Traffic modeling for WWW | |
| Author | Panjai Tantatsanawong |
| Call Number | AIT Diss. no.CS-00-01 |
| Subject(s) | World Wide Web Data Traffic Management System (Computer system) Web servers |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Engineering, School Advanced Technologies |
| Publisher | Asian Institute of Technology |
| Series Statement | Dissertation ; no. CS-00-01 |
| Abstract | An analytical modeling approach is introduced in this research to capture actual characteristics of the WWW traffic. The model consists of two submodels, one for the client side and the other for server side (WWW server or proxy cache server). The workload on the client side can be characterized by type of the requested pages (HTML, inlined image and CGI) and on the server side by the requested file types. Zipf's law is applied to investigate the popularity of the workload classes. From workload characteristics of the WWW sites considered, the WWW server can be classified into two types: a static WWW server, which provides static information WWW pages and an interactive WWW server, which provides real time processing such as multi-tier or E-commerce applications. The client and the server models are synchronized with hourly time intervals. On the client side, serveral TCP performance effects are included in the model such as TCP connection setup (three-way handshake), persistent connection and slow-start algorithms. On the server side, modeling and forecasting of hourly requested transactions are made using a combination of Box-Jenkins models with seasonality removal. Fourier analysis na standardization are used to render seasonal data into non-seasonal data. However, seasonality can not be completely removed by this approach, so both seasonal and non-seasonal models are employed and the best models are selected using posterior possibility criterion (PPC). In order to improve the forecasting accuracy, forecast values obtained by the developed models are adjust6ed hby the difference percentage of the day of the week and by using this percentage adjusted values in place of the actual data, which are not available at the time of forecasting. With this adjustment, the developed models can provide very accurate forecast values, except for the case wherer the higher level of hierarchical proxy is shared by shared by child proxies located in a different time zone. Finally, multiple-class closed and opened queueing network models are developed to estimate response time at the client side and the server side respectively. The proposed model can satisfactory predict the impact on the client response time and the network utilization as a function of HTTP (version 1.0 and 1.1), implementing proxy cache, hierarchical proxy cache and improving the cache with a prefetch mechanism. |
| Year | 2000 |
| Corresponding Series Added Entry | Asian Institute of Technology. Dissertation ; no. CS-00-01 |
| Type | Dissertation |
| School | School of Advanced Technologies (SAT) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Computer Science (CS) |
| Chairperson(s) | Kanchana Kanchanasut; |
| Examination Committee(s) | H.N.Phien;Tabucanon, M.T.;Anuchit Anuchitanukul;Yamaguchi, Suguru; |
| Scholarship Donor(s) | The Royal Thai Government; |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2000 |