KR100216027B1 - File system suitable for processing mass data - Google Patents

Abstract

A method for constructing a continuous medium file suitable for processing a large amount of data may be performed by formatting the disk area of the continuous medium in the order of a boot block, a super block, an inode block, an index block, and a data block, and the size of the index block and the data block in a large capacity. By appropriately extending the access to the data block to be performed in a single indirect designation through the index block, the system supports a maximum number of streams to satisfy the real-time transmission requirements of the compressed multimedia data stream to the subscriber and a large amount of data. Allows efficient storage and retrieval of.

Description

File system and file construction method for continuous media suitable for large data processing

Figure 1 is a file format layout of the file system according to the Jongle technique.

FIG. 2 shows a search path of blocks in inode list 6 for installing data block 8 in FIG.

3 is a diagram showing a data input / output structure in a file system according to the prior art.

4 is a file format layout diagram of a file system according to the present invention.

5 is a diagram illustrating a search path of an inode block 510 for installing a data block 530 through an index block in FIG.

6 is an overall structural diagram of a continuous medium file system according to the present invention.

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to file systems for continuous media, such as video-on-demand systems, and more particularly, to support the maximum number of streams that satisfy the real-time transmission requirements of compressed multimedia data streams to subscribers and to efficiently store and retrieve large amounts of data. The present invention relates to a file system for continuous media, and a method for constructing the file.

In general, file systems for continuous media such as video-on-demand systems use a disk as a recording medium and are known as a device for searching and reading a desired video information through a disk and transmitting it through a network. There is an urgent need to develop a file system for a large-capacity continuous media capable of meeting the demands and speeding up data search to perform real-time transmission to more subscribers.

A conventional file system for continuous media has to perform a search path as shown in FIG. 2 by having a file format as shown in FIG. 1, which is complicated by a search time and transmission process of data requested by the subscriber. Not only did it not meet service requirements, but there was a configuration problem of the system that could not process a large amount of data efficiently. Reference is made to the attached first, second and third views to explain the above reason occurring in a conventional system. Referring to FIG. 1, which shows a file format layout of a file system according to the prior art, boot block 2 is located at the beginning of a file and typically occupies the first sector of the disk. Bootstrap code for booting or initializing the operating system of the file system is generally located in Bootblock 2. The super block 4 located after the boot block 2 is a block indicating the state information which shows how big a file of the file system is, how many files can be stored, or where to find free space in the file. Inode list 6 shows a list of inodes located in the file format after the superblock 4 and consists of direct and indirect blocks. The data block 8 located after the inode list 6 is a part containing actual file data and management data. The conventional system having the file format as shown in FIG. 1 has a search path as shown in FIG. 2 when installed. 2 shows a search path of blocks in inode list 6 for installing data block 8 in FIG. In Figure 2, the direct 0-9 blocks in inode list 6 actually contain information that points to the disk block containing the data. In addition, a single indirect block has indirect information indicating a block having a list of numbers of the direct blocks. To find the data in the data block through this single indirect block, the contents of the indirect block 20 are read from the disk to find the disk block number of the direct block, and then the direct block is accessed to find the desired data. The double indirect block has a pointer to the list of indirect block numbers, and the triple indirect block has a pointer to the list of double indirect block numbers. As such, in order to find the data of the corresponding data block through the direct block, the path L1 needs to go through the path only once, but through the path L1 and L2 through the single indirect block, the path L1, L2, and L3 through the double indirect block. You must find the data in order. Therefore, in the conventional system, multiple accesses must be performed to find a data block in a disk that actually contains data. Therefore, there is a problem in that the performance of the file system is deteriorated due to the long search time of the data. This problem is more clearly identified by referring to FIG. 3 is a diagram illustrating a data input / output structure in a file system according to the related art, and includes a user mode unit 310 and a kennel mode unit 320. The user mode unit 310 is composed of a user process 30 and a system call unit 31 having an application unit A1 and a buffer B1, and the kennel mode unit 320 includes a protocol 33, a function call unit 34, a device driver 35 and buffers B2 and B3. have. The network interbase 330 is connected to the kennel mode unit 320 through lines L1, L2, and L3. Lines L1, L2, and L3 are interrupts, register accesses, and direct memory access lines, in turn. In the configuration of FIG. 3, it can be seen through the memory illuminating line that the user process 30 needs to perform two or more memory copy operations in order to transmit the video data requested by the subscriber to the network interface 330. This, in turn, is a factor in the efficiency of the network and, in turn, the performance of the system. In addition, the conventional file system of the existing operating system is used as it is for continuous media service systems such as video-on-demand systems, in which case the file system is a unit of a block having a relatively small size, for example 1-8 kilobytes of capacity The disk files were managed in blocks. This is because the conventional file system is designed for relatively small file management. Therefore, when such a system is applied to a large-capacity continuous media file service system having a file capacity of several tens of megabytes to several gigabytes as it is, the number of disk accesses is excessive and the file installation time is increased. This, in turn, is directly related to the problem of degrading system performance.

