Hi , I am working on developing a distributed process scheduling algorithm for linux as a part of my master's thesis. Unfortunately I am a newbie to kernel programming and the project which I am working on is difficult for people who understand kernel scheduler well. Still I am very passionate about learning new things. My approach for designing a distributed linux kernel is to assume that there are no network connections involved in the processes, there are no open files,etc to make the job simpler. I am ready to take even more assumptions to get a simple basic implementation of a disitributed linux process scheduling algorithm. Other design aspect which I was thinking was to run each process in a container and migrate the process along with the container. To schedule the processes across distributed system, I was thinking of applying work stealing since it saves lot of communication costs and it is scalable. What help I need is to help me to integrate this distributed algorithm with the Completely fair scheduler algorithm. The book by Robert Love does explain nicely about Completely Fair Scheduler code but I wish to understand all the intricacies of the code. Any other design aspect is welcomed. Nitin
