| At present,due to the rapid popularization of the Internet,a large number of end users use mobile devices such as mobile phones to watch videos.At the same time,due to the high scalability and low cost of peer-to-peer network technology(Peer-to-Peer,P2P),it has been widely used by researchers.attention.In order to provide better video services to end users,cloud service providers and video service providers combine to build a highly available peerto-peer network video-on-demand cloud platform.Generally,cloud service providers deploy a large number of edge cloud CDN nodes in different geographic areas and provide video services to end users by renting highly available ISP links.First,the video service provider uploads all video resources to the cloud data center,and then the cloud data center distributes the video clips to the edge cloud CDN node through the leased ISP line.However,because a large amount of video data is transmitted in the network link,a large amount of video transmission cost will be incurred.According to relevant data,video transmission costs account for 15% of cloud service providers' operating costs.Therefore,for cloud service providers,how to reduce the cost incurred when the cloud data center distributes video resources to other edge cloud CDN nodes is an urgent problem to be solved.On the other hand,with broadband access,more and more users choose to watch videos online in real time.When a user wants to watch a certain video,he will request the required video resource from a nearby edge cloud CDN node,and decode and play it locally in real time.However,the edge cloud CDN node may be missing the video resource,and will request the video resource from other cloud nodes,which not only incurs a large amount of video transmission costs,but also reduces the quality of video service.Therefore,how to reduce the video transmission cost of cloud service providers while ensuring service quality is also a problem that needs to be solved at present.This paper addresses the above two issues and combines the related technologies of the P2 P video-on-demand cloud platform,and proposes corresponding mathematical models and algorithms.The main research contents are as follows.First,based on the bandwidth peak charging model,this paper proposes a set of Netcutway algorithms based on the network maximum flow minimum cut.First,the video service provider uploads popular video resources to the cloud data center,and then the cloud data center distributes the video resources it owns to edge cloud CDN nodes in different geographic regions,and the edge cloud CDN node provides video to end users service.So that when a large amount of video data is transmitted on the link,a large amount of video transmission cost will be generated.In order to reduce the cost of video transmission,the cloud data center first divides the video into video clips of appropriate size according to the historical bandwidth peaks of different links,and pushes the divided video clips to other edge cloud CDN nodes on the network in a minimal cost manner.When the edge cloud CDN node receives the video clip,it first caches it to the local streaming media source server,then copies the video clip and pushes it to other edge cloud CDN nodes in a minimal cost way.At the same time,the edge cloud CDN nodes can request missing video clips from other cloud nodes at low cost.The algorithm is mainly divided into three parts.In the first stage,the cloud data center obtains the historical maximum bandwidth peak value,and uses the peak value to divide the video into appropriate clips at low cost.In the second stage,the maximum bandwidth that can be allocated to the link is calculated,and according to the maximum flow that can be allocated,the video clip is transmitted as far as possible without exceeding the peak of the maximum historical bandwidth of the link.In the third stage,when the edge cloud CDN node lacks a certain video clip,it needs to select the lowest cost path to request the video clip,and the cached video clip can be pushed to other nodes in the lowest cost way.Finally,edge cloud CDN nodes distributed in different geographical areas provide video services to end users.Second,this paper proposes the Netdmc algorithms.When a terminal user wants to watch a certain video,the video data will be requested from a nearby edge cloud CDN node,but the edge cloud CDN node may not have the video clip.When the edge cloud CDN node does not cache the video data requested by the user,the edge cloud CDN node will request the missing video resources from other edge cloud CDN nodes and cloud data centers.Therefore,when the edge cloud CDN node lacks the video data requested by the end user,the video transmission cost of the cloud service provider will be incurred.Secondly,due to a certain transmission delay caused by the video request,the quality of the video service is reduced.In order to minimize the video transmission cost of cloud service providers generated by edge cloud CDN nodes and ensure Qo S(Quality of Service),we have proposed a set of video transmission algorithms called Netdmc.The proposed algorithm can be divided into two parts.The first part is the least-cost path algorithm for non-emergency video clips.This algorithm reduces the cost when edge cloud CDN nodes request missing video clips.The second part is a low-latency algorithm for requesting urgent video clips to ensure the quality of end-user video service. |
PDF Full Text Request