2017년 3월 26일 일요일

Grid computing

Grid computing

The grid computing joins calculation resources (information storage domains such as computing power or the hard disks such as CPU's) which are on the network of wide areas such as the Internet together and is structure providing service as one computing system which compounded. The offered service is classified roughly into calculation processing and preservation, the use of data mainly. Large-scale calculation processing as I run out and a large quantity of data are developed as means to do it for preservation, the profit with the super computer of one place of calculation center and one set.

Table of contents

Summary

Of "the resource" to be necessary, and to provide the information service that the grid computing is necessary for anywhere when is necessary aim at the information infrastructure in the next generation of the Internet that aim at "make it a utility", and is realized as infrastructure on the Internet. The name comes from an electricity grid (power grid) which is the infrastructure which such a system has already established as of the early 21st century [1]. The explanation such as "the thing which settled a calculation power of the not being used PC, and performed a high-performance calculation" was reported without arresting only a certain one side [2].

The processing capacity of the computer improved year by year, but means such as the parallel computing, distributed computing has been developed to handle a large-scale calculation as possible in a short time. Call the thing which improved the processing capacity of one computer using such a technique drastically with a super computer, and the thing which I integrate plural computers and, as a whole, gave processing capacity to is called a computer cluster. In addition, the thing which can be equipped with the CPU about 100 and memory of number 100GB with the personal computer-based calculator exists.

The grid computing allows you to use a super computer and the computer cluster for these large-scale calculations to lie scattered in each place integrally and aims at the increase of a more overall processing throughput. Because specifications are different, you had to prepare, but the grid computing omits the trouble, and, as for the calculation resources such as a super computer or the computer cluster put in the calculation center of each place, each can treat the calculation processing that I put together in an individual specification when I use those plural numbers conventionally in each integrally.

The early grid was "a computing grid" aiming at performing large-scale calculation processing in plural calculations center (calculation grid), but data grid" proposes it in the field of scientific research to need it to treat large-scale data and is developed, and there are many things comprising both functions of calculation processing and data preservation, the use now. In addition, there are many things which the tool kit developed for industry makes much of throughput improvement of processing more plural than the calculation disposal of simple substances, and are developed (this becomes parallel of the processing, and it is a cause that the case which application in itself does not support for decentralization is considered).

Constitution of the grid

The grid computing is realized by middleware to use calculation resources and it. It absorbs the difference of the specifications of each calculation resource by applications software and software located midway of the operating system with the middleware and provides information to use them again integrally.

As for becoming the element of resources offered to Grid, it is in a domestic computer and plug computer, the scale of the cluster of the super computer, and it is thought that I share powerful memory resources and information and data or information to occur dynamically.

It is the structure that the resources which are necessary by requiring resources, and giving Grid processing are provided.

  • Side to use
    • I can use it from a normal computer.
    • If it leads to the network of the grid and implements a protocol, I can use it.
  • Side to provide
    • The simple computer to use can provide empty resources on the grid.
      • A calculation power and information, data, dynamic information (images of the video camera)
    • Calculation power of the high rank
      • Plural calculation powers offered on the grid are in combination available. I request a necessary calculation power, and an empty calculation power is informed of it and I send data and processing there and process it and am performed automatically. I seem to merely handle it in one's desktop from a user.
      • The offer of the calculation resources such as computer clusters.
    • The offer of memory resources.
      • The preservation of the backup of data becoming the second memory and the huge calculation result.
      • From the second memory, the transmission and reception of data is carried out to the calculation resources on the direct grid.

Middleware

Because I do not often limit the computer becoming the element of the grid to specific specifications, and an element building up by a feel again is big, it is explored that I build up the free operating system of the open source to a base. It is planned that it is in the specifications that shared a source code depending on the case that shared the specifications that are not limited to the specific CPU architecture and operating system as for the grid computing by doing so it.

Globus tool kit

The Globus tool kit is becoming the virtual standard as middleware to constitute a grid [3]. The service that Globus provides includes the following things.

  • Resource management (a grid resource management protocol: GRAM, Grid Resource Management Service).
  • An information service (monitoring and detective service: MDS - Monitoring and Discovery Service).
  • Data movement and management (wide area access to the second memory: GASS - Global Access to Secondary Storage).
  • GridFTP

Other tool kit

A following middleware system tool kit exists as well as Globus tool kit. But it is the tool kit for computer clusters, and there is the thing which does not adapt to a concept of the grid computing closely even if labelled as a grid.

The tool kit which Sun Microsystems develops by leadership. An open source.
A tool kit for macOS which Apple develops.
Middleware for Linux developed by the new information processing development system which was already dissolved or FreeBSD and system tool group. PC cluster consortium maintains it now.
A middleware product for Windows that Dai Nippon Printing developed. I am suitable for the use with the local grid.

Allied item

References

  • Ian Foster "What is the Grid? A Three Point Checklist" [1] [2] (English sentence)
  • "- - to spin the future of grid - - information-intensive society" (ISBN 4-621-07430-X)

Footnote

  1. ^ "- - pp.3 - 4 to spin the future of grid - - information-intensive society"
  2. ^ "- - p.3 to spin the future of grid - - information-intensive society"
  3. ^ http://www.globus.org/Globus

Outside link

It is acquired by "https://ja.wikipedia.org/w/index.php?title= grid computing &oldid=62405682"

This article is taken from the Japanese Wikipedia Grid computing

This article is distributed by cc-by-sa or GFDL license in accordance with the provisions of Wikipedia.

Wikipedia and Tranpedia does not guarantee the accuracy of this document. See our disclaimer for more information.

In addition, Tranpedia is simply not responsible for any show is only by translating the writings of foreign licenses that are compatible with CC-BY-SA license information.

0 개의 댓글:

댓글 쓰기

피드 구독하기: 댓글 (Atom)