Delivering Remote Desktop Services to Remote Users: A QoE Perspective

Organizing a university exam has become increasingly difficult in recent years due to new requirements introduced by modern university courses to provide students with modern working environments. All this is coupled with the usual logistical problems. The COVID-19 pandemic, moreover, introduced the necessity to move services, tools, and resource-intensive programs from university laboratory workstations to students’ personal devices. Born to solve this intricate puzzle, the CrownLabs project is a modern cloud-native technology allowing the delivery of remote desktop services to a heterogeneous plethora of devices by offloading the program execution, hence the resource-related requirements, to remote servers. The project was born during the COVID-19 pandemic in 2020, by a group of students and volunteers at the Polytechnic University of Turin. The initial intent was to provide a remote-laboratory service to students to let them practice from home. In 2021 it was enriched with scalability capabilities, making it suitable for a larger number of users, and enabling it as an exams platform. With respect to other production-grade Virtual Desktop Infrastructure (VDI) products, CrownLabs is the only solution extended to be well-integrated with the wider exam ecosystem. This thesis work consists of a CrownLabs exams QoE extension. It aims at providing additional observability tools, allowing students and examiners to track the remote environment’s health status. The implemented module enriches examiners’ experience with extended control over the exam instances with real-time metrics to track performances and connections throughout the instances. At the same time, students' experience is now improved with non-invasive tools to understand at any time the quality of their connection to the remote system and if they are properly using the available resources. The monitoring infrastructure is designed to be well-integrated within the existing CrownLabs architecture and open to extensions. The infrastructure includes a Metrics Collector server to retrieve metrics of interest, an Aggregator providing complete pieces of information to the client, and a variety of Frontend views exposing metrics to clients. Following the approach used so far, the module's implementation makes use of cutting-edge cloud technologies and development techniques.