Pipeline Data Modeling and Process
Majdi Zahran Intergraph Corporation Huntsville, Alabama 35894 Email: mmzahran@ingr.com Abstract This paper focuses on how various enterprise-system components impact data modeling and discusses specific data-modeling factors to consider in order to achieve a seamless workflow environment, where business processes deal with provisioning and sustaining the pipeline and involve engineering, operations, construction, dispatching, mobile computing, and maintenance activities. I refer to this integrated environment as a Geospatial Resource Management (GRM) system. This paper will provide examples of requirements that impact the data model and basic steps that must be addressed prior to initiating an intense datamodeling effort. Introduction In an AM/FM/GIS, as with most other systems, data represents the major investment and needs to be leveraged and accessed at an enterprisewide level to support decision making, sales, marketing, operations, emergency response, customer care, and other critical needs and interfaces. An enterprisewide focus and data distribution will encompass a wide range of users with varying levels of needs and geographic experience. This gives rise to the issue of data modeling and how the designed logical and physical data model can support these different requirements and data relations. Where to start? Before initiating data modeling and design, it is important to identify the required components of the intended GRM system, including interfaces, business workflows and system requirements. If network analysis is part of the solution, then considerations need to be made for all features or facilities that impact or influence gas flow – what are they and the detailed attributes needed to perform network analysis? How will you obtain the necessary connectivity-based information and pass it to network analysis? Another data modeling consideration might include an interface to asset management, where attributes, such as asset ID’s, need to be associated with all features maintained within the asset management system. The issue then is how best to manage the status of the asset and other related information. Data modeling Once the GRM system and its components are defined, the requirements completed, and business workflows established, we focus on the data model. A frequently asked question is whether or not to normalize the data model. There are both advantages and disadvantages to normalizing. In fact, there is no magic formula or standard answer. Highly normalized data models will result in some performance degradation; however, zero normalization can result in unnecessary duplication of data and less efficient use of the database. Normalization needs to be evaluated and implemented with special consideration of the target system and its requirements. Logical Data model: An important prerequisite to logical data modeling is to have all the business requirements and workflows defined. The logical data model addresses blocks of information and how they relate based on established business workflows and constraints, independent of the database choice. An Entity Relationship Diagram (ERD) is the initial step in building the logical data model. There are data modeling tools available to assist in the construction and maintenance of the ERD. Physical Data model: Establishes full attribution of the model based on defined data structures with consideration for database of choice, implementation platform, client/server environment, data segmentation, user outputs, performance and other user needs. Pre-developed Models Today, leading software vendors and implementers offer pre-developed data models, which include pre-configured workflows and interfaces to popular third-party software applications. These “standardized” data models are the result of knowledge gained by vendors’ years of experience designing and implementing the integrated workflows demanded by clients. The advantages of a pre-configured model are many and include standardization, ease of implementation, and a robust starting point from which to arrive at a unique data model that meets your specific organizational needs. Conclusion Careful planning of the essential data model foundation is necessary to successfully implement a GRM and achieve the many possible quantifiable benefits of GRM. Through a properly developed data model, data can be efficiently exchanged between independent information systems and applications and interfaces can be added according to your technical, organizational, and financial priorities. The data model must match each company’s business strategy – that is, there is not one data model that will ensure success. However, there is a methodology to help determine the specific data model that will best meet your company’s objectives. | ||
|
|