Anton Beloglazov, Jemal H. Abawajy, Rajkumar Buyya: Energy-aware resource allocation heuristics for efficient management of data centers for Cloud computing . Future Generation Comp. Syst. 28(5): 755-768 (2012) 1. 本文首先提出了一个节能云计算(energy-efficient Cloud computing)的架构框架(architectural framework)和原则。 下图 系统架构图 (S3.1) 如上图所示,分为四层,其中Green Service Allocator一层角色众多,承担了对服务用户请求进行分析、SLA协商、服务定价、VM调度、VM管理等功能。 2. (S3.2)中则提出了Power model,建立了能耗与CPU利用率的管理: 其中Pmax是服务器完全利用时最高的能耗,k是空闲服务器效率的能源比例(通常为70%),u则是CPU的利用率。 (能耗跟内存、磁盘、网络使用情况也有关系,但主要与CPU有关,所以上面的公式里只体现了CPU利用率,这个论断见S3.2) 3. 在此基础上,(S4)提出了资源供给与分配的算法,以改进云计算环境的节能。该算法是启发式的(heuristics),可以在确保满足客户QoS的前提下,使得数据中心的节能效果得到改进。 (S4.1) “VM placement”: 是关于针对创建新的VM请求,分配到哪台PM使得能耗增加最小的问题。此问题被建模为“bin packing problem with variable bin sizes and prices”,使用了修改版的“Best Fit Decreasing (BFD)”算法。 (S4.2)“VM selection”: 是关于如何优化当前的VM分配,以优化能耗的问题。主要分成两个步骤:首先选择迁移对象(一组VMs);然后使用MBFD算法,确定选出的迁移对象将被放置到哪些PMs上。 4. 针对“何时选择哪些迁移对象的问题”,(S4.2)中提出了三种选择策略(基本思想差不多): the minimization of migrations(MM) policy the highest potential growth(HPG) policy the random choice(RC) policy 以MM策略为例,下图是该策略的规则: 大意如下: 如果某一台PM的CPU利用率过高超出了上限,则找出一组个数最少的VM,这组VM就是迁移对象; 如果某一台PM的CPU利用率过低达不到下限,则该PM上所有VM都是迁移对象。 5. (S5)进行了实验验证 (S5.1)中介绍了性能度量指标: total energy consumption SLA violation percentage number of VM migrations initiated by the VM manager average SLA violation 实验是在模拟平台(CloudSim toolkit)上进行的。具体的实验过程和结果详见原文。 6. 作者在(S6)提出了相关的open challenges optimization of VM placement according to the utilization of multiple system resources optimization of virtual network topologies optimization of thermal states and cooling system operation efficient consolidation of VMs for managing heterogeneous workloads a holistic approach to energy-aware resource management via: http://www.cnblogs.com/ykt/archive/2013/02/14/2892562.html
Most every hot topic coming to my mind these days, I will check our social media system to see how social media reflects it. Word clouds are intriguing vehicles to present the common social image. Most word clouds generated by other systems are based on statistics of keywords mentioned together with the topic query, but our system can mine and present word clouds of various types of classified data, including positive/negative emotions towards a topic, pros and cons (or likes and dislikes) of the topic, etc. They reflect the social images of a topic from different perspectives. Today is Christmas Day, so it naturally becomes the topic of the day. The last post presents the social image of Xmas in Chinese social media , how about the word clouds in the English social media world? How about French? Here it is, as a way of celebrating the holiday season. Emotion-wise, as expected, Christmas time is full of joy and happy spirit, reflected by all the big green fonts of the positive sentiments expressed in social media. Negative sentiments do exist, but they are much fewer and scattered in smaller (red) fonts. The major complaints seem to be caused by the expectations not being met, disappointment in lack of fun and excitement. As for pros and cons, also as expected, pros are overwhelming. Both the emotion word cloud above and the pros-cons word cloud below seem to be bothered by some issues from some Christmas sweater party. In the French world known for its romance and passion, aimer+ and aimer- (love or not love) is a most striking sentiment word pair.. 【置顶:立委科学网博客NLP博文一览(定期更新版)】
CLOUD DELPHI METHOD, International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems(IJUFKS) Volume: 20, Issue: 1 (2012) pp. 77-97 DOI: 10.1142/S0218488512500055 2010年4月6日投稿 2010年10月4日一审结果,大修 2010年11月20日提交第一次修改版 2011年10月4日accepted with revision 2011年11月2日提交第二次修改版 2011年11月17日Accept 2012年2月7日在线出版 Abstract: Group decision making is an important category of problem solving techniques for complicated problems, among which the Delphi method has been widely applied. In this paper an improved Delphi method based on Cloud model is proposed in order to deal with the fuzziness and uncertainty in experts' subjective judgments. The proposed Cloud Delphi Method (CDM) describes experts' opinions by Cloud model and we aggregate the experts' Cloud opinions by synthetic algorithm and weighted average algorithm. Another key point of CDM is to stabilize and accommodate the individual fuzzy estimates by the defined stability rules rather than having to force them to converge, or reduce. The Cloud opinions and aggregation results can be exhibited in a graphically way leading experts to judge intuitively and it can decrease the number of repetitive surveys and/or interviews. Moreover, it is more scientific and easier to represent experts' opinion base on Cloud model which can combine fuzziness and uncertainty well. A numerical example is examined to demonstrate applicability and implementation process of CDM. Keywords: Decision making; Cloud Delphi method; Cloud model; fuzzy; uncertai nty 全文链接
