AN ENSEMBLE ALGORITHM ON P2P COMMUNICATION AND NETWORKING

ABSTRACTEdge cameras are ubiquitous, together with the recent boom in computer vision technology, and a variety of video analytics tasks are being processed at the edge. It is challenging to support more complex video analytics tasks on edge servers with unpredictable request loads and limited resources. However, most of these works use only a single optimization approach, focusing only on the improvement of a certain performance metric in a single processing stage, ignoring the balance of other performance metrics, and the space available for optimization is often very limited. Especially when dealing with video analytics tasks that need to be divided into two GPU‐CPU stages for completion, this unidirectional focus may lead to execution performance imbalance or even negative quality of service (QoS) optimization. In addition, to fully utilize the valuable resources on the edge servers, it is often necessary to schedule multiple types of video analytics tasks on the edge servers. However, most of the existing scheduling strategies only focus on how to allocate computational resources for end‐to‐end tasks. They lack the awareness and consideration of the execution of tasks in different execution stages, as well as the mutual interference among tasks. These scheduling strategies, lacking stage‐sensitivity and interference‐sensitivity, may cause performance conflicts in environments running multiple tasks involving GPU‐CPU dual‐stage processing, thus affecting the overall QoS. To address these challenges, we first evaluate the impact of batch processing, frame rate control, resolution selection, and CPU concurrency processing on throughput, latency, and accuracy when running dual‐stage tasks on edge platforms. Then, we propose DualRT, a soft real‐time video analytics framework for dual‐stage tasks, to optimize the QoS of dual‐stage tasks while avoiding request stacking on edge platforms. In the scheduling module of DualRT, we design a scheduling method using a multi‐agent deep reinforcement learning algorithm and a variable time window approach to schedule multiple dual‐stage tasks with joint control of batch size, resolution, frame rate, and CPU concurrency for each task. Our experimental results show that DualRT improves QoS by an average of 13.3% and maximum throughput by an average of 24.6% compared to state‐of‐the‐art solutions.

Video/Audio packet transport over IP (ver. 4) networks is not guaranteed for real-time communication if network congestion occurs due to excessive traffic load. Congestion avoidance control is thus necessary for real-time video/audio communication over IP networks. An adaptive video bit rate control scheme that detects the onset of congestion is designed to achieve real-time visual communication and efficient data transport. Congestion over a CSMA/CD LAN is detected from the number of retransmissions caused by collisions. The packet delay is used to detect congestion over connected LANs. The effectiveness of the adaptive video bit rate control scheme is verified in experiments that use a number of video/audio communication terminals and data terminals on IP networks. The interaction of the adaptive video bit rate control and TCP congestion control schemes are investigated and suitable bit rate control scheme is discussed. The adaptive bit rate control scheme expands the possibility of visual communication over IP networks.

The support for multiple high-definition video streams in wireless home networks requires appropriate routing and rate control measures, ascertaining the reasonable links for transmitting each stream and the rate of the video to be delivered over the chosen links. In this paper, we invest the combination of the routing and rate control in a united convex optimization formulation and propose a distributed joint solution based on cross-layer design. We first develop a distortion model which captures both the impact of encoder quantization and packet loss due to network congestion on the overall video quality. Then, the optimal joint rate control and routing scheme is realized by adapting its rate to the time-varying traffic and minimizing the overall network congestion. Furthermore, simulation results are provided, which demonstrate the effectiveness of our proposed joint routing and rate control scheme in the context of wireless home networks.

