Web Browsing Quality of Experience
Introduction to Web Browsing QoE
The ability to monitor web browsing quality of experience (QoE) is essential to determining when and where network conditions are contributing to an impaired user experience. Understanding the relationship between web QoE and network factors helps communications service providers (CSPs) to identify problems, to understand causes and contributing factors, and to evaluate potential solutions.
There have been recent attempts to approximate web browsing QoE by collecting and correlating data associated user behaviors. However, these approaches fall short for a number of reasons, and have only been considered as proxies for a real metric because the most important measurement of all—page load wait time—has been difficult for anyone, let alone CSPs, to obtain across the network and for each individual subscriber.
Measuring Web Browsing Quality of Experience
In this paper, we explore the challenges of monitoring web browsing QoE and propose requirements for those who wish to produce solutions in this space. We show that by meeting a relatively small list of specific requirements, a real-world solution can deliver meaningful insight into web browsing quality of experience.
Building a Web Browsing QoE-Monitoring Solution
Sandvine’s web browsing QoE solution allows CSPs to monitor, in detail, the web browsing experience of their subscribers.
The solution builds a comprehensive library of web page anatomy profiles, each associated with a web page that is either popular with subscribers or otherwise designated as important by the operator. Using these anatomy profiles, and combining with real-time measurements made as web browsing traffic flows through the network, the solution determines the time it takes to load each page.
This page load time measurement is turned into a web browsing QoE score (on a scale up to five) that corresponds to subscribers’ real-world experiences.
To maximize utility for the CSP, each web browsing QoE score is accompanied by more than 20 associated attributes that can be used to diagnose issues and to identify quality-related trends.