Web applications are collections of files for the purpose of responding to user requests in client-server mode over the World Wide Web. Most web applications have a connection to one or more databases for storing and delivering data as required by user requests. Some include a web server extension. There are a number of such systems, each belonging to a set of archetypes. Important factors in selecting a web application include conformance to principles of software engineering, ability to enforce application-level security, and resource conservation or high performance. Existing experimental studies do not provide the required data, owing to variations in configuration, aggregated, statistically mixed workloads, and lack of numerical performance criteria for the application alone. This research gained important insights from the application of scientific methods and numerical, objective criteria to the comparison of web applications. This project conducted architectural analyses, security tests and performance measurements of selected archetypes. This work assessed state saving, separation of concerns, security and performance with standardized tests administered in an invariant environment, i.e., host hardware, operating system, web server and database server. Architectural analysis was applied to state maintenance and separation of concerns. Interactive simulations tested abilities of archetypes to enforce security and to protect themselves from hacker attacks. Automated load tests measured resources consumed and application latency as a function of HTML request traffic. HTML page requests consisted of a standard set of queries against the same dataset. A load test application was developed to manage the testing and data-recording automatically. Results consist of rankings of web application archetypes according to a metric developed for this purpose. Supporting data consist of graphs of latency vs. time and page request traffic vs. resources consumed. Results show the capability of archetypes to enforce security, maintain data integrity, conserve server resources, and to make timely and efficient delivery of client-requested data
| Identifer | oai:union.ndltd.org:nova.edu/oai:nsuworks.nova.edu:gscis_etd-1441 |
| Date | 01 January 2003 |
| Creators | Cane, John W. |
| Publisher | NSUWorks |
| Source Sets | Nova Southeastern University |
| Detected Language | English |
| Type | text |
| Source | CEC Theses and Dissertations |
Page generated in 0.0022 seconds