Fairness-Efficiency Scheduling for Cloud Computing with Soft Fairness Guarantees

Fairness and efficiency are two important metrics for users in modern data center system. Due to the heterogeneous resource demands of CPU, memory, and network I/O for users' tasks, it cannot achieve the strict 100% fairness and the maximum efficiency at the same time. Existing fairness-efficiency schedulers(e.g., Tetris) can balance such a tradeoff by relaxing fairness constraint for improved efficiency using the knob. However, their approaches are unaware of fairness degradation under different knob configurations, which makes several drawbacks. First, it cannot tell how much relaxed fairness can be guaranteed given a knob value. Second, it fails to meet several essential properties. To address these issues, we propose a new fairness-efficiency scheduler, QKnober, to balance the fairness and efficiency elastically and flexibly using a tunable fairness knob. QKnober is a fairness-sensitive scheduler that can maximize the system efficiency while guaranteeing the \theta-soft fairness by modeling the whole allocation as a combination of fairness-purpose allocation and efficiency-purpose allocation. Moreover, QKnober satisfies fairness properties of sharing incentive, envy-freeness and pareto efficiency given a proper knob value. We have implemented QKnober in YARN and evaluated it using both testbed and simulated experiments. The results show that QKnober can achieve good performance and fairness.