Scalable video coding (SVC) has become more and more important with the enrichment of multimedia data and the diversification of network and terminal devices. In current MPEG SVC activities, a scalable extension of H.264/AVC, called scalable video model (SVM), is proposed by HHI and has shown further coding efficiency improvement and scalability functionality. However, the SVM model doesn't provide an efficient rate control scheme now, and rate control is achieved through a full search for selecting a suitable quantization parameter (QP). That is very inefficient and much time-consuming. In this paper, an efficient rate control scheme is proposed for the SVM, which is derived from the state-of-the-art hybrid rate control schemes of JVT with some considerations for scalable video coding. In the proposed rate control scheme, the rate distortion optimization (RDO) involved in the step of encoding temporal subband pictures is only implemented on the low-pass subband pictures, and rate control is independently applied to each spatial layer. For each spatial layer, the rate control is implemented at GOP, picture and basic unit levels. Furthermore, for the temporal subband pictures obtained from the motion compensation temporal filtering (MCTF), the target bit allocation and quantization parameter selection inside a GOP could make full use of the hierarchical relations inherent from the MCTF. The proposed rate control scheme has been implemented into SVM3.0 and experiment results show that the proposed algorithm can achieve the target bit rate with little bit rate fluctuation and keep fine image quality at the same time, but the computation complexity is reduced heavily.

Rate control schemes for panoramic video coding

There are several ways to improve the quality of experience (QoE) of mobile user equipments (UEs), among which latency reduction has been paid less attention compared to throughput enhancement. In this paper, we consider a new cellular network architecture that puts edge servers closer to UEs (e.g., nearby BS) and present a suitable rate control algorithm, so that we can achieve both lower controllable latency and better fairness without any throughput loss, compared to traditional rate control schemes such as TCP. In the proposal, BS decides the queue threshold to keep a desirable queue length and delivers congestion notification based on the queue differential to the edge server, so the edge server can adjust its rate to reflect wireless channel capacity while maintaining a target queueing delay at BS. Our scheme is beneficial due to (i) UE-agnostic approach and (ii) potential suitability to 5G network architecture under standardization by 3GPP. Via NS-3 simulation we demonstrate that our scheme outperforms popular TCP variants in terms of latency and fairness metrics in several network scenarios.

A rate control (RC) algorithm is a primitive block of video encoders that fulfills the bandwidth and buffer constraints for given channel and application properties. State-of-the-art RC schemes perform inefficient in terms of buffer and quality smoothness when handling varying rate distortion characteristics of different picture types (I, P, B) and different MBs in one picture (e.g. bright, textured, static/moving MBs) while posing a high computational overhead. In this paper, we propose a novel RC scheme that covers GOP, picture/ slice, and basic unit levels. It treats different picture types (I, P, B) in a non-linear fashion with consideration of whether they are referenced or non-referenced pictures. Our novel RC scheme prioritizes Macroblocks depending upon their spatial and temporal characteristics (considering eye-catching regions) for refined Quantization Parameter allocation. Compared to RC-Mode-3 (i.e. the latest RC Mode in JM reference software), our RC achieves up to 77.8% and 72.4% reduced buffer-and quality fluctuations, respectively. Compared to RC-Mode-0, our RC provides 2.97dB (i.e. 7.2%) better PSNR for the mixed Susie sequence. Moreover, our proposed RC is 16.6× faster than the RC-Mode-0 when executing on Intel Core2Duo T5500 (1.66 GHz).

The authors consider network utility maximisation problem in orthogonal frequency division multiple access (OFDMA) networks to study cross-layer, fair and efficient resource allocation. Assuming knowledge of the instantaneous channel gains, this problem is decomposed into rate control and scheduling problems at the transport and medium access control/physical layers, respectively. In contrast to the rate control problem that is solved using subgradient method, the scheduling problem has high computational complexity owing to optimising integer and continuous variables simultaneously. Based on the results from analysing the integer relaxed scheduling problem, computationally efficient adaptive scheduling (CEAS) and opportunistic time division multiple access (Opp-TDMA) scheduling schemes are proposed to joint subcarrier assignment and power allocation. Simulation results demonstrate that aggregate utility achieved in the network with the cooperation between rate control and proposed scheduling schemes outperforms those of previously proposed joint channel-aware and queue-aware scheduling schemes. Also, through comparison with the optimal solution, the authors conclude that CEAS is applicable for OFDMA real-time scheduling due to low computational complexity and high performance.

