一种共享存储视频服务器集群的请求调度策略

在VOD服务器集群中,对用户服务请求的合理调度是提高集群整体性能的关键技术之一.本文针对共享存储结构下的服务器机群,在请求调度算法LoadCache-rep基础上进行改进,提出一种基于视频节目点播集中度的调度策略,该策略通过将相近的点播请求调度至相同服务器上以充分利用服务器的缓存机制,同时兼顾均衡各服务器间的负载.并根据实时负载变化对请求进行迁移以消除VCR操作对负载分布的影响.仿真试验表明,该策略能有效提高视频服务器集群的运行性能.     

视频流调度策略的研究

视频流调度策略是视频点播系统中有效使用存储资源和网络资源的关键技术。该文研究了已有的多种流调度技术并分析了各种调度方法的优劣，在此基础上以补丁算法为基础提出了一种支持零延迟交互式操作的调度方法，并实验正明了这种方法在用户延迟、带宽消耗等方面的优越性。     

无线多跳网络上基于网络编码的多源多播算法

考虑到在无线多跳网络中存在多个相互独立的多播流的调度问题,首先提出了一个基于网络编码的动态多源多播模型,并根据该模型设计了一个适用于802.11DCF MAC的多播算法MulSrc。MulSrc以队列积压作为流调度的主要依据,以网络编码作为数据传输和本地恢复的主要手段,无固定的传输路径,在无线网络中实现以丰富的多路径来克服相关丢失、降低延迟的目标,适用于无线网络中大量存在的随机丢包情况。实验证明,在存在多个多播流的网络上,MulSrc可以取得比同类协议CodeCast和MMForests更好的性能。     

基于中央服务器模型的蓝牙多播技术

为解决蓝牙多播技术在移动智能终端应用的问题,分析蓝牙协议的特点及其通信机制,综合考虑通信设备的数量和设备间通信时间的延迟等因素,提出一个基于中央服务器的多播通信模型。通过改善蓝牙的基带层协议,打破传统蓝牙协议仅能支持7个设备的限制,实现一个可以同时支持更多设备进行通信的中央服务器设备,其对设备的调度采用的是基于活跃度的平等调度算法 EASA （equality of active scheduling algorithm）,该算法是对 RR （round Robin）调度算法的改进,以活跃度为指标,保证调度的公平性,缩短活跃度高的设备间通信时间的延迟。实验结果表明,该模型可以支持多个设备,保证通信的实时性。     

一种新的QoS覆盖多播路由协议的研究

研究了QoS覆盖多播路由问题，对度约束模型进行扩展，提出了一个新的支持QoS路南的覆盖多播网络模型．基于此模型，提出丁一个新的QoS覆盖多播路南协议QOS覆盖多播树协议（QOMTP）．该协议采用分布式和树优先的策略，使多播组成员之间能自组织地构建一棵基于源的、满足QoS约束的覆盖多播树．该协议采用了一种新的启发式局部优化算法，通过调节启发因子，能灵活地在延时和带宽之间进行均衡．仿真实验表明，采用按请求带宽转发媒体流，并选择适当的启发因子，QOMTP协议能减少多播树上节点的网络资源占用越，同时获得较高的节点接纳率和较好的动态适心性，从而证明了该协议机制及其算法的有效性．     

采用新的流调度策略的视频点播系统研究

客户多播补块流调度策略——CMP(Client Multicast Patching)——是基于代理缓存的有效的视频点播策略，它采用用户参与流调度的方式，某些用户在播放节目的同时，又作为服务器为后续用户多播一部分节目，因此，在不增加用户I／O带宽的前提下，通过利用局域网络丰富的带宽．有效地节省了服务器及主干网络的资源，并对用户请求提供即时服务，该文给出CMP方案的实现算法，并对系统进行仿真及性能分析，结果表明了CMP方案的有效性和可行性。     

IP／DWDM光因特网中的一种支持柔性QoS的并行一体化多播路由算法

本文研究了IP／DWDM光因特网中支持柔性QoS的并行一体化多播路由算法。对IP／DwDM光因特网中的多播请求及用户提出的端到端延迟需求区间，提出的算法一体化地解决路由选择和波长分配问题。目标是在考虑网络负载均衡的前提下，寻找一棵费用次优的多播树，并且满足用户QoS需求。该算法基于粗粒度并行遗传模拟退火算法构造多播树，基于波长图思想在多播树上进行波长分配。仿真研究表明，该算法是可行的，并且具有较好的性能。     

一种面向重尾分布的SSD磁盘调度算法

存储设备上的大量文件其长度呈重尾态分布,IO请求的响应延迟和请求大小有着密切关系,并且固态硬盘的IO操作不对称。基于以上几点,在内核NOOP调度算法的基础上提出一种针对重尾数据分布下的IO调度算法。该算法通过减少大量小片请求的等待时间,提高固态硬盘的性能。经实验验证,相比内核的NOOP调度算法,平均响应时间减少17%。     

