2016/12/12 by S. Amir Hosseini, Hosseini, S. Amir, Shivendra S. Panwar +1
Computer Science · Decision Sciences · Engineering · #Advanced Bandit Algorithms Research #Advanced Wireless Network Optimization #Caching and Content Delivery #FOS: Computer and information sciences #Networking and Internet Architecture (cs.NI)
paper · pdf · doi:10.48550/arxiv.1612.03985
openalex publication_date 2016/12/12 · openalex created_date 2022/10/05 · openalex updated_date 2026/07/28
In this paper, we consider the problem of optimal scalable video delivery to\nmobile users in wireless networks given arbitrary Quality Adaptation (QA)\nmechanisms. In current practical systems, QA and scheduling are performed\nindependently by the content provider and network operator, respectively. While\nmost research has been focused on jointly optimizing these two tasks, the high\ncomplexity that comes with a joint approach makes the implementation\nimpractical. Therefore, we present a scheduling mechanism that takes the QA\nlogic of each user as input and optimizes the scheduling accordingly. Hence,\nthere is no need for centralized QA and cross-layer interactions are minimized.\nWe model the QA-adaptive scheduling and the jointly optimal problem as a\nRestless Bandit and a Multi-user Semi Markov Decision Process, respectively in\norder to compare the loss incurred by not employing a jointly optimal scheme.\nWe then present heuristic algorithms in order to achieve the optimal outcome of\nthe Restless Bandit solution assuming the base station has knowledge of the\nunderlying quality adaptation of each user (QA-Aware). We also present a\nsimplified heuristic without the need for any higher layer knowledge at the\nbase station (QA-Blind). We show that our QA-Aware strategy can achieve up to\ntwo times improvement in user network utilization compared to popular baseline\nalgorithms such as Proportional Fairness. We also provide a testbed\nimplementation of the QA-Blind scheme in order to compare it with baseline\nalgorithms in a real network setting.\n