Fairness Efficiency Scheduling for Cloud Computing with Soft Fairness Guarantees
ABSTARCT :
Fairness and efficiency are two important metrics for users in modern data center computing system. Due to the heterogeneous resource demands of CPU, memory, and network I/O for users’ tasks, it cannot achieve the strict100%fairness and the maximum efficiency at the same time. Existing fairness-efficiency schedulers (e.g., Tetris) can balance such a tradeoff elastically by relaxing fairness constraint for improved efficiency using the knob. However, their approaches are unawareof fairness degradation under different knob configurations, which makes several drawbacks. First, it cannot tell how muchrelaxedfairness can be guaranteed given a knob value. Second, it fails to meet several essential properties such as sharing incentive. 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 afairness sensitive scheduler that can maximize the system efficiency while guaranteeing the soft fairness by modeling the whole allocation as a combinationoffairness-orientedallocation an deficiency orientedallocation. Moreover, QKnober satisfies fairness properties of sharing incentive, envy-freeness and pare to efficiency given a proper knob value. We have implemented QKnober in YARN and evaluated it using both tested and simulated experiments. The results show that QKnober outperforms its alternatives DRF and Tetris by31.2%and4.5%, respectively.
EXISTING SYSTEM :
Fairness and efficiency are two important metrics for users in modern data center computing system. Due to the heterogeneous resource demands of CPU, memory, and network I/O for users’ tasks, it cannot achieve the strict100%fairness and the maximum efficiency at the same time. Existing fairness-efficiency schedulers (e.g., Tetris) can balance such a tradeoff elastically 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 such as sharing incentive. To address these issues,
DISADVANTAGE :
Resource allocation strategy using service level agreement (SLA) mainly depends on Response time, throughput and QoS parameters.
PROPOSED SYSTEM :
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 soft fairness by modeling the whole allocation as a combinationoffairness-orientedallocation and efficiency oriented 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 tested and simulated experiments. The results show that QKnober can achieve good performance and fairness.
ADVANTAGE :
Scheduling, load balancing, resource provisioning and traffic control are some of the ways of achieving QoS in cloud environment where, scheduling and load balancing goes hand in hand.
|