一种支持视频VCR操作的共享动态缓存算法

支持VCR操作是流媒体服务的重要功能,现有流媒体服务器中的常见缓存算法均未能充分考虑对VCR操作的支持。在性能良好的SRB算法的基础上,结合补丁技术提出了支持VCR操作的SRC-SV算法。仿真结果表明,SRC-SV算法能够较好支持用户的VCR操作,并能有效地降低服务器开销以及主干网的带宽消耗。     

优化的多媒体流多播选路的试探算法

在网络实时应用如远程教育、视频会议等系统中,对多播多媒体流的每一个流,不仅有严格的端到端时延限制,还有网络带宽的约束。网络在满足应用服务请求的同时,还应尽可能地降低资源的消耗。因此需要研究优化网络资源利用的多媒体流多播路由算法。论文采用优化的试探方法解决多播路由多媒体流分发问题。实验结果表明优化后的算法,与已有算法比较起来,使客户端接收视频流的满意度提高5%～8%,链路带宽节余8%～20%。     

Range Multicast for Video on Demand

We explore a communication paradigm for video on demand, called Range Multicast. This schemeis a shift from the conventional thinking about multicast where every receiver must obtain the same data packet at any time. A range multicast allows new members to join at their specified time and still receive the entire video stream without consuming additional server bandwidth. Clients enjoy better service latency since they can join an existing multicast instead of waiting for the next available server stream. We also present techniques to support video-cassette-recorder-like interactivity in this environment. Unlike existing methods which require clients to cache data in a private buffer, the Range Multicast solution utilizes the shared network storage to make more efficient and cost-effective use of the caching space. Furthermore, since a range multicast can accommodate clients with different play points in the video, a client has a better chance to join an on-going multicast for normal playback after finishing a VCR operation. This strategy avoids the need for a new server stream, and thus further alleviates the server load. Our simulation results confirm the aforementioned benefits.     

Internet TV中受控组播的设计与实现

在Internet TV中，为解决现有的组播协议不支持用户认证，缺乏对数据访问的控制，在流量扩散安全性，频道切换延时等方面上存在的问题，通过扩展现有的组播协议，提出了完整的网络接入认证流程，加入组播组流程，退出组播组流程，异常／强制退出组播组流程，断网流程，实现了受控组播协议，使得对数据的访问得到控制，整个服务完全受控于服务方，减少了频道切换的响应时间，并通过SR仿真试验，验证了受控组播协议在Internet TV中的可实用性。     

Multicast protocols for scalable on-demand download

Previous scalable protocols for downloading large, popular files from a single server include batching and cyclic multicast. With batching, clients wait to begin receiving a requested file until the beginning of its next multicast transmission, which collectively serves all of the waiting clients that have accumulated up to that point. With cyclic multicast, the file data is cyclically transmitted on a multicast channel. Clients can begin listening to the channel at an arbitrary point in time, and continue listening until all of the file data has been received.This paper first develops lower bounds on the average and maximum client delay for completely downloading a file, as functions of the average server bandwidth used to serve requests for that file, for systems with homogeneous clients. The results show that neither cyclic multicast nor batching consistently yields performance close to optimal. New hybrid download protocols are proposed that achieve within 15% of the optimal maximum delay and 20% of the optimal average delay in homogeneous systems.For heterogeneous systems in which clients have widely varying achievable reception rates, an additional design question concerns the use of high rate transmissions, which can decrease delay for clients that can receive at such rates, in addition to low rate transmissions that can be received by all clients. A new scalable download protocol for such systems is proposed, and its performance is compared to that of alternative protocols as well as to new lower bounds on maximum client delay. The new protocol achieves within 25% of the optimal maximum client delay in all scenarios considered.     

Minimizing bandwidth requirements for on-demand data delivery

Two recent techniques for multicast or broadcast delivery of streaming media can provide immediate service to each client request, yet achieve considerable client stream sharing which leads to significant server and network bandwidth savings. The paper considers: 1) how well these recently proposed techniques perform relative to each other and 2) whether there are new practical delivery techniques that can achieve better bandwidth savings than the previous techniques over a wide range of client request rates. The principal results are as follows: First, the recent partitioned dynamic skyscraper technique is adapted to provide immediate service to each client request more simply and directly than the original dynamic skyscraper method. Second, at moderate to high client request rates, the dynamic skyscraper method has required server bandwidth that is significantly lower than the recent optimized stream tapping/patching/controlled multicast technique. Third, the minimum required server bandwidth for any delivery technique that provides immediate real-time delivery to clients increases logarithmically (with constant factor equal to one) as a function of the client request arrival rate. Furthermore, it is (theoretically) possible to achieve very close to the minimum required server bandwidth if client receive bandwidth is equal to two times the data streaming rate and client storage capacity is sufficient for buffering data from shared streams. Finally, we propose a new practical delivery technique, called hierarchical multicast stream merging (HMSM), which has a required server bandwidth that is lower than the partitioned dynamic skyscraper and is reasonably close to the minimum achievable required server bandwidth over a wide range of client request rates     

一种TCP友好的Internet DTV组播发送速率控制机制

当前的Internet没有对流媒体应用提供足够的QoS的保证,同时Internet DTV的组播发送速率必须满足以下两点要求:1)能够自适应网络拥塞的变化;2)能够适应节目码率的要求,因此需要在服务器端进行组播发送速率控制.分析了Internet DTV组播的难点,提出了一种基于缓冲区管理的具有网络自适应特性的组播发送速率控制方法.通过合理地控制发送方的发送速率,既能自适应网络状况的变化,又能满足流媒体实时播放的需求.仿真实验结果表明,该算法在保证服务器端发送缓冲区不溢出前提下,与单纯采用TFMCC方法比较,减少了接收端的平均丢包率,提高了网络电视节目流的传输质量.     

