Weizhan Zhang

An overlay multicast protocol for live streaming and delay-guaranteed interactive media

In many collaborative multimedia applications, there is often a requirement for simultaneously supporting live streaming and shareable interaction. A major challenge in designing such an application by overlay multicast is how to simultaneously provide scalable live streaming and delay-guaranteed interactive media. Live streaming by overlay multicast incurs additional application-layer latency, which conflicts with the delay-sensitive property of interactive media. To handle this dilemma, in this paper, we propose a layered degree-constrained overlay multicast protocol, which organizes the overlay multicast tree as a layered degree-constrained core tree and an extended tree. The core tree maintains available resources in its top layers for subsequent two-way interaction, whereas the extended tree expands the core tree for one-way live streaming. Our simulation and experimental results show that the proposed overlay multicast protocol can simultaneously provide delay-guaranteed interactive media as well as scalable live streaming.

Tree-aware selective frame discard for P2P IPTV system on set-top boxes

Peer-to-peer (P2P) streaming provides an effective means to enlarge the service scale for set-top box (STB) based internet protocol television (IPTV). A major challenge of designing such a P2P IPTV system is how to provide high-quality P2P video streaming over resource constrained devices. In this paper, we present a P2P IPTV system, which tailors to support live video streaming on resource constrained STB devices. Specifically, we propose a tree-aware selective frame discard algorithm for the P2P IPTV system, which performs a differentiated frame discard strategy by taking into account the location of a STB node in P2P multicast tree. Simulation results demonstrate that our proposed scheme enhances the media quality of STB based P2P streaming on congested links.

Open-LTE: An Open LTE simulator for mobile video streaming

Simulation is the optimal means to evaluate the booming researches on how to enhance the end-to-end service reliability of mobile video streaming over LTE network. However, to the best of our knowledge, all existing LTE network simulators provide simulations of relatively closed virtual networks, in which only meaningless tracing data can be simulated being delivered. Researches on mobile video streaming have not yet been fully supported. In light of this, herein, the open LTE simulator Open-Sim is made to provide the simulation of virtual LTE network with the ability to connect actual hosts over real wired link in realtime. The transport and application layer related logics of video streaming can be deployed on remote hosts and will no longer be limited by the simulator framework. Open-LTE is thus compatible with experimental studies on most aspects of mobile video streaming. Open-LTE is simple to use by providing a centralized configuration file to set up the LTE channel fading scenarios and interconnect real traffic with the virtual LTE network. We will demonstrate our work with QoE experiments on a live video streaming application.

A behavioral sequence analyzing framework for grouping students in an e-learning system

Grouping of students benefits the formation of virtual learning communities, and contributes to collaborative learning space and recommendation. However, the existed grouping criteria are mainly limited in the learning portfolios, profiles, and social attributes etc. In this paper, we aim to build a unified framework for grouping students based on the behavioral sequences and further predicting which group a newcomer will be. The sequences are represented as a series of behavioral trajectories. We discuss a shape descriptor to approximately express the geometrical information of trajectories, and then capture the structural, micro, and hybrid similarities. A weighted undirected graph, using the sequence as a node, the relation as an edge, and the similarity as the weight, is constructed, on which we perform an extended spectral clustering algorithm to find fair groups. In the phase of prediction, an indexing and retrieval scheme is proposed to assign a newcomer to the corresponding group. We conduct some preliminary experiments on a real dataset to test the availability of the framework and to determine the parameterized conditions for an optimal grouping. Additionally, we also experiment on the grouping prediction with a synthetic data generator. Our proposed method outperforms the counterparts and makes grouping more meaningful.

Workload modeling for virtual machine-hosted application

We provide a workload model for VM-hosted application based on grey system theory.The correlation analysis approach is used to determine the bottleneck resource.Relation between the workload and the bottleneck resource consumption is modeled.We apply our model to a video-on-demand system as a case study. The workload complexity of virtual machine-hosted (VM-hosted) applications widens the gap between resource demand and allocation, which could lead to either resource underutilization or application overload. This study addresses this issue by providing a correlation analysis approach and a novel workload model for VM-hosted applications. Based on the grey system theory, the correlation analysis approach is used to determine the bottleneck resource of a given VM-hosted application, and the Workload Conversion Model (WCM) constructs a quantitative relationship between the application workload and the consumption of the bottleneck resource by modeling the resource consumption. We apply our model to a video-on-demand system as a case study. The model identifies that the VM memory is a key bottleneck resource of the system, and a bimodal relationship exists between the workload and VM memory consumption. The experimental results also show that the precision of the WCM is high, with a relative percentage error less than 5% and a variance ratio less than 25%.

