计算机科学
隐藏物
计算机网络
概率逻辑
虚假分享
智能缓存
GSM演进的增强数据速率
以信息为中心的网络
缓存算法
内容交付
互联网
带宽(计算)
核心网络
分布式计算
CPU缓存
电信
操作系统
人工智能
作者
Haibo Wu,Jun Li,Junge Zhi,Yanjie Ren,Lingling Li
标识
DOI:10.1145/3360468.3368187
摘要
Information-centric Networking (ICN) has become a promising paradigm towards the future Internet. In-network caching is one of its prominent features. Existing cache schemes roughly group into deterministic caching and probabilistic caching. However, most current studies tend to employ a sole caching strategy throughout the entire network, which is hard to meet both users' and ISPs' demands. We propose a hybrid caching strategy (HCache), which uses different caching methods in different regions and combine their merits. Specifically, the cache network is divided into the core and edge areas. The edge areas apply deterministic caching to reduce access delay from user's angle, while the core area utilizes probabilistic caching to enrich the cache diversity and reduce inter-domain traffic from ISP's angle. Simulations indicate HCache outperforms traditional probabilistic and deterministic caching schemes in terms of access delay, cache hit ratio and bandwidth savings. Moreover, HCache is efficient in content delivery and friendly to both users and ISPs.
科研通智能强力驱动
Strongly Powered by AbleSci AI