Scientific African (Mar 2024)
Modelling the response rate of Apache web server using extreme value theory
Abstract
Delay in response is associated with servers used in processing data particularly when there is large number of users. Consequently, the performance of web servers plays a crucial role in the digital ecosystem, necessitating analysing their tail behaviours. Models such as queue models, neural networks have been used for modelling the response time of web servers that are characterised with heavy-tailedness. Such models cannot sufficiently describe the behaviours in the tail area of the distributions. Therefore, a proper model is required to fit these extreme responses to have a comprehensive insight into server performance. In this study, an explorative data analysis was conducted on a selected web servers, and the Generalized Extreme Value (GEV) and Generalized Pareto Distribution (GPD) models were fitted to the data. The outcome of the study showed practical applicability of extreme value models to model tail events of response time of a web server relative to the existing studies in literature that ignored extreme behaviours of server response rate. The study highlights the point at which the referenced server would break down, which could account for a total breakdown of work in an organization. Website operators should leverage on these analyses and results to device measures to improve server performance and prevent breakdowns. This study recommends the adoption of extreme value models for modelling web server behaviours, providing valuable insights for enhancing server performance through hardware, software, or configuration adjustments. Additionally, proactive measures like load balancing, alerts, and performance monitoring are suggested to avert extreme events and ensure a reliable and efficient web server operation. These findings contribute to optimizing web server performance and ultimately elevating users' experience and site reliability in the digital landscape which aligns with the Africa's Union's Agenda 2063 (Goal,2) and Sustainable Development Goals (SDGs, 9) that underpins education and skills revolution with advanced technology and innovations.
