KR20020096027A - Grid filesystem that uses the network and storage resource of personal computer that resident remotely in the world - Google Patents

Abstract

PURPOSE: A grid file system for storing the files and offering a service by using the storage of a personal computer and the network resources distributed worldwide is provided to transmit the large capacity of the data regardless of a bandwidth by dispersing the bandwidth requested in the file transfer depending on the number of the personal computers in a grid area. CONSTITUTION: A grid file system server fragments the file of the large capacity and replicates the fragment files. The replica files are dispersed in the personal computers participating in the grid file system service. If the specific data are requested, the file search and inquiry are requested for a central server having the file distribution information of the grid file system via a grid file system agent. The files dispersed in the various places are downloaded sequentially, and added to each other to obtain a complete file. Herein, the parallel processing is possible and the network environment of a high bandwidth is not needed. Thus, a high-quality file service is realized.

Description

Grid filesystem that uses the network and storage resource of personal computer that resident remotely in the world}

In the past, technologies such as clustering, distributed database systems, and distributed file systems have been used to service large amounts of text or multimedia data. However, this technique is an extension of server resources, and there is a limit to server resource expansion. In addition, in order to transmit data of a large amount of multimedia data to a client, there is a problem that a high bandwidth network is required for a cluster constituting a server. To solve this problem, grid computing technology is used to distribute files using the storage resources of computers connected to the Internet and participate in grid services, and to send files to clients requesting specific files. Instead of retrieving the file from the centralized server, the original file is obtained by directly merging the disassembled data from each personal computer participating in the distributed grid service. This allows the centralized server not to consider the extra network bandwidth needed for file transfer.

In the past, technologies such as clustering, distributed database systems, and distributed file systems have been used to service large amounts of text or multimedia data. However, this technique is an extension of server resources, and there is a limit to server resource expansion. In addition, in order to transmit data of a large amount of multimedia data to a client, there is a problem that a high bandwidth network is required for a cluster constituting a server. In order to solve this problem, this patent applies grid computing technology.

1 is an overall system diagram constituting the grid file system

Figure 2 shows the centralized data server and grid that make up the grid file system area.

Schematic diagram of a personal computer participating in the file system

1. Compress multiple binary files to reduce the file size and decompose them into specified sizes (1, 5, 10, etc.)

2. Each piece of disassembled file is encrypted.

3. Make multiple copies of each fragmented file fragment that has been encrypted.

4. Distribute copies of the created disassembled and encrypted file fragments in storage on the personal computer participating in the grid service area.

5. Information about distributed files is stored in a centralized data server.

3 is a schematic diagram showing the interaction of a grid file system and a grid file system agent.

1. The client requests a search query for a desired file through a grid file system agent or an index of stored files by category or organized folder method.

2. The grid file system agent retrieves the file information stored in the centralized server at the client's request and obtains decomposition information, distributed layout information, and status information of individual computers participating in the grid service that can be connected.

3. The grid file system agent uses the obtained information to access a plurality of personal computers participating in the grid service, download and merge the disassembled files sequentially, and reconstruct the original file.

4. During the file download, a candidate list of other computers with the same disassembled file item that can be currently connected is created in case the current connection is lost. Connect and continue to download files.

The grid file system server, shown at the top of FIG. 2, fragments a large number of binary files to create multiple replicas to personal computers participating in grid file system services as represented at the bottom of FIG. When requesting specific data by distributed arrangement, use the Grid Filesystem Agent application program as shown in FIG. 3 to request a file search and inquiry operation from the central server that has file layout information of the grid filesystem and use the results obtained. By disassembling in advance, files placed in various places are downloaded sequentially and merged into one to get the perfect file you want. In addition, since the file download can be parallelized and the server does not need the high-bandwidth network environment required for transferring a large amount of files from the server, the high-quality files available only at high-cost clustering file system services at low cost due to the spread of high-speed Internet. Service is available.

The grid file system agent program can be developed and serviced in various forms such as a standalone application program or a virtual drive device. In the latter form, users can use the files stored in the grid storage service as if they were connected directly to their computers.

We solved the problem of high bandwidth and the need for large server resources. In other words, centralized file server requires high bandwidth for data transmission, but using this grid file system, bandwidth required for file transfer is distributed as much as the number of personal computers participating in grid area. Large data transfer is possible. This method does not change data frequently but frequently reads or adds data, and has relatively low security requirements, such as VOD services of broadcasting stations or various video production companies, or cultural assets at home and abroad. It is suitable for various types of large-capacity multimedia data services such as multimedia data containing various information and multimedia data connected to an encyclopedia. Further consideration of security issues, such as increasing the strength of encryption of decomposed data in the future, will be widely applicable in more diverse fields. In particular, it would be more useful to use data storage for public interest by using the storage space of personal computers of public organizations including government and local government.

Claims (2)

Method of decomposing and distributing data by using storage resources of a personal computer distributed in a remote place (FIG. 2) Requesting data retrieval from the central server using the grid file system agent, receiving the results, and downloading data by merging and disassembling the data stored in a personal computer distributed remotely (Fig. 3).