With the widespread use of streaming media application on the Internet, a significant change in Internet workload will be provoked. Caching is one of the applied techniques for enhancing the scalability of streaming s...With the widespread use of streaming media application on the Internet, a significant change in Internet workload will be provoked. Caching is one of the applied techniques for enhancing the scalability of streaming system and reducing the workload of server/network. Aiming at the characteristics of broadband network in community, we propose a popularity-based server-proxy caching strategy for streaming medias, and implement the prototype of streaming proxy caching based on this strategy, using RTSP as control protocol and RTP for content transport. This system can play a role in decreasing server load, reducing the traffic from streaming server to proxy, and improving the start-up latency of the client. Key words streaming server - proxy - cache - streaming media - real time streaming protocol CLC number TP 302 - TP 333 Foundation item: Supported by the National High Technology Development 863 Program of China (2001AA111011).Biography: Tan Jin (1962-), male, Ph. D candidate, research direction: network communications, multimedia technologies, and web caching.展开更多
It is expected that by 2003 continuous media will account for more than 50% of the data available on origin servers, this will provoke a significant change in Internet workload. Due to the high bandwidth requirements ...It is expected that by 2003 continuous media will account for more than 50% of the data available on origin servers, this will provoke a significant change in Internet workload. Due to the high bandwidth requirements and the long-lived nature of digital video, streaming server loads and network bandwidths are proven to be major limiting factors. Aiming at the characteristics of broadband network in residential areas, this paper proposes a popularity-based server-proxy caching strategy for streaming media. According to a streaming media popularity on streaming server and proxy, this strategy caches the content of the streaming media partially or completely. The paper also proposes two formulas that calculate the popularity coefficient of a streaming media on server and proxy, and caching replacement policy. As expected, this strategy decreases the server load, reduces the traffic from streaming server to proxy, and improves client start-up latency.展开更多
Streaming media data have strict delay and loss requirements, which cannot be adequately supported by the current Internet. Proxy caching technology is introduced to alleviate the effect of the bottleneck bandwidth on...Streaming media data have strict delay and loss requirements, which cannot be adequately supported by the current Internet. Proxy caching technology is introduced to alleviate the effect of the bottleneck bandwidth on the transporting quality. A new architecture of multimedia proxy caching is given to solve the problem of the traditional Web caching technology for the multimedia streaming. The mixed media management module and the (resource) management module are added to this architecture, the prefetching and replacement method is studied considering the characteristics of streaming media data, the concept of weighed popularity is given to enhance the (exactness) of the replacement algorithm.展开更多
提出了一种基于段流行度的移动流媒体代理服务器缓存算法P2CAS2M2(proxy caching algorithm based on segment popularity for mobile streaming media),根据移动流媒体对象段的流行度,实现了代理服务器缓存的接纳和替换,使移动流媒体...提出了一种基于段流行度的移动流媒体代理服务器缓存算法P2CAS2M2(proxy caching algorithm based on segment popularity for mobile streaming media),根据移动流媒体对象段的流行度,实现了代理服务器缓存的接纳和替换,使移动流媒体对象在代理服务器中缓存的数据量和其流行度成正比,并且根据客户平均访问时间动态决定该对象缓存窗口大小。仿真结果表明,对于代理服务器缓存大小的变化,P2CAS2M2比A2LS(adaptive and lazy segmentation algorithm)具有更好的适应性,在缓存空间相同的情况下,能够得到更大的被缓存流媒体对象的平均数,更小的被延迟的初始请求率,降低了启动延时,而字节命中率接近甚至超过A2LS。展开更多
