Real-Time Display of Dublin Traffic Information on the Web
TCD-CS-2000-46.pdf (PDF) 3.301Mb
Many large-scale and highly specialised distributed systems such as Urban Traffic Control (UTC) systems produce considerable quantities of data on a real-time basis. This data is then frequently locked in a proprietary format that makes it difficult to utilise for external purposes such as historical data analysis or additional end-user applications. The purpose of this dissertation is to examine the techniques and issues involved in extracting, processing and presenting such large quantities of data to a variety of end-user applications in a meaningful and scalable fashion. In order to gain further insights into this subject it was decided to develop an end-user interface for a real-life example of a large-scale distributed system that produces high volumes of data. The example chosen was SCATS, the city of Dublin?s UTC system. This system measures the traffic flows passing through several hundred junctions around the city and automatically adjusts the timing of the each intersection?s traffic lights on a minute-by-minute basis. Three central issues present themselves in this project. The first of these is the question of how to extract data from a proprietary system and convert it into a format that can be universally read and understood. The second issue is how an algorithm can be created that takes the available data and generates a meaningful measure of road traffic congestion at any given point in time. The third problem involves presenting such large quantities of data to a wide variety of end-users in a manner that is clear, meaningful and intuitive. The three tier application resulting from the research carried out during the course of the project is evaluated under several headings. Concurrency issues along with the performance, scalability and usability of each part of the implementation are discussed and evaluated. The congestion algorithm is compared against other measures that are available in SCATS, and formulae from the existing literature.
Author: Dineen, Mark
Availability:Full text available