A Segment-Based Storage and Transcoding Trade-off Strategy for Multi-version VoD Systems in the Cloud

Multi-version video-on-demand (VoD) providers either store multiple versions of the same video or transcode video to multiple versions in real time to offer multiple-bitrate streaming services to heterogeneous clients. However, this could incur tremendous storage cost or transcoding computation cost. There have been some works regarding trading off between transcoding and storing whole videos, but they did not take into account video segmentation and internal popularity. As a result, they were not cost-efficient. This paper introduces video segmentation and proposes a segment-based storage and transcoding trade-off strategy for multi-version VoD systems in the cloud. First, we split each video into multiple segments depending on the video internal popularity. Second, we describe the transcoding relationships among versions using a transcoding weighted graph, which can be used to calculate the version-aware transcoding cost from one version to another. Third, we take the video segmentation, version-aware transcoding weighted graph, and video internal popularity into account to propose a storage and transcoding trade-off strategy, which stores multiple versions of popular segments and transcodes unpopular segments. We then formulate it as an optimization problem and present a heuristic divide-and-conquer algorithm to get an approximate optimal solution. Finally, we conduct extensive simulations to evaluate the solution; the results show that it can significantly lower the storage and transcoding cost of multi-version VoD systems.

A version-aware computation and storage trade-off strategy for multi-version VoD systems in the cloud

Nowdays, many Video-on-Demand (VoD) providers offer multiple-quality video streaming services to heterogeneous clients, called as multi-version VoD. Some researches focus on video transcoding in real-time or video layered encoding/decoding, but they are not widely used in VoD industry. Storing multiple versions of the same video is an easy solution, but it consumes lots of storage space. Although there are also a few works about trading-off between transcoding and storage, they did not utilize the transcoding relationships among different versions and took the video popularity into account, which bring that they may have little cost-efficiency for multi-version VoD systems. To minimize the cost, in this paper, we propose a version-aware transcoding computation and storage trade-off strategy for multi-version VoD systems in the cloud. Firstly, it utilizes the transcoding weight graph to describe the transcoding relationships among different versions of a video. According to the graph, the transcoding computation cost from one version to another version can be calculated. Secondly, it takes the video popularity of different versions, the prices of storage and computation resources in the cloud into account to decide which versions of which videos should be stored or transcoded. We then formulate it as an optimization problem and present a heuristic approximate optimal solution. Finally, we conduct extensive simulations to evaluate our strategy and solution, and the results show that they can significantly lower the cost of multi-version VoD systems.

SAMP: supporting multi-source heterogeneity in mobile P2P IPTV system

Peer-to-Peer (P2P) streaming provides an effective method to deploy large-scale internet protocol television (IPTV) application over today¿s Internet. A major challenge in designing such an application is how to deal with content delivery among heterogeneous clients. Especially, with the development of the mobile internet and smart phones, how to provide an effective multi-source heterogeneous P2P IPTV solution becomes of great importance. Due to the special requirement of supporting differentiated multiple video streaming for different users, already existing P2P live video streaming techniques are inappropriate to be adopted directly. Therefore, this paper proposes a source-aware mobile P2P IPTV solution (SAMP) to provide multi-source live streaming for heterogeneous clients. In detail, SAMP includes a source-aware P2P topology construction algorithm to deduce the bandwidth occupation and the control overhead among the P2P clients, and contains a source-aware P2P streaming scheduling algorithm to further enhance the delay performance of the P2P IPTV system. The evaluation results show that the proposed SAMP scheme can provide a differentiated multi-source live streaming for heterogeneous clients with reduced bandwidth consumption, control overhead, and delivery delay.

Modeling and Predicting the Active Video-Viewing Time in a Large-Scale E-Learning System

Many studies of the mining of big learning data focus on user access patterns and video-viewing behaviors, while less attention is paid to the active video-viewing time. This paper pinpoints this completely different analysis unit, models the extent to which factors influence it and further predicts when a user permanently leaves a course. The goal is to provide new insights and tutorials regarding data analytics and feature subspace construction to learning analysts, researchers of artificial intelligence in education and data mining communities. To this end, we collect video-viewing data from a large-scale e-learning system and use the Cox proportional hazard function to model the leaving time. The models mainly include the interactions between variables, non-linearity assumption and age segmentation. Finally, we use the collected hazard ratios of model covariates as the learning features and predict which users tend to prematurely and permanently leave a course using efficient machine learning algorithms. The results show that, first the modeling can be used as an efficient feature extraction and selection technology for classification problems and that, second the prediction can effectively identify users’ leaving time using only a few variables. Our method is efficient and useful for analyzing massive open online courses.

CBC: Caching for cloud-based VOD systems

Cloud-based video on demand (VOD) service is a promising next-generation media streaming service paradigm. Being a resource-intensive application, how to maximize resource utilization is a key issue of designing such an application. Due to the special cloud-based VOD system architecture consisting of cloud storage cluster and media server cluster, existing techniques such as traditional caching strategies are inappropriate to be adopted by a cloud-based VOD system directly in practice. Therefore, in this study, we have proposed a systemic caching scheme, which seamlessly integrates a caching algorithm and a cache deployment algorithm together to maximize the resources utilization of cloud-based VOD system. Firstly, we have proposed a cloud-based caching algorithm. The algorithm models the cloud-based VOD system as a multi-constraint optimization problem, so as to balance the resource utilization between cloud storage cluster and media server cluster. Secondly, we have proposed a cache deployment algorithm. The algorithm further manages the bandwidth and cache space resource utilization inside the media server cluster in a more fine-grained manner, and achieves load balancing performance. Our evaluation results show that the proposed scheme enhances the resource utilization of the cloud-based VOD system under resource-constrained situation, and cuts down the reject ratio of user requests.