P2P流媒体cache是一种有效减少带宽开销、提高对象利用率的技术,通常采用FIFO,LRU等算法置换内容.然而,流媒体不同于Web对象,P2P网络也有别于客户/服务器模式.在分布式应用中这些算法可能影响系统的性能,为此,分析了FIFO和LRU置换算法,...P2P流媒体cache是一种有效减少带宽开销、提高对象利用率的技术,通常采用FIFO,LRU等算法置换内容.然而,流媒体不同于Web对象,P2P网络也有别于客户/服务器模式.在分布式应用中这些算法可能影响系统的性能,为此,分析了FIFO和LRU置换算法,提出了基于供求关系的SD算法,以及基于分片副本数量的REP算法,并对其进行评估和比较.针对不同的节点到达间隔,将SD和REP同FIFO,LRU进行比较,发现在启动延迟、媒体副本数量和根节点依赖度方面SD和REP几乎均优于FIFO和LRU.同LSB(least sent bytes)算法相比,某些场景中SD的启动延迟减少了约40%,而REP在副本数量方面远远超过LSB的结果,说明在P2P网络流媒体服务中使用SD和REP缓存置换算法有助于提高系统性能.展开更多
提出一种适用于移动通信网的两级缓存流媒体系统结构2CMSA(two-level cache mobile streaming archi-tecture),它突破了移动流媒体系统中终端缓存空间小、无线接入网带宽窄的局限;针对2CMSA结构设计了基于两级缓存的移动流媒体调度算法2...提出一种适用于移动通信网的两级缓存流媒体系统结构2CMSA(two-level cache mobile streaming archi-tecture),它突破了移动流媒体系统中终端缓存空间小、无线接入网带宽窄的局限;针对2CMSA结构设计了基于两级缓存的移动流媒体调度算法2CMSS(two-level cache based mobile streaming scheduling algorithm),建立数学模型分析了其性能;仿真实验证明,与原有的移动流媒体系统相比,使用2CMSS调度算法能够有效地节省网络传输开销,降低用户启动时延。展开更多
该文构造了一种新的流媒体缓存效用函数,该函数综合考虑流媒体节目的流行度特性及传输网络的代价参数;设计了一种针对多视频服务器、基于网络代价的流媒体缓存分配与替换算法(Network Cost Based cache allocation and replacement algo...该文构造了一种新的流媒体缓存效用函数,该函数综合考虑流媒体节目的流行度特性及传输网络的代价参数;设计了一种针对多视频服务器、基于网络代价的流媒体缓存分配与替换算法(Network Cost Based cache allocation and replacement algorithm,NCB)。仿真实验结果显示,NCB算法有效提高了缓存命中率,降低了传送流媒体所消耗的总体网络代价;该算法在网络结构复杂、节目数量庞大的Internet流媒体应用环境中表现出较优越的性能。展开更多
介绍了一种结合了CDN和P2P互补优势的流媒体混合内容分发网络(HyCDN)。针对HyCDN不同区域提出了相应的缓存算法,域内用户端综合考虑了流媒体前缀字节的有用性、文件的传输代价及点播热度,在此基础上提出缓存替换算法(Comprehensive Valu...介绍了一种结合了CDN和P2P互补优势的流媒体混合内容分发网络(HyCDN)。针对HyCDN不同区域提出了相应的缓存算法,域内用户端综合考虑了流媒体前缀字节的有用性、文件的传输代价及点播热度,在此基础上提出缓存替换算法(Comprehensive Value Cache Replacement Algorithm for P2P,CVCR4P2P);对域间边缘服务器采用补丁预取与调度算法(Dynamic Scheduling Algorithm for Proxy Caching,DSA4ProxyC),通过基于用户访问情况自适应伸缩缓存的分配方案,使流媒体后缀部分在边缘服务器中缓存的数据段与其流行度成正比。理论分析及实验结果表明,混合流媒体缓存调度策略的实施能有效地降低骨干网络带宽资源消耗,对用户请求到达速率的变化具有良好的适应性。展开更多
文摘With the widespread use of streaming media application on the Internet, a significant change in Internet workload will be provoked. Caching is one of the applied techniques for enhancing the scalability of streaming system and reducing the workload of server/network. Aiming at the characteristics of broadband network in community, we propose a popularity-based server-proxy caching strategy for streaming medias, and implement the prototype of streaming proxy caching based on this strategy, using RTSP as control protocol and RTP for content transport. This system can play a role in decreasing server load, reducing the traffic from streaming server to proxy, and improving the start-up latency of the client. Key words streaming server - proxy - cache - streaming media - real time streaming protocol CLC number TP 302 - TP 333 Foundation item: Supported by the National High Technology Development 863 Program of China (2001AA111011).Biography: Tan Jin (1962-), male, Ph. D candidate, research direction: network communications, multimedia technologies, and web caching.
文摘It is expected that by 2003 continuous media will account for more than 50% of the data available on origin servers, this will provoke a significant change in Internet workload. Due to the high bandwidth requirements and the long-lived nature of digital video, streaming server loads and network bandwidths are proven to be major limiting factors. Aiming at the characteristics of broadband network in residential areas, this paper proposes a popularity-based server-proxy caching strategy for streaming media. According to a streaming media popularity on streaming server and proxy, this strategy caches the content of the streaming media partially or completely. The paper also proposes two formulas that calculate the popularity coefficient of a streaming media on server and proxy, and caching replacement policy. As expected, this strategy decreases the server load, reduces the traffic from streaming server to proxy, and improves client start-up latency.
