Abstract | ||
---|---|---|
Video streaming is one of the most popular Internet services which may use thousands of servers. Current video streaming scheduling algorithms do not distinguish long streaming tasks from short ones which may result in sub-optimal energy consumption. In this paper, we observe that task length has strong correlations with user access profile, which can be used to predict the length of a given streaming task. Based on the predicted task length, we propose a series of heuristics algorithms that form a more power-efficient scheduling scheme. Experiments show that our approach is about 10 % to 160 % more power efficient than current scheduling approaches. |
Year | Venue | Field |
---|---|---|
2015 | ICA3PP | Power efficient,Scheduling (computing),Computer science,Video streaming,Server,Computer network,Heuristics,Energy consumption,Distributed computing,The Internet |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
17 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yunyun Jiang | 1 | 0 | 0.34 |
Tian Xiao | 2 | 17 | 1.36 |
Jidong Zhai | 3 | 340 | 36.27 |
Ying Zhao | 4 | 902 | 49.19 |
Wenguang Chen | 5 | 1014 | 70.57 |