From wikipedia: Criticism of the term This article's Criticism or Controversy section(s) may mean the article does not present a neutral point of view of the subject . It may be better to integrate the material in those sections into the article as a whole. (March 2010) During a video interview, Forrester Research VP Frank Gillett expresses criticism about the nature of and motivations behind the push for cloud computing. He describes what he calls cloud washing in the industry whereby companies relabel their products as cloud computing resulting in a lot of marketing innovation on top of real innovation. The result is a lot of overblown hype surrounding cloud computing. Gillett sees cloud computing as revolutionary in the long term but over-hyped and misunderstood in the short term, representing more of a gradual shift in our thinking about computer systems and not a sudden transformational change. Larry Ellison , CEO of Oracle Corporation has stated that cloud computing has been defined as everything that we already do and that it will have no effect except to change the wording on some of our ads. Oracle Corporation has since launched a cloud computing center and worldwide tour. Forrester Research Principal Analyst John Rymer dismisses Ellison's remarks by stating that his comments are complete nonsense and he knows it. Richard Stallman said that cloud computing was simply a trap aimed at forcing more people to buy into locked, proprietary systems that would cost them more and more over time. It's stupidity. It's worse than stupidity: it's a marketing hype campaign, he told The Guardian. Somebody is saying this is inevitable and whenever you hear somebody saying that, it's very likely to be a set of businesses campaigning to make it true.
Cloud Indexing by Seth Maislin (September 24, 2010 - 10:14 GMT) Last week, Seth Earley blogged about the inefficacy of social tagging , but there's one scenario in which social tagging will breathe new life into an esoteric, 200-year industry: book indexing. I've written hundreds of book indexes, presided over the American Society for Indexing, managed an international indexing partnership, taught courses, established standards, built tools, and consulted with a lot of influential folks, so trust me when I tell you that it pains me to see this happening. I believe with every fiber of my professional being that the human work of subject indexing is and will continue to be superior in quality to every alternative ever imagined. Oh well. There is just too much information to index by hand, period. Books, periodicals, websites, blogs, messages, and documents are being produced or transformed too quickly for humans to keep pace, regardless of training and tools. Perhaps in response, the use of search algorithms becomes ever more popular, while overly optimistic expectations of retrieval quality grows increasingly preposterous. A more realistic response would be an increase in subject indexers' fees -- after all, demand is outpacing supply at an astounding rate -- but indexers haven't experienced a rate increase since the 1990s. The truth is that editorial indexing and all smart hands-on tagging is disappearing in favor of automatic approximations. And it is a reasonable argument that the substandard tagging of millions of pages and documents is better than leaving most of them without any subject metadata whatsoever. As industries grow more and more interested in taxonomy and semantics and autoclassification, it is the book publishing industry that will experience the most pronounced change. Printed books will disappear, and electronic book readers will become networked. It is intriguing that books, with their inherent depth and complexity, can be completely deconstructed into chapters, pages, and sentences, and then subjected to the same social ingenuity applied to photographs and songs. Individuals will be able to mark up (tag) their copies of books, and those notes will feed into a cloud of content that can be shared among all readers. Following in the footsteps of social bookmarking comes social book- marking . Readers of popular books (and in particular works of fiction) will enjoy having access to deep and faceted lists of keywords, for use in both search- and browse-based user environments. These cloud indexes can be combined arbitrarily to create global indexes for entire book collections, personal or otherwise, and these global indexes can be refined using standard book metadata (e.g., publisher), personal metadata (e.g., date of first read), cloud metadata (e.g., Lady Gaga's favorites), and thesauri or taxonomies. Books that aren't popular are doomed in the cloud indexing paradigm. Publishers and authors must work hard to encourage readers to contribute, especially because traditional SEO techniques won't work well with book content. Book publishing will remain as competitive as always. And for the first time in the history of book publishing, it will be impossible to judge a book by its cover. Instead, we'll use the index.
