This project aims to provide predictable latency between Xen guests on the same machine or on different hosts connected with a highspeed, low-latency connection. The planned execution of this project entails selecting or creating a queuing discipline, modification of the networking framework, and assistance with the theory required to ensure latency guarantees given the aforementioned modifications. Prior kernel work will be valuable but not required for a motivated student. Extra time would allow for extension to include high-latency or jittery connections between hosts.
This project has applications to realtime applications in the booming cloud industry. Examples includes hard real-time applications in industry and the military where timeliness of data or actions is of utmost importance. Other uses include soft real-time uses in just about any cloud computing or virtual machine application, for example ensuring adequately timely frames for streaming video and/or audio or the "killer app" of cloud computing, cloud-based gaming.
The student will gain more mastery with C/C++ programming, Linux kernel programming, and Linux networking. The student will collaborate on the project with a doctoral student and the faculty.