We propose a scheme for efficient transport of video over wireless networks using a novel rate control scheme. The explicit bit rate (EBR) control specifies the use of a variable rate channel and provides an adaptively varying requisite bandwidth to carry varying video encoded output. Due to the use of prediction and variable length codes (VLC) for video codecs, variable-bit-rate (VBR) encoding is considered logical in all advanced video encoders, such as H.263 and H.264. The VBR method of rate control which is currently employed in 3G networks requires the use of a constant bit rate (CBR) physical channel to carry video transmissions to ensure a fixed bandwidth channel to the user during a video call. However, a constant bit rate channel can result in delivering video frames in unequal time intervals, delays in transmission and poor video quality at the end user. We show the application of an explicit bit rate (EBR) control mechanism for high efficiency and quality video transmission for both cdma2000 1x and WCDMA wireless networks.

More load, more differentiation — Let more flows finish before deadline in data center networks

Rate control is a central part of standard video codecs. Many rate control schemes including some for JVT video coding (MPEG-4 AVC /H.264) employ a quadratic formulation of the rate-distortion (R-D) function. The rate control for JVT video coding is rather complex. In this paper, we analyse the problems existing in the quadratic R-D model and in the recent rate control scheme for JVT video coding. According to the analysis and experimental results, we present an improved quadratic R-D model and a frame-layer rate control scheme for JVT video coding based on the model. This quadratic R-D model is utilized for QP determination in rate control scheme, where the model parameters are estimated using statistical linear regression analysis. The model is also used for computing the starting quantization step (QP) for the first frame in the sequence. The Experimental results show that due to the accuracy of the improved quadratic R-D model, the rate control is quite effective, where the generated bit rates are very close to the target bit rates while achieving good R-D performance. And the starting QP is very close to the average QP, by which the performance is also improved.

Rate control scheme plays a vital role in the circumstances of region-of-interest (ROI) based video coding, aiming to allocate a higher bit rate to the ROI while setting a relative lower bit rate ratio in non-ROI to maintain the total bit rate close to the assigned target value. Among various rate control models, the R-Lambda based rate control model outperforming other models. Unfortunately, the R-Lambda model fails to reflect the content of video, and the consequence is a downgraded achievement of the rate control. This paper is proposed to design a new rate control scheme based on the saliency map, aiming at extending saliency based rate control algorithm to both frame level and coding tree unit (CTU) level. Experimental results indicate that the proposed bit allocation structure provides an improvement in the region of interest by employing the peak signal-to-noise ratio (PSNR) quality and the subjective assessment.

We consider the utility-maximization problem in the downlink of wireless mesh networks with orthogonal frequency-division multiple access (OFDMA). We formulate this problem as a cross-layer design of joint rate control and OFDMA scheduling to efficiently utilize the scarce radio resources. The problem is decoupled into a rate-control problem at the transport layer and a channel-aware and queue-aware scheduling problem at the Media Access Control/physical layer (MAC/PHY). The rate-control problem adjusts arrival rates to the base station (BS) queues, and the scheduling problem determines link rates, i.e., departure rates from the BS and other network nodes, through subcarrier and modulation rate assignment. Although the rate-control problem is locally solved at the BS, we propose a greedy algorithm that solves the scheduling problem in a distributed manner, at network nodes. Furthermore, we propose a heuristic algorithm for fast execution of the scheduling scheme at individual nodes. Numerical results show that the heuristic algorithm performs comparably with the greedy algorithm, whereas it has lower computational complexity. In addition, our proposed scheduling scheme, when it cooperates with the rate-control mechanism, improves the network performance in terms of end-to-end delay, aggregate utility, and fairness.

Effcient video compression and improving quality of video in communication for computer endcoding applications

This paper considers optimal rate control and routing schemes for multipath networks which can be formulated as multipath network utility maximization problems. In these schemes, maximizing the aggregated user utility over the network with multipath routes under the link capacity constraints is the objective of utility maximization problems. By adopting the Lagrangian method, sub-problems for users and paths are deduced and interpreted from an economic point of view. In order to obtain the optimal rate allocation, a novel distributed primal-dual algorithm is proposed, and the performance is evaluated through simulations under two different fairness concepts. Moreover, window-based flow control scheme is also presented since it is more convenient to realize in practical end-to-end implementation than the rate control scheme.