文摘Streaming media data have strict delay and loss requirements, which cannot be adequately supported by the current Internet. Proxy caching technology is introduced to alleviate the effect of the bottleneck bandwidth on the transporting quality. A new architecture of multimedia proxy caching is given to solve the problem of the traditional Web caching technology for the multimedia streaming. The mixed media management module and the (resource) management module are added to this architecture, the prefetching and replacement method is studied considering the characteristics of streaming media data, the concept of weighed popularity is given to enhance the (exactness) of the replacement algorithm.
文摘提出了一种基于段流行度的移动流媒体代理服务器缓存算法P2CAS2M2(proxy caching algorithm based on segment popularity for mobile streaming media),根据移动流媒体对象段的流行度,实现了代理服务器缓存的接纳和替换,使移动流媒体对象在代理服务器中缓存的数据量和其流行度成正比,并且根据客户平均访问时间动态决定该对象缓存窗口大小。仿真结果表明,对于代理服务器缓存大小的变化,P2CAS2M2比A2LS(adaptive and lazy segmentation algorithm)具有更好的适应性,在缓存空间相同的情况下,能够得到更大的被缓存流媒体对象的平均数,更小的被延迟的初始请求率,降低了启动延时,而字节命中率接近甚至超过A2LS。
文摘P2P流媒体cache是一种有效减少带宽开销、提高对象利用率的技术,通常采用FIFO,LRU等算法置换内容.然而,流媒体不同于Web对象,P2P网络也有别于客户/服务器模式.在分布式应用中这些算法可能影响系统的性能,为此,分析了FIFO和LRU置换算法,提出了基于供求关系的SD算法,以及基于分片副本数量的REP算法,并对其进行评估和比较.针对不同的节点到达间隔,将SD和REP同FIFO,LRU进行比较,发现在启动延迟、媒体副本数量和根节点依赖度方面SD和REP几乎均优于FIFO和LRU.同LSB(least sent bytes)算法相比,某些场景中SD的启动延迟减少了约40%,而REP在副本数量方面远远超过LSB的结果,说明在P2P网络流媒体服务中使用SD和REP缓存置换算法有助于提高系统性能.
文摘提出一种适用于移动通信网的两级缓存流媒体系统结构2CMSA(two-level cache mobile streaming archi-tecture),它突破了移动流媒体系统中终端缓存空间小、无线接入网带宽窄的局限;针对2CMSA结构设计了基于两级缓存的移动流媒体调度算法2CMSS(two-level cache based mobile streaming scheduling algorithm),建立数学模型分析了其性能;仿真实验证明,与原有的移动流媒体系统相比,使用2CMSS调度算法能够有效地节省网络传输开销,降低用户启动时延。
文摘该文构造了一种新的流媒体缓存效用函数,该函数综合考虑流媒体节目的流行度特性及传输网络的代价参数;设计了一种针对多视频服务器、基于网络代价的流媒体缓存分配与替换算法(Network Cost Based cache allocation and replacement algorithm,NCB)。仿真实验结果显示,NCB算法有效提高了缓存命中率,降低了传送流媒体所消耗的总体网络代价;该算法在网络结构复杂、节目数量庞大的Internet流媒体应用环境中表现出较优越的性能。
文摘介绍了一种结合了CDN和P2P互补优势的流媒体混合内容分发网络(HyCDN)。针对HyCDN不同区域提出了相应的缓存算法,域内用户端综合考虑了流媒体前缀字节的有用性、文件的传输代价及点播热度,在此基础上提出缓存替换算法(Comprehensive Value Cache Replacement Algorithm for P2P,CVCR4P2P);对域间边缘服务器采用补丁预取与调度算法(Dynamic Scheduling Algorithm for Proxy Caching,DSA4ProxyC),通过基于用户访问情况自适应伸缩缓存的分配方案,使流媒体后缀部分在边缘服务器中缓存的数据段与其流行度成正比。理论分析及实验结果表明,混合流媒体缓存调度策略的实施能有效地降低骨干网络带宽资源消耗,对用户请求到达速率的变化具有良好的适应性。
