Provisioning cost-effective mobile video caching

Provisioning cost-effective mobile video caching
复制标题

DOI:
10.1109/icc.2016.7511549
复制
发表时间:
2016-05
期刊:
2016 IEEE International Conference on Communications (ICC)
影响因子:
--
通讯作者:
Seyed Ehsan Ghoreishi;V. Friderikos;Dmytro Karamshuk;Nishanth R. Sastry;H. Aghvami
Seyed Ehsan Ghoreishi;V. Friderikos;Dmytro Karamshuk;Nishanth R. Sastry;H. Aghvami
中科院分区:
其他
文献类型:
--
作者:
Seyed Ehsan Ghoreishi;V. Friderikos;Dmytro Karamshuk;Nishanth R. Sastry;H. Aghvami

文献摘要

被引文献

相似文献

移动的视频流量的爆炸式增长要求在移动的运营商网络中部署内容缓存。通过网络内缓存,用户对流行内容的请求可以从部署在终端用户附近的移动的网关处的内容缓存来提供,因此大大减少了内容服务器和运营商网络骨干上的负载。在实践中,内容高速缓存可以安装在运营商网络内部的多个级别(例如,服务网关、分组数据网络网关、RAN等),从而产生了分级网络内视频高速缓存的思想。为了评估分层缓存的优点和缺点,在本文中,我们制定了一个缓存配置问题,其目的是找到最佳的权衡缓存存储和带宽节省的成本之间的分层缓存。更具体地说,我们的目标是找到最佳大小的视频缓存在不同层次的分层网络缓存架构,最大限度地减少传输带宽成本的存储成本的比率。我们克服了我们的问题,这是制定为一个二进制整数规划(BIP)的复杂性,通过使用规范对偶理论(CDT)。使用入侵杂草优化(伊沃)获得的数值结果表明,可以实现重要的收益,效益成本比和成本效率分别提高了43%和38%以上。
The exploding volumes of mobile video traffic call for deploying content caches inside mobile operators network. With in-network caching, users' requests for popular content can be served from a content cache deployed at mobile gateways in vicinity to the end user, therefore considerably reducing the load on the content servers and the backbone of operator's network. In practice, content caches can be installed at multiple levels inside an operator's network (e.g., serving gateway, packet data network gateway, RAN, etc.), leading to an idea of hierarchical in-network video caching. In order to evaluate the pros and cons of hierarchical caching, in this paper we formulate a cache provisioning problem which aims to find the best tradeoff between the cost of cache storage and bandwidth savings from hierarchical caching. More specifically, we aim to find the optimal size of video caches at different layers of a hierarchical in-network caching architecture which minimizes the ratio of transmission bandwidth cost to storage cost. We overcome the complexity of our problem which is formulated as a binary-integer programming (BIP) by using canonical duality theory (CDT). Numerical results obtained using the invasive weed optimization (IWO) show that important gains can be achieved, with benefit-cost ratio and cost efficiency improvements of more than 43% and 38%, respectively.