Cloud Computing refers to shared computing resources accessed remotely as services (via the Internet, Intranet, dedicated network etc) , that can be provisioned on demand and billed according to usage . Most current Cloud Service providers (such as Amazon and Google) primarily gear their services at SMEs. However, many larger Enterprises are now beginning to think of: a) how they can implement Cloud Computing internally - Cloud Private b) how they can extend their Cloud Private to make use of available Cloud services from external vendors - Hybrid Cloud c) or how they can offer their own Cloud services - Cloud Public The term Cloud refers to on-demand services accessed over the Internet, however these services can be classified into 3 layers , each layer defining a Cloud of services on its own: 1. The Infrastructure Cloud provides on-demand access to server capacity, storage space, network connection and middleware software , with flexibility to increase or decrease capacity as required 2. IT Services or the Service Cloud provides a set of software components from which the Independent Software Vendors (ISVs) can create end-user applications. These components are generally exposed as web services. 3. Software-as-a-Service or the Application Cloud is a software application that is licensed for use as a service provided to customers on demand Usually, the term Cloud Computing refers to the Infrastructure Cloud. This set of Cloud services includes on demand access to server, storage, network services as well as the ability to design and deploy complex IT configuration which supports the business applications. Platform- as-a-Service is another service within Cloud Computing which offers developers a full suite of tools to design their applications without struggling with deployment details. When companies are considering whether or not to build a Cloud Private (also called Internal Cloud), it refers to a Cloud Computing platform. What has Virtualisation got to do with Cloud Computing? Cloud Computing relies on Virtualisation. However it requires an additional layer of automation and intelligence to deliver the on demand services. Virtualisation is the enabling technology whilst a Cloud Computing platform provides the level of automation and autonomics required to offer a zero touch self service Data Centre to its users. Public Cloud, Private Cloud and Cloud Hybrid Cloud Computing Principles: Shared computing resources, virtualised and accessed as a service Public Cloud Open market for on demand computing and IT resources Amazon, Google, IBM... Huge adoption by SMEs and currently tested by Enterprise However Maturing and Limited SLA Reliability and availability to be improved Security to be demonstrated Trust and confidence to be acquired Corporates and large Enterprises have existing Data Centres running their mission critical applications Large Enterprises are not likely to move their core or strategic applications onto Public Cloud in the short to medium time frame Private Cloud For Enterprises and Corporates with large scale IT For their core and strategic applications To take Cloud Computing benefits Data Centre consolidation and reduction of Capex Opex reduction by automation Flexibility, cycle time improvement and internal quality of service Cloud Hybrid Enterprises will continue to test and adopt Public Cloud Overflow management and peak adaptation via Public Cloud Hybrid Cloud = Private Cloud with connections to Public Cloud Cloud Private - internal and secure for your Enterprise. The Benefits of Cloud Private Cloud Private enables a more dynamic and cost effective use of IT infrastructure and resources. The top line benefits to your organisation include: Reduced Capital investment On demand scalability scale up when you need it, scale back in a downturn Get processes right first time around Reduce business process cycles Location independence for customers and Data Centre assets Lower price point for hardware/software usage Pay as You Go (PAYG) model Increased Enterprise agility Usage billing (PAYG model in IT Units) Automated provisioning Reference: cloud essential: http://www.cloudessential.com/cloudcomputing/ Cloud Computing Explained.pdf
维基百科(Wikipedia.org)对云计算的定义: 云计算是分布式计算技术的一种,其最基本的概念,是通过网络将庞大的计算处理程序自动分拆成无数个较小的子程序,再交由多部服务器所组成的庞大系统进行搜寻和计算分析,最后将处理结果回传给用户.通过这项技术,网络服务提供者可以在数秒内,处理数以千万计甚至亿计的信息,达到和超级计算机同样强大效能的网络服务. 加州大学伯克利分校(University of California at Berkeley)的Michael Armbrust等对云计算的定义: 云计算包括互联网上各种服务形式的应用以及这些服务所依托数据中心的软硬件设施,这些应用服务一直被称作软件即服务(SaaS),而数据中心的软硬件设施就是所谓的云, 云计算就是SaaS和效用计算.云分为公共云(Public Cloud)和私有云(Private Cloud). 云计算(Cloud computing)是一种基于互联网的超级计算模式,其原理非常类似于网格计算.它是把存储在大量分布式计算机产品中的大量数据和处理器资源整合在一起协同工作.作为一种新兴的共享基础架构的方法,它可以将巨大的系统池连接在一起以提供各种IT 服务.这使得企业能够将资源切换到需要的应用上,根据需求访问计算机和存储系统. 云安全联盟(Cloud Security Alliance, CSA)的研究报告提出的SPI模型(SPI Model)把云计算的服务形式分为基础架构即服务(Infrastructure as a Service,IaaS),平台即服务(Platform as a Service,PaaS)和软件即服务(Software as a Service,SaaS)三大类。开放云宣言组织(Open Cloud Manifesto)则根据研究的需要把云计算细分成终端用户―云、企业―云―终端用户、企业―云(综合)、企业―云―企业、企业―云(便携式)、私有(内部)云等6种模式。 作为网络时代的新计算形式,云计算是以数据、用户和服务为三大中心为导向的.在功能方面,真正的云计算平台应该能具备以下三方面的功能特性. ⑴提供资源――包括计算、存储及网络资源,需要服务提供者架设出规模巨大的全球化的数据库及存储中心,能够实现海量的存储、出色的安全性和高度的隐私性和可靠性.此外,它还应是高效的、低价的、节省能源的. ⑵提供动态的数据服务――包括原始数据、半结构化数据和经过处理的结构化数据,一个优秀的云计算架构一定要有提供大规模数据存储、分享、管理、挖掘、搜索、分析和服务的智能. ⑶提供云计算平台――包括软件开发API、环境和工具.云计算需要真正形成一个有生命力、有黏性、可持续发展的生态系统.
