Performance Improvement of a Front-end only Web

Abstract

In this fast-paced era where there are billions of active internet users worldwide, there is a huge demand for video streaming services, social networking sites, and communication using the internet. The Internet is being used for transactions, official email, or entertainment purposes. Users having short attention spans quickly move to different web pages if web pages lag by a few seconds, This results in loss of customers for e-commerce websites or loss of viewers in a video streaming platform. Many businesses are solely dependent on the internet for whom the loss of customers makes a huge difference due to slow websites. Hence, nowadays performance of websites is as important as the website itself. In this paper, we discussed how the performance of a website can be improved using performance enhancement techniques in the web application. There are many front-end only websites due to having Backend as a Service (BaaS) infrastructure and the ease of managing the backend using BaaS. We found there’s a performance gap in such websites that results in users leaving the website. Our work is based on how to improve this performance to retain website traffic. We have studied how this performance gap can be reduced and explored some of the topics mentioned widely in papers. Therefore, we discussed various techniques which can be applied in a front-end only application to increase the performance. Here we implemented a few techniques such as caching, pseudo streaming, image optimization, etc. on a completed video streaming web application and performed testing on the modified application, and saw how much the changes affected the performance of the website. We used average response time, Error rate, Maximum response time and throughput as performance metrics to measure the performance and then analyzed the result with the previous performance to figure out which techniques created more impact on the overall improvement of performance. Performance of the application was improved up to 17.13% in response time and 51.76% in throughput on a modules using the selected performance techniques. Also some relation with content size and performance was found in this research.