The art of RT-XEN

• Main Author: Lim, Zhi Cheng
• Other Authors: School of Computer Engineering
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Computer science and engineering
• Online Access: http://hdl.handle.net/10356/59134
• Institution: Nanyang Technological University

Computer virtualization had gained a lot of popularity over the years. It has become one of the most important assets to businesses and had helped many enterprises to save hardware cost, maximize operating efficiency as well as to achieve better integration across multiple operating systems [1]. Virtualization in computing refers to the method of sharing hardware resources across multiple operating systems on a single physical machine [2]. While ‘full virtualization’ [3] has evolved into the most common and successful methodology used in real world now, there are actually many more remarkable techniques that has been researched on, to improve virtualization experiences. The idea of virtualizing real-time operating system (RTOS) is one of many remarkable innovations. RTOS is designed to serve real-time operations, typically without missing a deadline. It is commonly used in embedded systems and applications that requires real-time functionally. The introduction of RT-XEN, originated from XEN (an open source virtualization platform), is one of widely used real-time hypervisors on the market [4]. Recently, It has launched a new multi-core virtualization platform for RTOS, which has proven to increase performance as a whole [5]. In this report, we will install RT-XEN in details, and understand its’ architecture behind it. Using Litmus-RT [6], we will create a guest virtual machine with different real-time tasks and configurations to determine RT-XEN’s functionally. It is found that the RT-XEN’s schedulability for real-time tasks is very dependable on the “ Control Domain” [2.3.1 – Control Domain (Dom0)]. Using a deterministic approach, the optimal Virtual Central Processing Unit (VCPU)’s parameters to be defined for guest virtual machine can also be observed for a random task created.