2012年中国网络广告市场热点聚焦：破冰：网络广告产业链逐步打通，RTB产业链方兴未艾；转变：视频媒体、电商平台在网络广告市场中份额显著上升；迁徙：用户行为转移，多维互动需求将迎来动移动互联网爆发增长；整合：多屏切换，优势互补，跨媒体营销才刚刚开始；涅槃：广告营销与互联网的融合，面临的困惑与挑战。从产业链角度来看，Ad Exchange、RTB等新模式受到广泛关注，目前这些新型模式及其市场在国内都刚开始起步，广告主的认知程度还比较低。艾瑞分析认为：（1）越来越多的品牌广告主认识到新模式将带来非常好的效果；（2）更多媒体加入新模式的阵营，使得广告资源日益丰富；（3）新模式下受众对于广告的体验更好。因此，新模式网络营销市场发展前景广阔。

智慧城市是当今全球范围内出现的新理念和新实践。世界各国，尤其是欧、美、日、韩等发达国家和地区，都在积极开展各项研究与应用，研发城市智慧应用与系统，建立相应的城市试点进行实践。在我国，从中央到地方都在积极探讨发展智慧城市。

通过农村信息化通道和平台的融合发展，开发和推广实用的信息系统与应用，并进行相应的信息化应用教育，使农村信息化成为传统农村生产和生活方式变革的动力、工具与支撑，从而不断提升农村信息化的使用效率和使用价值。

The electronic security industry hosts a gamut of products which are primarily used for monitoring or surveillance, intrusion detection, fire of Carbon Monoxide detection, safe access, contraband detection and others.It also entails a variety of products designed to meet the specific requirements of the homes, commercial spaces such as offices and complexes, public places such as malls and large industries and institutions.The industry draws its revenues primarily from the sales,installation and replacement markets.Additionally,some of the electronic security systems also involve monitoring expenses which are a major source of the increasing recurring monthly revenue(RMR) of the dealers and integrators.

在科学研究、计算机仿真、互联网应用、电子商务等诸多应用领域,数据量正在以极快的速度增长,为了分析和利用这些庞大的数据资源,必须依赖有效的数据分析技术.传统的关系数据管理技术(并行数据库)经过了将近40年的发展,在扩展性方面遇到了巨大的障碍,无法胜任大数据分析的任务；而以MapReduce为代表的非关系数据管理和分析技术异军突起,以其良好的扩展性、容错性和大规模并行处理的优势,从互联网信息搜索领域开始,进而在数据分析的诸多领域和关系数据管理技术展开了竞争.关系数据管理技术阵营在丧失搜索这个阵地之后,开始考虑自身的局限性,不断借鉴MapReduce的优秀思想改造自身,而以MapReduce为代表的非关系数据管理技术阵营,从关系数据管理技术所积累的宝贵财富中挖掘可以借鉴的技术和方法,不断解决其性能问题.面向大数据的深度分析需求,新的架构模式正在涌现.关系数据管理技术和非关系数据管理技术在不断的竞争中互相取长补短,在新的大数据分析生态系统内找到自己的位置.

提出一种基于语义的恶意代码行为特征提取及检测方法,通过结合指令层的污点传播分析与行为层的语义分析,提取恶意代码的关键行为及行为间的依赖关系；然后,利用抗混淆引擎识别语义无关及语义等价行为,获取具有一定抗干扰能力的恶意代码行为特征.在此基础上,实现特征提取及检测原型系统.通过对多个恶意代码样本的分析和检测,完成了对该系统的实验验证.实验结果表明,基于上述方法提取的特征具有抗干扰能力强等特点,基于此特征的检测对恶意代码具有较好的识别能力.

网络异常检测对于保证网络的可靠运行具有重要意义,而现有的异常检测方法仅仅单独利用流量的时间相关性或空间相关性.针对这一不足,同时考虑流量矩阵的时空相关性,提出了一种基于MSPCA的全网络异常检测方法.该方法综合利用小波变换具有的多尺度建模能力和PCA具有的降维能力对正常流量进行建模,然后采用Shewart控制图和EWMA控制图分析残余流量.此外,还利用滑动窗口机制对MSPCA异常检测方法进行在线扩展,提出了一种在线的MSPCA异常检测方法.因特网实测数据分析和模拟实验分析表明:MSPCA算法的检测性能优于PCA算法和近期提出的KLE算法；在线MSPCA算法的检测性能非常接近MSPCA算法,且单步执行时间很短,完全满足实时检测的需要.

