Please note: This master’s thesis presentation will be given online.
Ashraf Abdel-hadi, Master’s candidate
David R. Cheriton School of Computer Science
Supervisor: Professor Samer Al-Kiswany
We present Falcon, a novel scheduler design for large-scale real-time data analytics and cloud services. Falcon achieves the scheduling accuracy of modern centralized schedulers while supporting clusters with hundreds of thousands of nodes. At Falcon’s core is a novel scheduler design that leverages modern programmable switches. Falcon supports a FIFO scheduling policy, as well as data locality-based, task priority-based and task resource constraint-based policies.
Our prototype evaluation on a cluster with a Barefoot Tofino switch shows that the proposed approach can reduce scheduling overhead by 120 times and increase the scheduling throughput by 100 times compared to state-of-the-art centralized and decentralized schedulers.