The magnitude of the project required that it be endorsed at the highest levels of City management to insure that resources (both staff and money) would be available. To achieve this critical endorsement GIS was defined as an essential “enabling technology” and presented to the City Manager and City Council along with other ISS strategic initiatives, including the establishment of citywide networking. In addition, ISS management pre-defined some key requirements and objectives for their Enterprise GIS project:
The GIS project was officially started in the spring of 1997 with a high-level needs assessment. Since the City of Fort Worth did not have the staff or the expertise to prepare this document a GIS consulting firm was hired. An important benefit to hiring a consultant was that a neutral, 3rd party expert was much better positioned to achieve inter-departmental consensus on potentially contentious issues related to project phasing and data ownership. After interviewing various departments, a GIS Strategic Plan was created that projected time frames, project phasing and dollar amounts needed to complete the foundation of the enterprise GIS. It was decided that this foundation would provide the GIS data common to a majority of City work functions (these common features were defined as parcels, lots, street centerlines, and zoning) and establish a master GIS data server that all departments could access. The planning document was necessary to present to upper management and City Council so that money could be appropriated to proceed with the project. The development of the high-level strategic plan took approximately two months to complete. The next step was to complete the GIS Strategic Plan by preparing a detailed needs assessment and implementation strategy that involved all seventeen (17) of the City’s departments that would use the GIS. Each department’s by-in for maintaining the various levels of data in the enterprise wide GIS was defined in the detail plan along with the sources of that data. GIS applications were also identified, described and prioritized on a citywide basis. This step was completed in the summer of 1997. With a detailed plan in place the next critical step was the preparation of a database design. Key objectives of the City’s Enterprise GIS are integration with existing city information systems and ensuring the database supports a wide range of City work functions. A formal database design process was absolutely essential to the successful realization of these two objectives. Key elements of the City’s database design process include:
A data migration specification to achieve these objectives was prepared and an RFP for data migration/conversion was released December 1997. A data conversion vendor was selected and GIS database construction started in June of 1998. Data migration and conversion of the first phase of GIS data components was completed in the summer of 1999. During this time, the City of Fort Worth worked closely with its Tarrant Appraisal District to ensure that the most current parcel data was converted. The Tarrant Appraisal District (TAD) is an important player in the City’s overall Enterprise GIS system concept. TAD is responsible for maintaining and providing the City digital updates to their parcel data. The City is responsible for providing TAD digital updates of City approved platted land (subdivision lots). Currently, ISS is helping its departments implement key GIS data maintenance applications to update the GIS database. ISS is also actively involved in assisting departments to plan, design and develop additional components of the GIS database – data elements specific to the City’s utility infrastructure, parks, police, planning, and development services. While the City’s ISS Department is responsible for managing and administering the GIS database, its user departments are responsible for developing their unique data elements, for the maintenance of these data elements and for the purchase or creation of applications to support their work functions. The ISS Department will play a pivotal role in technically coordinating and orchestrating this on-going development. Critical issue in achieving enterprise GIS Implementing Enterprise GIS is not simply a matter of selecting and installing software. It requires a concerted effort that involves strategic planning, organizational development, proper design, implementing requisite computing infrastructure and establishing appropriate institutional mechanisms. The following discusses what we consider are essentials for the successful implementation of Enterprise GIS. The Enterprise GIS Strategic Plan Preparing a strategic plan for Enterprise GIS is a first and highly important step. The strategic plan has several important goals: developing multi-department consensus on priorities and responsibilities, winning management approval, defining the tasks and technical strategy for implementing the system and preparing a planning level budget estimate. Achieving consensus and management approval is essential, and often difficult to accomplish using in-house resources. As a result, using a respected, experienced consultant to develop the strategic plan is often a good idea. The strategic plan must define a technical strategy and develop a planning-level budget for approval. This requires a level of detail that is often missing from many strategic plans. A plan that only states at a high-level desired outcomes or makes a list of the software and data elements that must be developed is not enough to deliver a sound technical strategy and budget estimate. Important elements that should be developed as part of a strategic plan include:
