Advances, Systems and Applications
Table 7 Makespan-aware schedulers and their properties
From: MapReduce scheduling algorithms in Hadoop: a systematic study
Reference | Year | Key Ideas | Advantages | Disadvantages | Comparison Algorithms | Evaluation Techniques |
|---|---|---|---|---|---|---|
Varalakshmi et al. [76] | 2022 | Optimized scheduling of multi-user Map-Reduce jobs in heterogeneous environment | Reducing the makespan, Improving resource utilization, near-zero wait time | Not considering data locality. Not considering adaptable dynamic environment | FIFO | Implementation |
Maleki et al. [77] | 2021 | SPO: A Secure and Performance-aware Optimization for MapReduce Scheduling | Minimizing makespan Maximizing security with risk-limitation constraint Traffic reduction | No particular | FIFO, HCS | Simulation |
Maleki et al. [78] | 2020 | TMaR: a two-stage MapReduce scheduler for heterogeneous environments | Improving the power consumption of cluster Minimizing the makespan of a batch of tasks Reducing the network traffic Considering heterogeneous environments | No particular | Hadoop-stock, Hadoop-A | Simulation |
Jiang et al. [79] | 2017 | Minimizing Makespan for MapReduce systems with different servers | Minimizing the makespan of servers | Not considering Heterogeneity of resources and jobs | FIFO, Fair, LPT | Simulation: cluster with 50 nodes |
Verma et al. [80] | 2013 | Using a heuristic, called BalancedPools to optimize the jobs execution | Improving makespan Increasing resource utilization | Not considering jobs with dependency, e.g., MapReduce workflow Not considering heterogeneity | Min strategy, Max strategy, MinSim, MaxSim | Implementation: on Hadoop cluster. using Hadoop 0.20.2 Simulation: SimMR |
Yao et al. [81] | 2015 | Using slot ratio between Map and Reduce tasks as a tunable knob to improve the makespan Using Workload Monitor and Slot Assigner to automate the slot assignment ratio | Minimizing makespan Increasing resource utilizations Dynamically allocates slots to tasks Considering the heterogeneity | Not considering data locality | FIFO | Implementation: in Hadoop V0.20.2 |
Zheng et al. [82] | 2016 | Proposing joint scheduling optimization of overlapping Map and shuffle phases | Minimizing makespan Reducing average job execution time Increasing resource utilization | Leading to an overly large job waiting time | MaxDiff, Pairwise, MaxShuffle, MaxSRPT | Conducting real data-driven experiments |
Tang et al. [83] | 2016 | Using an optimized MapReduce workflow scheduling algorithm | Minimizing makespan Low schedule length Increasing parallel speedup Higher efficiency Considering data locality | No particular | MRWS_NPI, SWS, WS_NWH | Implementation: in a Heterogeneous cluster |