It giant competing for Grid Computing

It giant competing for Grid Computing The third wave of Internet information technology will appear in 20 trillion and will form an industry with an annual output value of US dollars.

"Using IT resources in the future is like using water and electricity, just as you can use it." This is William M, senior vice president of IBM Global server group and head of group administration. zeitler recently took part in the first sentence of IBM's "E-Commerce Server and total storage solutions Emy" forum in Shanghai. According to market analysts, the true purpose of William M. zeitler's trip is to describe grid computing as the third wave of the Internet. Before that, IBM announced a $4 billion investment plan to build 40 data centers around the world and build them into an experimental network to study and develop grid computing. Not only IBM, but other IT giants have also taken actions to seize the potential market of grid computing. Sun not long ago released its enterprise version of Grid Engine software and formed a consortium with Globus, an open-source software standard organization that enjoys a high reputation in the distributed computing field. Japan's NTT Data plan will join intel and SGI in a three-month grid computing experiment in mid-2002, with 1 million computers, including families, enterprises and academic institutions, its total processing capability will be five times faster than the existing fastest supercomputer. In addition, Compaq, which was later involved in the grid computing field, also said that it is currently developing a grid computing solution globally, to provide hardware and software products and service support to customers seeking for grid computing systems. According to Forbes ASAP, grid technology will reach its peak in 2005 and bring new opportunities to the Internet. If grid technology enables the market to grow at an expected annual growth rate of 17%, a large industry with an annual output value of 2020 US dollars will be formed in 20 trillion. What is grid computing? Grid computing is a new concept proposed by scientists to solve some of today's scientific difficulties, it is a technology that uses the Internet or a dedicated network to interconnect various computing resources that are geographically widely distributed, these computing resources include supercomputer, Computer Cluster, storage system, and visualization system. When necessary, this technology can also selectively treat these computing resources as separate and unified computing resources. As a "broker" for different computing resources, grid computing service providers can intelligently allocate computing resources, which brings two benefits: first, they can optimize existing computing resources, it solves engineering and design problems faster. Second, it can adjust each part of the application to the system that is most suitable for it, so as to solve the entire problem in a shorter time and at a lower cost. The idea of grid computing comes from another major "Power Grid )". The original intention of power grid is that power suppliers supply electricity according to user needs, consumers only need to pay for the part of the electricity they use. As a result, the basic idea of grid computing has been extended to obtain high-performance computing capabilities just as people get power from the grid in their daily lives. William M. zeitler believes that "the future will enter such an era ". As the grid computing environment can be connected to heterogeneous "isolated islands" of different standards in a wide area to form a huge global computing system, it is an advanced form of Internet development, it has been highly valued by countries and organizations around the world. At present, many forums have been set up through the Internet to build an experimental environment and carry out research projects, most of which can be shared online, this greatly facilitates research and utilization of related resources. Commercial enterprises such as IBM and Sun want to introduce such super computing capabilities into daily office areas, so that the company's IT department can send computing tasks that are not competent for local computers to the grid computer system, to save system deployment costs. According to the idea, grid computing will go through four stages to solve the problems of capacity, data, high availability and utilities respectively. It is understood that the capacity problem of grid computing has been solved, and data and high availability will be achieved by the end of this year and next year respectively. William M. Zeit ler believes that we also need to introduce corresponding service providers. Industrial Applications in China Huge demand for Grid Technology The development of grid computing will be a long process. Li guojie, academician of the Chinese Emy of engineering and director of the Institute of Computing Technology of the Chinese Emy of sciences, believes that "we must start from now" Because Grid technology is most likely to form a standard in recent years. Currently, China has carried out two projects: "national high-performance computing environment" and "advanced computing infrastructure Beijing Shanghai pilot project" to study grid computing, the two projects were led by the Chinese Emy of sciences and Tsinghua University respectively. The calculation Institute of the Chinese Emy of Sciences is carrying out the "Vega plan) the Research on Grid computing is based on metadata, component framework, intelligent body, grid Public Information Protocol, and grid computing protocol. Cao zhonglin, A pramem software company in Beijing, predicts that many industries in China, such as energy, transportation, meteorology, water conservancy, agriculture and forestry, education, and environmental protection, there will be huge demands for high-performance computing grids, that is, information grids. Currently, grid computing provides the following services in China: · Resource data Over the past few decades, the country has spent a lot of financial resources to collect data on underground, terrestrial, and atmospheric resources, the methods used include underground exploration, ground manual surveying and monitoring, aerial remote sensing, and satellite remote sensing. However, because the data is scattered in different places, sharing is difficult and utilization is low. Sharing them on a grid brings many benefits. · High-performance computing The establishment of a computing grid facilitates the sharing of resources among various computing centers, making full use of hardware and software resources, and saving costs. It can play an unprecedented role in basic research, automobile, large-scale hydropower engineering, petroleum exploration, meteorological climate, aviation, transportation, finance, medical and other fields. · Biological information The informatization and globalization of biological science has become the trend of the times. Our life scientific community also has a strong call for grid. Some biological research institutions led by Peking University life college have joined relevant international organizations and shared modern biological information resources on the grid to improve work efficiency. · Environmental Protection China has already established several environmental data collection networks and computer networks, but it is inconvenient to use because it has not achieved unified online sharing and management. After the grid is created, it can solve this problem and provide a variety of new application forms such as the cache service, directory service, data mining and classification service of news media. International commercial applications Criteria are key to success Just as TCP/IP is the core of the Internet, standard protocols and services need to be defined to build grid computing. Currently, standardization groups, including the Global Grid Forum, Object Management Organization OMG, W3C devoted to the study of network services and semantic WWW, and globus.org, have started planning. OMG, W3C, grid for um, and other standardization organizations held a "Software Service Grid Seminar" with people from academic and commercial fields on last July to accelerate the formulation of ggg standards for global big grids. Next, Globus, another open source code grid standard organization, also organized a conference to study the infrastructure of high-performance computing through WAN connections. Globus is currently developing standard grid architectures and other technologies. So far, there is no formal standard for grid computing, but William M. zeitler said that the relevant organizations and enterprises have reached a consensus on the core technology: globus Toolkit developed by Argonne National Laboratory in the United States and the Information Science Institute of Southern California (ISI) has become a de facto standard for grid computing, 12 computer and software vendors, including IBM, Entropia, Microsoft, Compaq, Cray, SGI, sun, Dian, Fujitsu, Hitachi, and NEC, have announced the use of globus toolkit. As an open architecture and open standard infrastructure, Globus Toolkit provides many basic services required to build grid applications, such as security, resource discovery, resource management, and data access. Currently, all major grid projects are built based on the protocols and services provided by Globus Tookit. In addition to standards, the lack of security, manageability, and IT talents is also an urgent problem for Grid Computing. Otherwise, it will not become the business architecture of enterprises. Some problems that are often ignored in the internal system environment, such as security, authentication, and reliability, must be solved in any distributed environment. It is understood that at present, each enterprise only provides IT solutions for grid construction within the University or the company for academic research or internal use, and is connected to the private grid, in order to form a shared grid with more powerful computing power, a great risk is required. In addition, are competing grid providers willing to sell their excess resources as needed? Whether the large amount of data and computing involved in grid applications can be securely shared among organizations is not what the current Internet and network infrastructure can do. Therefore, there are still many problems to be solved before grid computing can be used for commercial purposes.