当前位置: X-MOL 学术Comput. Netw. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Optimized Content Centric Networking for Future Internet: Dynamic Popularity Window based Caching Scheme
Computer Networks ( IF 4.4 ) Pub Date : 2020-07-23 , DOI: 10.1016/j.comnet.2020.107434
Sumit Kumar , Rajeev Tiwari

Content-Centric Networking (CCN) has emerged as a promising future Internet architecture that concentrates on content-centric information retrieval instead of current host-based infrastructure. In-network content caching has been integrated with CCN that allows the transformation from IP-based architectures to a completely decentralized environment which is based on the Content-Centric approach for future Internet architecture. In traditional caching schemes, each intermediate router caches a copy of content before forwarding it towards the requester. However, the caching capacity of routers is much smaller as compared to the forwarded traffic through them. For comprehensive utilization of available cache space, a novel content caching scheme called DPWCS (Dynamic Popularity Window-based Caching Scheme) has been proposed that efficiently places contents in the network cache. By intelligently determining the size of popularity window using request rate and the total number of distinct contents in the network, DPWCS is able to improve network performance by caching popular contents in the routers. The proposed scheme also translates the threshold determination issue into an optimization problem and propose cache size and Zipf distribution based heuristics for its computation. Extensive simulation results demonstrate the necessity of DPWCS, as the default caching scheme is unable to achieve significant network performance with realistic request rates, cache sizes, and requests distribution patterns. It is shown that the proposed scheme increases the cache hit-ratio, hop reduction ratio and server-load reduction ratio up-to ≈ 13%, ≈ 11% and ≈ 13% from peer competing caching schemes respectively with widely accepted Least-Recently-Used (LRU) cache replacement policy and ≈ 23% reduction in average network hop count under various simulation scenarios. This makes it suitable for Industry 4.0 and futuristic Internet architectures with 5G.



中文翻译:

针对未来Internet的优化内容中心网络:基于动态流行度窗口的缓存方案

以内容为中心的网络(CCN)已经成为一种有前途的未来Internet体系结构,该体系结构专注于以内容为中心的信息检索,而不是当前基于主机的基础结构。网络内内容缓存已与CCN集成在一起,它允许从基于IP的体系结构转换为完全分散的环境,该环境基于Content-Centric方法用于将来的Internet体系结构。在传统的缓存方案中,每个中间路由器都会在将内容副本转发给请求者之前缓存内容副本。但是,与通过路由器转发的流量相比,路由器的缓存容量要小得多。为了全面利用可用的缓存空间,已经提出了一种新颖的内容缓存方案,称为DPWCS(基于动态流行度窗口的缓存方案),该方案可以有效地将内容放入网络缓存中。通过使用请求率和网络中不同内容的总数智能地确定流行窗口的大小,DPWCS可以通过在路由器中缓存流行内容来提高网络性能。所提出的方案还将阈值确定问题转化为最优化问题,并提出基于缓存大小和基于Zipf分布的启发式算法进行计算。大量的仿真结果证明了DPWCS的必要性,因为默认的缓存方案无法通过现实的请求速率,缓存大小和请求分发模式来实现显着的网络性能。结果表明,所提出的方案从同等竞争的高速缓存方案中分别将高速缓存命中率,跳数减少率和服务器负载减少率分别提高到≈13%,≈11%和≈13%,这被广泛接受的“最近最少”在各种模拟情况下,使用(LRU)缓存替换策略,平均网络跳数减少≈23%。这使其适用于工业4.0和具有5G的未来互联网架构。

更新日期:2020-07-23
down
wechat
bug