本文是计算机专业的留学生essay范例，题目是“Concepts and Research into Green Networking（绿色网络的概念与研究）”，由于其可能带来的经济效益，节约过多的能源消耗成为网络建设中的一个关键问题。这些问题通常被认为是“绿色网络”，涉及到在策略、设备和网络协议中插入能源意识。
Saving of excessive energy consumption is becoming a key concern in networking, because of theprobable economical benefits. These concerns, usually argued to as “green networking”, relate to inserting energy-awareness in the strategy, in the devicesand in the protocols of networks.
In this work, I first formulate a more precise definition of the “green” attribute. I further more classify a few standards that are key enablers of energy-aware networking research. I then overview the up-to-date state of the art and offer a catalog of the relevant work, with a superior focus on networking.
Recent studies shows that Information Technology and communication advances are responsible for significant amount of world electric power consumptions which ranges from 2% to 10%, that is one of the contributing factor for global warming, via gases release from greenhouse and from the growth of demand of internet applications and services. Therefore, for these reasons energy efficient and sustainable networking often called “Green Networking”, has become a hot issue in the last few years.
Definition: Green Networking is the practice of selecting energy-efficient networking technologies and products, and minimizing resources use whenever possible. 
All facets of Information Technology and Communication are under supervision, from energy-saving design of all networking devices, to strategies which reflect the entire network’s energy depletion in the planning, design, implementation and management points, to new approach for long-run sustainability of the networking which covers reformed attitudes of users’ as well as smart energy mowing techniques.This special concern on Green Networking intentions at providing revolutionary influences to the research and development of energy-efficient networking solutions and approaches for network sustainability.
In this paper, the authors considered energy-aware network devices (e.g., routers, switches, etc.) able to trade their energy consumption for packet forwarding performance by means of both low power idle and adaptive rate schemes. The proposed analytical model is able to capture the impact of power management capabilities on network performance metrics. The analytical framework considers stochastic incoming traffic at the packet level with Long Range Dependency (LRD) properties.
On the basis of the analytical model, authors have chosen the parameters characterizing the joint usage of Adaptive Rate(AR) and Low Power Idle(LPI) energy-aware capabilities by optimizing the desired tradeoff between energy consumption and Quality of Service(QoS) while at the same time enforcing the satisfaction of given upper bounds on both. Since the performance and cost indicators used in the optimization depend on incoming traffic volumes and statistical features (notably, burst inter-arrival time and average burst length), researchers repeat the optimization periodically under updated estimations of these quantities.
The modeling and control framework has been validated experimentally by using a Linux-based open software router with AR and LPI primitives under traffic generated by real-world traces; the results demonstrate how the proposed model can effectively represent energy-aware and network-aware performance indexes. Therefore proposed model, is efficient and addressing green networking maintaining the Quality of Service (QoS) in the network.
Traditionally, networking systems are designed and dimensioned according to principles that are inherently in opposition with green networking objectives: namely, over-provisioning and redundancy. On the other hand, due to the lack of Quality of Service (QoS) provision from the Internet architecture, over-provisioning is a common practice: networks are dimensioned to sustain peak hour traffic, with extra capacity to allow for unexpected happenings. As a result, through low traffic periods, over-provisioned networks are also over-energy-consuming. Furthermore, on behalf of resiliency and fault-tolerance, networks are also deliberate in a redundant manner.
Devices are added to the structure with the sole purpose of taking over the duty when another device fails, which further adds to the overall energy ingesting. These objectives, drastically divergent to the environmental ones, make green networking an interesting, and technically challenging, research arena. A major change is indeed needed in networking research and development to introduce energy-awareness in the network design, deprived of compromising either the Quality of Service (QoS) or the network consistency.
This section illustrates a few key paradigms that the network infrastructure can exploit to reach the green objectives formalized above. We individuate three classes of solution, namely resource consolidation, virtualization and selective connectedness . These three categories represent three research directions, which may find further detailed applications in device and protocol design.
Regroups all the dimensioning strategies to reduce the global consumption due to devices underutilized at a given time. Given that the traffic level in a given network approximately follows a well-known daily and weekly behavior , there is an opportunity to “adapt” the level of active over-provisioning to the current network conditions.
In other words, the required level of performance will still be guaranteed, but using an amount of resources that is dimensioned for current network traffic demand rather than for the peak demand. This can, for example, be achieved by shutting down some lightly loaded routers and rerouting the traffic on a smaller number of active network equipment. Resource consolidation is already a popular approach in other fields, in particular data centers and CPU.
Regroups a set of mechanisms allowing more than one service to operate on the same section of hardware, hence refining the hardware operation. It results in a lowered energy consumption, as long as a single machine under high load consumes less than several lightly loaded ones, which is generally the case. Virtualization can be applied to multiple kinds of resources, comprising network links, storage hardware, software resources, etc. A typical example of virtualization consists in sharing servers in data centers, thus reducing hardware costs, improving energy management and reducing energy and cooling costs, ultimately reducing data center carbon footprint. In the current context, virtualization has already been deployed with success: e.g., the US Postal Service has virtualized 791 of its 895 physical servers . As virtualization is a more mature research field, we refer the interested reader to  for a detailed survey of virtualization techniques from a computer architecture perspective, and to  for a networking perspective. At the same time, it should be noted that a virtualization solution designed explicitly to reduce network energy consumption has yet to appear.