from http://www.bnl.gov/newsroom Extensive field study may offer clues to climate change, weather UPTON, NY, During May and June 2009, scientists from the U.S. Department of Energys (DOE) Brookhaven National Laboratory, Argonne National Laboratory, and the University of Colorado at Boulder will use high-tech scanners analogous to those in medical settings to make observations of clouds. The research, conducted at the Climate Research Facility (ACRF) in Ponca City, Oklahoma, run by DOEs Atmospheric Radiation Measurement (ARM) program, could lead to more accurate weather forecasts and predictions about climate. Clouds play a critical role in Earths weather and climate, said Brookhaven atmospheric scientist Dong Huang, lead researcher for this study. But poor understanding of clouds has long limited scientists ability to make accurate predictions about weather and climate change. One major challenge is the shear scale of the problem: Cloud processes occur on spatial scales ranging from sub-micrometers (millionths of a meter) to thousands of kilometers. The typical probes used by scientists sample a tiny volume. Using these methods, it would take hundreds of years to take readings from an entire cloud, while the typical lifetime of a cloud is just tens of minutes, said ARM chief scientist Warren Wiscombe, a co-investigator on this study. To scan a larger area in a short time, the team will use a novel cloud tomography approach to reconstruct three-dimensional cloud structure. Our approach is very similar to x-ray computed tomography (CT), used by doctors to diagnose disease, but this time the patients are clouds, said Huang. A CT scanner obtains x-ray data of the body slice by slice using detectors that rotate around the patient. Similarly a cloud tomography system uses multiple microwave sensors to scan clouds from several distinct ground locations. The interior structure of a cloud can then be inferred from the resultant radiometric measurements using sophisticated algorithms. The scientists will use a network of five microwave sensors to probe clouds thermal emission, or release of heat energy, along with two cloud radars, a variety of optical and infrared sensors, and weather balloons to measure other characteristics. Using this combination of instrumentation, we will be able to obtain three-dimensional maps of the distribution of clouds, atmospheric moisture, and other characteristics over a domain of about 10 kilometers, said Huang. The quality of the cloud and moisture reconstructions will be evaluated using concurrent cloud measurements collected by a research aircraft operated by another field experiment led by Brookhaven atmospheric scientist Andy Vogelmann. This study will collect long-term statistics on low-optical-depth clouds. The combined data will enable scientists to better understand the role of clouds in regulating Earths radiation energy budget or how the planet absorbs and re-radiates energy from the sun. These data will also be used to assess the validity of how various cloud processes are represented in computer models of cloud behavior. The combination of intensive field experiments, long-term observations, and modeling will provide important insights that could directly benefit weather forecasting and climate modeling, Huang said. This research is supported by DOEs Office of Science (Office of Biological and Environmental Research). Note to local editors: Dong Huang lives in Long Island, New York. Related Links: ACRF: http://www.arm.gov/acrf/ ARM: http://www.arm.gov/about/
The Open Cloud Consortium (OCC): supports the development of standards for cloud computing and frameworks for interoperating between clouds; supports the development of benchmarks for cloud computing; supports open source software for cloud computing; manages a testbed for cloud computing called the Open Cloud Testbed; sponsors workshops and other events related to cloud computing. 具体情况见网站: http://www.opencloudconsortium.org/index.html