基于并行复算的故障恢复技术,将故障恢复的计算任务分配至未发生故障的结点上并行执行,从而显著缩短复算时间,有效降低故障恢复开销,提高并行程序容错性能.基于该故障恢复技术,提出了一种针对OpenMP并行程序的容错机制PR-OMP,有效解决了分段复算、复算负载重分布等问题；此外,还扩展了传统编译数据流分析技术,提出了针对OpenMP并行程序的数据流分析技术,并基于该技术计算状态保存开销进行优化.设计实现了用于支持PR-OMP的编译工具GiFT-OMP,并通过实验证明了PR-OMP机制及其支持工具的有效性,评估并分析了其性能和可扩展性.

针对网格计算环境动态,异构和分布的特性以及网格资源分配中资源利用率低、效益不均等问题,结合微观经济学理论,建立了一种多赢家式的网格资源拍卖模型(muti-winners auction model,简称MWAM).将隐马尔可夫模型应用在网格用户t时刻出价状态预测方面,并结合分配算法计算出能够获得所需资源的概率；并且在原有资源分配机制的基础上,结合非完全信息纳什均衡理论设计了一种多赢家拍卖算法.从理论上证明了资源分配结束后系统收益最大,且本模型符合微观经济学中的激励相容性与个人理性准则.实验模拟在验证了隐马尔可夫预测的可行性的同时,又与几种具有代表性的算法相比较,从资源利用率、系统总收益等方面突显了本模型的优势.

基于自负载周期流技术,提出一种采用降速率包列的可用带宽测量方法SLDRT(self-loading decreasing rate train),并全面分析了该算法在多跳网络、突发性背景流的环境下的性能.SLDRT采用单条包列即可实现对可用带宽的高速测量,具备单次采样、准确测量的特性,可通过调整递减因子等参数,提高测量精度,降低测量负载.理论分析和不同背景流场景下的实验结果表明:在多跳、突发性背景流下,SLDRT具有较强的健壮性；与pathChirp,Pathload算法相比,不仅测量精度优良,而且大量缩短了测量时间,减轻了因测量而引入的额外负载.

As the complexity of machines and architectures has increased, performance tuning has become more challenging, leading to the failure of general compilers to generate the best possible optimized code. Expert performance programmers can often hand-write code that outperforms compiler-optimized low-level code by an order of magnitude. At the same time, the complexity of programs has also increased, with modern programs built on a variety of abstraction layers to manage complexity, yet these layers hinder efforts at optimization. In fact, it is common to lose one or two additional orders of magnitude in performance when going from a low-level language such as Fortran or C to a high-level language like Python, Ruby, or Matlab. General purpose compilers are limited by the inability of program analysis to determine programmer intent, as well as the lack of detailed performance models that always determine the best executable code for a given computation and architecture. The latter problem can be mitigated through auto-tuning, which generates many code variants for a particular problem and empirically determines which performs best on a given architecture. This thesis addresses the problem of how to write programs at a high level while obtaining the performance of code written by performance experts at the low level. To do so, we build domain-specific embedded languages that generate low-level parallel code from a high-level language, and then use auto-tuning to determine the best performing low-level code. Such DSELs avoid analysis by restricting the domain while ensuring programmers specify high-level intent, and by performing empirical auto-tuning instead of modeling machine parameters. As a result, programmers write in high-level languages with portions of their code using DSELs, yet obtain performance equivalent to the best hand-optimized low-level code, across many architectures. We present a methodology for building such auto-tuned DSELs, as well as a software infrastructure and example DSELs using the infrastructure, including a DSEL for structured grid computations and two DSELs for graph algorithms. The structured grid DSEL obtains over 80% of peak performance for a variety of benchmark kernels across different architectures, while the graph algorithm DSELs mitigate all performance loss due to using a high-level language. Overall, the methodology, infrastructure, and example DSELs point to a promising new direction for obtaining high performance while programming in a high-level language.

