Mobile HTTP-based streaming using flexible LTE base station control

Abstract

This paper investigates the advantage of adopting a flexible resource control scheme when performing HTTP-based adaptive streaming across LTE systems. To guarantee video fluidity, mobile video streaming is known to require a large bandwidth overhead with respect to the net encoded video rate. The quality of a received video stream is impacted by variations in the size of the transmitted video packets (chunks), and by statistical fluctuations in the data rate at which the allocated downstream wireless channel operates. First, in considering an illustrative video scenario, we show that the chunk size distribution is heavy-tailed, and is well fit by a Gamma distribution. Second, we employ a HAS based proxy video manager and resource controller at the base station node. Based on the channel quality observed and reported by a mobile client, the manager selects the proper channel bandwidth and data rate levels at which to transmit the stream's chunks, in accordance with the selected encoded video rate and the configured Quality of Experience (QoE) level at which the user is targeted to receive the video stream.The communications data rate is also set to assure an acceptable low video reception stall probability. To illustrate the performance of such a dynamic bandwidth allocation scheme, we compare it with an operation that employs a stationary setting of the channel bandwidth, and we compute the gain achieved when such adaptations are performed at the base station node on a chunk by chunk basis. We show by analysis, and confirm by simulations, the improvements achieved in the system's performance behavior through the use of the adaptive resource allocation scheme

CLEVER: A Cooperative and Cross-Layer Approach to Video Streaming in HetNets

“…Each version is encoded for reception at a prescribed targeted video quality level, and is accordingly characterized by its specific flow bit rate. Previous works (e.g., [7], [8]) have shown that by using resource allocation schemes that are executed by a manger that resides at the cellular eNB side or at a central Resource Manager (RM), one can provide effective QoE support of video streaming at the video users while simultaneously providing for efficient utilization of available wireless resources. In this work, we extend this approach by considering a heterogeneous cellular layout (HetNet) that includes micro and macro cells.…”

“…We have then considered different quantization levels (namely 25, 30 and 35) corresponding to the quality levels q = 2, 3 and 4. Moreover, the videos are characterized by different encoding bit rates and different Peak Signal to Noise Ratios (PSNRs), 7 which are summarized in Tab. 5.…”

“…Let us notice that even under the constant rate constraint, the video chunk sequence exhibit large fluctuation in size. In fact, as detailed in [7], the chunk size tends to follow a heavy tailed (e.g. Gamma) distribution, corresponding to a not negligible probability of occurrence of very large chunks.…”

“…The complexity of this operation is O(M × N ). Then, users may be offloaded to the MeNB (lines[5][6][7][8][9][10][11][12]. This operation, repeated for all the µeNBs, has a complexity of O(M ×N 2 ).…”

Joint Adaptive Rate and Scheduling for Video Streaming in Multi-Cell Cellular Wireless Networks

“…Encoder settings and resulting video rate R(q) relevant to quality q are summarized in TABLE I [21].…”