الفهرس | Only 14 pages are availabe for public view |
Abstract With the advances in web technologies, more and more information on the web is contained in dynamically-generated web pages. Dynamic web applications are characterized by capabilities for personalization and frequently updating of contents. One of the most recent caching mechanisms that are employed to enhance the performance of Dynamic web applications, is fragment-based caching mechanism. In order to serve a website with this mechanism, an efficient method is needed for fragmenting the web pages automatically rather than manual fragmentation which is expensive, error prone, and unscalable. This thesis introduces an automatic fragments detection technique that is capable of identifying fragments, which are shared among multiple pages and are cost-effective cache units. This technique is suitable for work in large collections of web pages and is more accurate in detecting the maximal shared fragments. A system called ASFD was proposed to implement this technique. This system fragments the pages of a given website and detect the shared fragments among them. The system also collects statistics about the fragments such as the size, how many pages share the fragment,..etc. These statistics aid the web administrator to decide whether to turn on the automatic fragmentation (for the given website to be served with fragment-based caching mechanism) or not. The main contributions of this thesis can be summarized as follows: Constructing a modified Augmented Fragment Tree (AF Tree) that makes the candidate fragments appropriate for being stored in database for further comparison. Proposing a new algorithm for Automatic Fragment Detection which is more accurate in detecting the maximal shared fragments. Generating statistics about the detected shared fragments such as the size, how many pages share each fragment,… etc. The results that came out from the implemented system was very promising, when compared our ASFD system against the Buckets-Based system. The comparison has emphasized the fastness and the stability of our system whatever the number of the given pages was. |