Scholarlink

<p>The importance of e-learning in modern education systems is increasing, and Server-Side Rendering (SSR) is making a comeback as a popular technique for building websites with dynamic content. This paper examines a real-world e-learning platform utilizing SSR and assesses its performance and environmental impact. An initial analysis of the platform’s usage was conducted to replicate a realistic workload during which we captured realistic user behavior. Since existing load-generation tools do not support all the requirements of modern websites, we developed a custom load-generation tool built on top of established open-source tools. We use our load-generation tool to emulate virtual users, using captured user behavior as a basis. The load generator we developed is very low-maintenance and allows developers to easily test different configurations of a hybrid website without having to adapt the load generator. Therefore, thanks to our flexible load generator, we were able to measure the resources and energy consumption of the real-world e-learning platform under varying virtual user loads in different configurations. The results showed that using a caching mechanism significantly lowered system load compared to a system without caching. Regardless of whether caching was enabled, external network traffic remained nearly identical as all required resources (database, static files, etc.) were hosted on the system. Energy consumption was lower with the cache enabled, aligning with other system resource usage patterns. Specifically, our system would produce 0.161t <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\hbox {CO}_2\)</EquationSource> </InlineEquation> emissions at idle in a year. Under permanent load with 600 concurrent users, the system would produce 0.192t <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\hbox {CO}_2\)</EquationSource> </InlineEquation> with caching enabled and 0.203t <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(\hbox {CO}_2\)</EquationSource> </InlineEquation> without caching.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Server-side performance and environmental impact evaluation of a hybrid e-learning website

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月20日

查看全文
  • Thomas Prantl,
  • Luis Wolf,
  • Jonas Schiller,
  • Marco Pruckner,
  • Samuel Kounev,
  • Tobias Dänzer,
  • Jóakim von Kistowski

摘要

The importance of e-learning in modern education systems is increasing, and Server-Side Rendering (SSR) is making a comeback as a popular technique for building websites with dynamic content. This paper examines a real-world e-learning platform utilizing SSR and assesses its performance and environmental impact. An initial analysis of the platform’s usage was conducted to replicate a realistic workload during which we captured realistic user behavior. Since existing load-generation tools do not support all the requirements of modern websites, we developed a custom load-generation tool built on top of established open-source tools. We use our load-generation tool to emulate virtual users, using captured user behavior as a basis. The load generator we developed is very low-maintenance and allows developers to easily test different configurations of a hybrid website without having to adapt the load generator. Therefore, thanks to our flexible load generator, we were able to measure the resources and energy consumption of the real-world e-learning platform under varying virtual user loads in different configurations. The results showed that using a caching mechanism significantly lowered system load compared to a system without caching. Regardless of whether caching was enabled, external network traffic remained nearly identical as all required resources (database, static files, etc.) were hosted on the system. Energy consumption was lower with the cache enabled, aligning with other system resource usage patterns. Specifically, our system would produce 0.161t \(\hbox {CO}_2\) emissions at idle in a year. Under permanent load with 600 concurrent users, the system would produce 0.192t \(\hbox {CO}_2\) with caching enabled and 0.203t \(\hbox {CO}_2\) without caching.