This report is intended as a response to some of the questions posed by the FCC regarding the upcoming TV-band incentive auction, given in their NPRM, as they relate to the television whitespaces. In particular, we argue (1) that channel 37 should be made available for whitespace use; (2) that the channels reserved for wireless microphones should be reserved on an as-used basis only; and (3) that the guard bands which will be created via the incentive auction must be considered as database-registration-requiring whitespace if unlicensed devices are authorized to use them. These three proposals have two common themes: (1) they each work toward the goal of making otherwise-wasted spectrum available as whitespace; and (2) in each case, the key concept is that the involved parties can (and in some cases must) register their devices and use geolocation of some sort. We will sketch each of our proposals and show how together they can make whitespace available for up to 10 million more Americans with minimal overhead while ensuring that licensed users receive the quality of service that they expect. As a result, essentially no one would be left without whitespace access.

Software bugs are inevitable in software-defined networking (SDN) control planes, and troubleshooting is a tedious, time-consuming task. In this paper we discuss how one might improve SDN network troubleshooting by presenting a technique, retrospective causal inference, for automatically identifying a minimal sequence of inputs responsible for triggering a given bug in the control software. Retrospective causal inference works by iteratively pruning inputs from the history of the execution, and coping with divergent histories by reasoning about the functional equivalence of events. We apply retrospective causal inference to three open source SDN control platforms---Floodlight, POX, and NOX---and illustrate how our technique found minimal causal sequences for the bugs we encountered.

Breadth-first search (BFS) is a fundamental graph primitive frequently used as a building block for many complex graph algorithms. In the worst case, the complexity of BFS is linear in the number of edges and vertices, and the conventional top-down approach always takes as much time as the worst case.A recently discovered bottom-up approach manages to cut down the complexity all the way to the number of vertices in the best case, which is typically at least an order of magnitude less than the number of edges. The bottom-up approach is not always advantageous, so it is combined with the top-down approach to make the direction-optimizing algorithm which adaptively switches from top-down to bottom-up as the frontier expands.We present a scalable distributed-memory parallelization of this challenging algorithm and show up to an order of magnitude speedups compared to an earlier purely top-down code. Our approach also uses a 2D decomposition of the graph that has previously been shown to be superior to a 1D decomposition.Using the default parameters of the Graph500 benchmark, our new algorithm achieves a performance rate of over 240 billion edges per second on 115 thousand cores of a Cray XE6, which makes it over 7 faster than a conventional top-down algorithm using the same set of optimizations and data distribution.

What if robots and automation systems were not limited by onboard computation, memory, or programming? This is now practical with wireless networking and rapidly expanding Internet resources. In 2010, James Kuffner at Google introduced the term “Cloud Robotics" to describe a new approach to robotics that takes advantage of the Internet as a resource for massively parallel computation and real time sharing of vast data resources. The Google autonomous driving project exemplifes this approach: the system indexes maps and images that are collected and updated by satellite, Streetview, and crowdsourcing from the network to facilitate accurate localization. Another example is Kiva Systems new approach to warehouse automation and logistics using large numbers of mobile platforms to move pallets using a local network to coordinate planforms and update tracking data. These are just two new projects that build on resources from the Cloud. Steve Cousins of Willow Garage aptly summarized the idea: “No robot is an island." Cloud Robotics recognizes the wide availability of networking, incorporates elements of open-source, open-access, and crowdsourcing to greatly extend earlier concepts of “Online Robots" and “Networked Robots". Cloud Robotics has potential to improve robot performance in at least five ways: 1) Big Data: indexing a global library of images, maps, and object data, 2) Cloud Computing: parallel grid computing on demand for statistical analysis, learning, and motion planning, 3) Open-Source / Open-Access: humans sharing code, data, algorithms, and hardware designs, 4) Collective Robot Learning: robots sharing trajectories, control policies, and outcomes, and 5) Crowdsourcing and call centers: offline and on-demand human guidance for evaluation, learning, and error recovery. This article surveys related work as of Fall 2012.

This dissertation explores the multifaceted interplay between efficient computation and prob-ability distributions. We organize the aspects of this interplay according to whether the randomness occurs primarily at the level of the problem or the level of the algorithm, and orthogonally according to whether the output is random or the input is random.