Edge Computing Assisted Adaptive Mobile Video Streaming

Abstract

Nearly all bitrate adaptive video content delivered today is streamed using protocols that run a purely client based adaptation logic. The resulting lack of coordination may lead to suboptimal user experience and resource utilization. As a response, approaches that include the network and servers in the adaptation process are emerging. In this article, we present an optimized solution for network assisted adaptation specifically targeted to mobile streaming in multi-access edge computing (MEC) environments. Due to NP-Hardness of the problem, we have designed a heuristic-based algorithm with minimum need for parameter tuning and having relatively low complexity. We then study the performance of this solution against two popular client-based solutions, namely Buffer-Based Adaptation (BBA) and Rate-Based Adaptation (RBA), as well as to another network assisted solution. Our objective is two fold: First, we want to demonstrate the efficiency of our solution and second to quantify the benefits of network-assisted adaptation over the client-based approaches in mobile edge computing scenarios. The results from our simulations reveal that the network assisted adaptation clearly outperforms the purely client-based DASH heuristics in some of the metrics, not all of them, particularly, in situations when the achievable throughput is moderately high or the link quality of the mobile clients does not differ from each other substantially. Index Terms-Server and network assisted DASH, multi-access edge computing (MEC), quality of experience, fairness, load balancing, integer nonlinear programming (INLP), greedy scheduling algorithm

Energy-Aware QoE and Backhaul Traffic Optimization in Green Edge Adaptive Mobile Video Streaming

“…Toward satisfying the requirements of 5G networks, mobile edge computing (MEC) concept has been proposed by the European telecommunications standard institute (ETSI) which enables moving the contents to the edges of the network nearby the end users [5]. Authors in [13] proposed the edge computing assisted system for DASH video streaming with the objective of jointly maximizing the QoE of the clients, fair bitrate allocation and balancing the utilized resources among multiple base stations. Along with MEC, the video content caching and retrieval at the network edge within the radio access network (RAN) has been shown to be a promising solution to alleviate significantly the traffic burden on the backhaul network [6], [14], [15], [16].…”

“…The available resource blocks at each time slot in any base station indicates the allocated bandwidth in the frequency domain based on the achievable throughput of the client and its assigned bitrate according to LTE 3GPP specifications [23], [25], [24]. For the sake of low complexity in the performance evaluation and following relevant research works [13], [26], the resource allocation to the clients at the base station is performed at every one-second time slot in our system model. However, our system is easily adoptable to smaller time scaling such as subframe without any modification to the model and the proposed solution.…”

“…Constraint (11) ensures that no stalling happens in the client's buffer during its whole video streaming session and (12) determines the clients to server mapping. Constraint (13) ensures that the client downloads its chunk from only one location at each time slot and (14) states that at each time slot, the transrating operation can be performed in only one edge server. Constraints (15)- (16) determine the range of decision variables.…”

“…The estimated throughput and the switching threshold δ Sw , which is used to control the switching level of the allocated bitrates to the consecutive chunks of the client, are then computed (line 4). Threshold δ Sw is computed knowing that the highest switching between the consecutive chunks of video happens when the bitrates are allocated merely based on the buffer level [13]. A fairness threshold δ F a is also used to control the fairness value in allocating the bitrates to the client with respect to other simultaneous clients at the same time slot.…”

Edge Computing: Applications, State-of-the-Art and Challenges

“…By defining a series of intrusion rules to determine intrusion behavior, the system has a high detection rate when detecting external attacks. In order to handle large-scale network data and application access to manage control traffic in the cloud, Gul et al [40] proposed a multi-threaded distributed intrusion detection system model in 2011. In this model, the cloud intrusion detection system can process large-volume data packets, analyze them, and generate reports efficiently.…”