[2015年第2期] 云计算发展态势与关键技术进展

development trend and key technical progress of cloud computing

        徐保民

         云计算，数据科学，大数据

        cloud computing, data science, big data

        近年来，随着信息技术的突破和数据科学的崛起，云计算引起了产业界、学术界和政 府部门的高度关注。 文章简要阐述了云计算的产生背景与面临的问题， 探讨了云计算的虚拟 化技术、编程模式、数据存储管理、节能技术、云计算安全、资源调度等若干关键技术的研究进 展，并对云计算的发展态势提出了自己的看法，如： 云计算的标准化、云计算的未来属于 PaaS、 混合云将成用户首选、图应用作为一种云服务等。

        As a new pattern of business computing, cloud computing is the development of grid computing, parallel computing, distributed computing, etc. Its ultimate goal is to regard computing, service, and application as a kind of public facilities which are available to the public. Cloud computing is internet based distributed computing model where for enabling convenient, on-demand network access to a shared pool of virtual resources and providing an elastic computing services. Users can enjoy the service even he knows nothing about the technology and the professional knowledge in this field. Cloud computing will bring a fundamental change in working methods and business models. So, cloud computing is often regarded as the third revolution of IT industry, following the personal computer revolution and the Internet revolution. With the breakthroughs of information technology and the rise of data science, cloud computing has attracted great attention from industries, academia, and governments. Government vigorously promotes the development of cloud computing and the innovation of cloud computing technology and applications through pilot demonstration and other activities. For example, in 2012, the Ministry of Science and Technology of China released the first ministerial-level cloud plan“12th five-year special plan of China cloud science and technology development”. It has an important significance for speeding up the cloud computing technology innovation and industrial development. IT giants have invested heavily in the layout of the cloud service and are committed to the research and application of advanced cloud computing technologies. Such as, in 2011, Baidu officially opened its cloud computing platform. It will continue to open IaaS, PaaS, and SaaS cloud platforms. Services provided by Baidu include cloud storage and virtual machines, application execution engine, intelligent data analysis and event notification service, network disk, open API, etc. Researchers focus on key technology breakthroughs, which solve some key science and technology issues, in cloud computing. For example, the key mechanism of cloud computing is to dispatch the computing tasks to resource pooling. It enables a variety of applications to gain computing power, storage, and a variety of software services according to their needs. To solve this problem, Xu et al. proposed for the first time an algorithm of job scheduling based on Berger model. The algorithm establishes dual fairness constraint to judge the fairness of the resources allocation in cloud computing. The paper firstly introduces the background of cloud computing and problems faced by cloud service provider and cloud. There are a number of underlying technologies and services that make cloud computing possible. Some of the most important technologies are resource management, resource scheduling, virtualization technology, programming mode, energy-saving technologies, and data storage management technology. These technologies are also explained in the paper. Meanwhile, many latest research results on these technologies are introduced and discussed in detail. Such as, the energy consumption has become the most important problem in cloud data center based on MapReduce framework. In this paper, we introduce the latest research work from four aspects of component level energy-saving, server consolidation, data replica management, and node startup and shutdown. Our latest research results are also introduced, namely, a hypergraph-based power saving algorithm based on data replica strategy in Hadoop. The algorithm has certain advantages such as fault tolerance and availability of data blocks. The energy efficiency is better than CS-k algorithm in the same operating environment. The paper also draws out the possible future research directions on some key technologies, such as cloud security technology and cloud resource scheduling. Finally, this paper discusses the hottest development features of cloud computing and helps the readers to understand the new trend developments in cloud computing.