Dimensioning the capacity of true video-on-demand servers

The need to reduce the huge bandwidth demand of video-on-demand (VoD) services has led to the conception of both multicast and broadcast based techniques for the deployment of such services on a large scale. Interactivity, a desirable feature for video services, includes the capacity to perform VCR operations. However, whenever a viewer requests the performance of a VCR operation, his/her video stream becomes unsynchronized with that of his/her multicast group, and a new channel must be allocated for the performance of this operation. The present paper introduces a novel approach for determining the number of video channels needed to support such interactivity. Moreover, it investigates the performance of interactive systems with a pool of channels reserved for the support of VCR operations. Systems with both batching and piggybacking are analyzed. Results indicate that for a medium to high number of users performing VCR operations the number of channels required to achieve target levels of quality-of-service is lower for systems with no pool of reserved channels than it is for systems with such a pool.     

P2MCMD: A scalable approach to VoD service over peer-to-peer networks

In this paper, we deal with the problem of providing Video-on-Demand service under Scheduled Video Delivery (SVD) paradigm, which allows users to request a video for instant viewing or later viewing. In this paradigm, a major concern for the server is how to schedule the delivery of requested videos to assure users a continuous playback while utilizing its limited channels more efficiently. We propose a novel approach, P2MCMD, to deal with this problem. P2MCMD treats the service of video delivery in finer granularity. What is scheduled is segment-based multicast groups instead of video-based multicast groups. With the aid of dynamic multicast, the server is enabled to allocate its delivery resources flexibly and efficiently. The uneven popularity of provided videos are also taken into account and exploited so that the produced delivery schedule has the potential to boost the degree of multicast, thereby making the server achieve higher channel utilization. Moreover, peer-to-peer cooperation is employed to alleviate the workload at the server. The server can be relieved of the retransmission of duplicate video segments by enabling a latecomer to appeal to early arriving peers for the missing initial video segments. The performance of P2MCMD is evaluated through simulation. The experimental results show that our proposed approach can scale to a larger number of requests and allow the provision of a wider variety of videos at the server as compared to the existing approaches for SVD paradigm.     

P2P transfer of partial stream in multimedia multicast

In this paper, we propose a new multicast scheme that is based on the client-initiated-with-prefetching (CIWP) and peer-to-peer (P2P) transfer of a partial multimedia stream. In the CIWP scheme, when a new client joins an ongoing multicast channel, the server has to create an extra unicast channel to retransmit the partial stream that has already been transmitted. However, the unicast channel consumes some of the I/O bandwidth of the server, as well as some of the network resources between the server and the client's Internet Service Provider (ISP). To solve this problem, we propose the use of the P2P transfer algorithm to deliver the partial stream from a client that has already joined the ongoing multicast session to the newcomer. This P2P transfer between clients is limited to clients belonging to the same ISP. To further improve the performance, a threshold is used to control the P2P transfer. We performed analytical studies to show that the proposed multicast scheme can reduce the consumption of the network resources of the server, by utilizing the client's disk space. We also performed various simulation studies to demonstrate the performance improvement in terms of the use of the server's bandwidth and the waiting time for the clients’ requests.     

A hybrid scheme for multicast authentication over lossy networks

For multicast communication, authentication is a challenging problem, since it requires that a large number of recipients must verify the data originator. Many of multicast applications are running over IP networks, in which several packet losses could occur. Therefore, multicast authentication protocols must resist packet loss. Other requirements of multicast authentication protocols are: to perform authentication in real-time and to have low communication and computation overheads. In the present paper, a hybrid scheme for authenticating real-time data applications, in which low delay at the sender is acceptable, is proposed. In order to provide authentication, the proposed scheme uses both public key signature and hash functions. It is based on the idea of dividing the stream into blocks of m packets. Then a chain of hashes is used to link each packet to the one preceding it. In order to resist packet loss, the hash of each packet is appended to another place in the stream. Finally, the first packet is signed. The proposed scheme resists packet loss and is joinable at any point. The proposed scheme is compared to other multicast authentication protocols. The comparison shows that the proposed scheme has the following advantages: first, it has low computation and communication overheads. Second, it has reasonable buffer requirements. Third, the proposed scheme has a low delay at the sender side and no delay at the receiver side, assuming no loss occurs. Finally, its latency equals to zero, assuming no loss occurs.