Applying the same base concept, selective connectedness of devices, as outlined in , , consists in distributed mechanisms allowing single pieces of equipment to go idle for some time, as clearly as probable for the rest of the networked devices. If the consolidation principle relates to resources that are shared within the network infrastructure, selective connectedness allows instead to turn off unused resources at the edge of the network. For instance, edge nodes can go idle in order to avoid supporting network connectivity tasks (e.g., periodically sending heartbeats, receiving unnecessary broadcast traffic, etc.). These tasks may have to be taken over by other nodes, such as proxies, momentarily faking identity of sluggish devices, so that no essentialmodification is required in network protocols
6.Green Networking With Packet Processing Engines: Modeling and Optimization带包处理引擎的绿色网络:建模与优化
With the goal of monitoring power consumption in metro/transport and main networks, the paper reflects energy-aware devices capable to shrink their energy chucks by adjusting their performance. In specific, the paper focuses on state-of-the-art packet processing engines, which normally characterize the most energy-consuming apparatuses of network devices, and that are often collected of a number of parallel pipelines to “divide and conquer” the received traffic load. The paper talk about goal to control both the power structure of pipelines and the way to issue traffic flows among them.
The authors proposed an analytical model to precisely represent the impact of green network technologies (i.e., low power idle and adaptive rate) on network-aware and energy-aware performance indexes. The model has been confirmed with experimental consequences, accomplished by using energy-aware software routers loaded by real-world traffic traces. The attained outcomes determine how the projected model can successfully epitomize energy-aware and network-aware presentation indexes. The method goals at dynamically adjusting the energy-aware device structure to lessen energy consumption whereas handling with received traffic signals and gathering network performance limitations. In order to genuinely comprehend the impact of such policy, a number of experiments have been executed by using experimental data from software router designs and real-world traffic traces.
Finally, as the ultimate goal of networking is to provide services to end-users, the quality of such services and of the user experience is a topic that spans over all the previous branches. Indeed, while energy efficiency is becoming a primary issue, it shall never be neglected that the energy gain must not come at the price of a network performance loss. This delicate tradeoff arises from opposite principles: indeed, while networked systems have traditionally be designed and dimensioned according to principles such as overprovisioning and redundancy, green networking approaches praise opposite practices such as resource-consolidation and selective-connectedness. The challenge lays in this case in applying the latter principles in a way that is as transparent as possible to the user – in other words, avoiding that resource consolidation translates into congestion, or that selective connectedness translates into unreachability. While the first wave of green studies focused more on the achievable energy gain, we believe that a systematic evaluation of networking performance from the user-perspective should be undertaken as well. Indeed, in all branches interesting questions remain, which deserve precise quantitative answers: Finally, we believe that while, for the time being, techniques of different branches have been studied in isolation, future research should address the combined impact of different techniques as well. Indeed, even though each of the above techniques alone do not constitute serious threats for the QoS perceived by the end-user, however it is not guaranteed that the joint use of several technique will not raise unexpected behaviour. Due to the current rise in green networking research and attention, it cannot be excluded that, in a near future, users will run Energy Aware Applications, in a home equipped with a green set-top-box implementing Interface Proxying functionalities, and will access the Internet through an Internet Service Provider implementing Energy Aware Routing in devices interconnected by Adaptive Link Rate lines – which opens a number of interesting questions that are so far all unexplored.
最后，由于联网的最终目标是向最终用户提供服务，因此此类服务和用户体验的质量是一个跨越前面所有分支的主题。事实上，虽然能源效率正在成为一个主要的问题，但绝不能忽视的是，能源收益不能以网络性能损失为代价。这种微妙的权衡产生于相反的原则:事实上，尽管传统上网络系统的设计和维度是根据过度供应和冗余等原则，但绿色网络方法赞扬相反的实践，如资源整合和选择性连接。在这种情况下，挑战在于以一种对用户尽可能透明的方式应用后一种原则——换句话说，避免资源整合转化为拥塞，或选择性连接转化为不可达。虽然第一波绿色研究更多地关注于可实现的能源收益，但我们认为，从用户角度对网络性能的系统评估也应该进行。的确，在所有的分支中都存在有趣的问题，这些问题值得作出精确的定量回答:最后，我们认为，虽然目前对不同分支的技术已经进行了孤立的研究，但未来的研究还应解决不同技术的联合影响。实际上，尽管上述每一种技术单独都不会对终端用户感知到的QoS构成严重威胁，但不能保证几种技术的联合使用不会引起意外行为。由于目前绿色网络研究和关注的增加，不能排除，在不久的将来，用户将运行能源感知应用程序，在一个配备绿色机顶盒实现接口代理功能的家庭，并将通过互联网服务提供商在自适应链路速率线连接的设备中实现能量感知路由(Energy Aware Routing)接入互联网——这打开了许多有趣的问题，这些问题迄今为止都没有被探索过。