Ironically, while the selection of GIS software was so critical during the 1980’s it is less so now because of an evolving adherence to standards and the increasing maturity of GIS software for supporting the demands of enterprise computing. The technical importance of using only a single GIS software product and hardware platform is also less important to achieving operational success than it used to be. While there is a cost to managing a heterogeneous GIS software and hardware environment, many Enterprise GIS systems around the country use GIS software from multiple vendors (see “Managing a Non-centrally Funded GIS: How the City of Fort Worth and Other U.S. Cities Make It Work,” by Finkle and Sanger in these proceedings). What is of critical importance to the success of Enterprise GIS is the availability of high bandwidth, local and wide are (LAN and WAN) network infrastructure. The City of Fort Worth has implemented a switched Ethernet 100 mbs backbone and a Sonet OC3 140 mbs ring that establishes WAN connectivity between City office buildings located in the downtown area. Connectivity with large outlying buildings is establishes using T-1 or Frame Relay and ISDN is used to connect small buildings. Local area networks serving individual departments use a mix of fiber and category 5 cable, operating at 10 mbs, with plans for upgrading to 100mbps. In today’s high demand, client-server computing environment, these are minimum requirements for good performance. Database Design Achieving Enterprise GIS requires an investment in a formal, deliberate and committed database design process. GIS implementations of the past have suffered from either the complete lack of a database design or a “monolithic” approach to database design that believed the database design not only had to be exhaustively comprehensive but that is was also immutable. The truth is, the Enterprise GIS database design must be thorough, well documented, incrementally modified and continually updated as new points of integration and new applications are encountered and the physical database is adjusted. This reality is well accepted in the traditional IT world, and in part explains why database designers and administrators are so well paid – it is an essential and on-going requirement to the operation of a sophisticated, enterprise-wide information system. The initial database design should follow formal IT information systems engineering methodology. Database design should consider all priority applications (business functions) the database will need to support and make sure it provides for data integration with any external information systems that are targeted as priorities. The resultant design is indispensable to the data conversion contractor because without it there is no targeted outcome and this makes developing reliable data conversion cost estimates nearly impossible. The database design then should become a living document that is kept up-to-date to accurately describe the content, meaning and structure of the physical database. To do this the database design document must also provide a data dictionary that explains the meaning and purpose of data elements and data relationships. There is nothing more frustrating to an application or system integration programmer than to discover the old database design document lacks a data dictionary or no longer reflects the physical database. Before the fun of programming can start, the design document must laboriously researched and updated. Building the Enterprise Institutional Mechanisms Successfully implementing Enterprise GIS also requires a lot of attention and time be spent in creating and using formal institutional mechanisms for consensus development, coordination, decision-making and approval. At the planning stage of the project this means using representative committees and perhaps techniques brought by an experienced consultant to achieve consensus on development objectives and priorities, and to win funding approval for the project. But this is only the start. On-going methods of communication and decision making must be established and put into routine use. Another critical institutional requirement is that the organization implementing Enterprise GIS provides funding for the staff needed to support the central functions of an Enterprise GIS. A review of several, established Enterprise GIS systems around the country reveals that a wide range of responsibilities and funding methods exist for these central, enterprise support functions (see “Managing a Non-centrally Funded GIS: How the City of Fort Worth and Other U.S. Cities Make It Work,” by Finkle and Sanger in these proceedings). But essential support functions commonly assigned to a central Enterprise GIS support group include:
Today, achieving Enterprise GIS is much less an issue of the technical capabilities of GIS software. Instead, it is very much an issue of proper planning, design, technical execution and institutional formation. In this paper we have related the successful approach taken by the City of Fort Worth to implement its Enterprise GIS. We have also attempted to extrapolate from this experience the critical decisions, methods and steps that we would recommend other organizations consider as they implement for the first time an Enterprise GIS system or embark on the expansion of their existing automated mapping system to realize the benefits of Enterprise GIS. | ||
| © GISdevelopment.net. All rights reserved. |