中文摘要：

互联网的飞速发展,也使网络能耗急剧增长.但目前网络设备能效低下,未实现能耗比例计算的理念.而网络却为峰值负载而设计,在众多时间处于低负载,存在巨大的节能契机.首先介绍网络设备的能耗模型,继而从两方面阐述网络能耗优化的理论与技术：一方面,在假设网络总流量无法改变的前提下,为网络设备增加能源和性能状态,并优化本地控制策略,从而使单个网络设备实现能耗比例计算,或者在不提高现有网络设备能效的前提下,通过网络范围的协同和流量工程,使网络整体实现能耗比例计算的理念;另一方面,为网络提供缓存能力可以降低或缓解网络流量,从而减少网络的传输能耗或缓解其增长速度,智能的缓存部署、内容存储和请求路由能够进一步优化网络的能耗.在上述基础上,比较了各种技术的优劣,并分析了未来的研究方向.

英文摘要：

The fast development of the Internet also results in sharp rise in network energy consumption. The energy efficiency of present network devices is very low, which is an obstacle in realizing energy-proportional computing. However, while networks are designed for peak hour load, actual network load can be quite low in a large fraction of time, leaving great opportunity for energy saving. After an introduction of the energy consumption model of network devices, theories and technologies for optimizing network energy consumption are studied from two aspects. First, when the total amount of network traffic cannot be changed, it is possible to add power and performance states to individual network devices and optimize local control policies to realize energy proportionality; or, without sophisticated power management capabilities, it is also possible for a network as an ensemble to realize energy proportionality through coordination and traffic engineering. Second, it is possible to reduce or alleviate the traffic by providing cache capability to the network, which in turn reduces or alleviates the growth of network transmission energy. Intelligent cache deployment, content cache and request routing policies can further optimize network energy consumption. Based on the above discussion, a thorough comparison between different technologies is provided and potential future research directions are analyzed.