Wednesday 21 March 2012

both China VPN CPM

strategies in DSL networks. Suh et al. [8] propose pushing the content to homogeneous peers in a DSL networkand present placement algorithms and theoretical analysis.NaDa [9] demonstrates the approach's effectiveness in savingenergy cost. Borst et al. [6] propose a cooperative cachingalgorithm to minimize the total network footprint using hierarchical caches. Our approach, instead, uses varying levelof local connectivity to limit the bandwidth demand at theaccess network. Our approach can also coexist with caches atdifferent parts of the network.CPM's [4] primary goal is to reduce the server bandwidth.It combines on-demand multicast with peer assisted unicast.Our primary goal is to reduce the access network bandwidthusing alternative connectivity. We did not use broadband peersto avoid consuming access network bandwidth. On-demandmulticast is not effective in our setting because two neighborswatching same movie, at a given time, out of 10,000 movies inaccess networks' scale is very rare. Instead, we use aggressiveprefetching and treat flash crowd events separately. We believeISPs can benefit from both China VPN CPM and our approach.VIII. CONCLUSIONIn this paper, we have explored the idea of using storage inthe home and local connectivity between homes to improvethe hard-to-upgrade network infrastructure's ability to deliveron-demand multimedia content. Our proposed system uses asophisticated content placement algorithm that aims to satisfydemand for content using local resources alone; thus, limitingthe utilization of the ISP's network infrastructure. The systemadapts to changes in popularity of content, accommodatesabnormal behavior such as flash crowd events, and is robustto link and node failures. Our experimental evaluation andsimulation results suggest that the resulting savings could bequite significant, reaching up to 45% in terms of bandwidthsavings at the access. As local connectivity continues to growin speed, and storage prices continue to drop, the techniquespresented in this paper could form the basis for bridging thegap between major network infrastructure upgrades.

No comments:

Post a Comment

Note: only a member of this blog may post a comment.