As described above, there is a problem in that the performance of the file system is deteriorated due to the disadvantage that the search time of the data takes a long time and the real time transmission is difficult due to the data copying operation.

Accordingly, an object of the present invention is to provide a file system for a continuous medium and a method for constructing the file, which can shorten the time required for file installation in order to solve the above problems.

Another object of the present invention is to provide a file system for a continuous medium suitable for processing large amounts of data and a method of constructing the file.

It is still another object of the present invention to provide a file system for a continuous medium capable of supporting a maximum number of streams while satisfying a real-time transmission requirement of a compressed multimedia data strip for a subscriber.

It is still another object of the present invention to provide a video system on demand and a file construction method capable of improving the performance of a file system by shortening data search time and performing real-time transmission during transmission.

According to the method for constructing a continuous medium file suitable for processing large data according to the present invention for achieving the above object, the disk area is formatted in the order of a boot block, a super block, an inode block, an index block, and a data block. And expanding the size of the data block to a large capacity so that the access of the data block is performed by a single indirect designation through the index block. In addition, according to the file system for continuous media suitable for large-capacity data processing, it is connected to a network connected to a plurality of subscribers to manage channel allocation and load, and the network manager unit for receiving and transmitting the desired continuous media data to the network; The application program is connected to the network manager via an established interface, and transmits a corresponding signal in response to the service request of the subscriber to the subscriber, and applies a control signal to the interface so that the subscriber can provide continuous data. And accessing the corresponding data block in the disk for continuous media in a single indirect designation method by the index block in accordance with the control signal applied through the interface, and providing the continuous media data desired by the subscriber to the network manager. year It has a configuration including at least a medium file system unit. According to the method and system described above, the time required for file installation is shortened, thereby enabling real-time transmission. In addition, to provide a file system for continuous media suitable for large data processing.

Hereinafter, a file system and a file construction method for a continuous medium according to a preferred embodiment of the present invention will be described with the accompanying drawings. Like elements in the accompanying drawings are labeled with the same or similar reference numerals or names for ease of understanding. In the following description, specific details are set forth by way of example and in detail in order to provide a more thorough understanding of the present invention. However, for those skilled in the art, the present invention may be practiced by the above description without these details. In addition, the structure and operation of the basic structure and components of conventional file systems for continuous media well known in the art are not described in detail in order not to obscure the subject matter of the present invention.

In the following description, a preferred embodiment of the present invention will be described by way of example only and with reference to the accompanying drawings. First, Figure 4 is a file format layout of the file system according to the present invention. FIG. 5 shows a path through which the inode block 510 searches for data through the index block to install the data block 530 in FIG. 6 is an overall structural diagram of a continuous medium file system according to the present invention. Before describing the overall technical gist, in the present invention, to continuously store the data of the continuous medium on the disk, and to reduce the number of times (access) to the disk to minimize the time required for searching, as shown in FIG. The size of the data block in the disk has been extended to 256 kilobytes, and access to the data block is made in a single indirect way through the index block of FIG. 4 in order to reduce the time taken for memory copying in the user mode. In addition, the index block can be prefetched into main memory such as SRAM when the file is opened, thereby improving the performance of direct designation.

Referring to FIG. 4, the disk area for the continuous medium is formatted in the following order: boot block 2, super block 4, inode block 510, index block 520, and data block 530, and the size of the index block 520 and data block 530 is determined. It expands to 256 kilobytes to accommodate large capacities and allows access to the data block 530 to be performed with a single indirect designation through the index block 520. Thus, the number of disk accesses is reduced, and the predictability of disk access time is increased by the expansion of 256 kilobytes. In addition, the index data of the index block is prefetched when the file is opened, thereby achieving the same effect as that of the direct specification.

Referring to FIG. 5, it can be seen that access to the data block 530 is performed in a single indirect manner through the index block 520. According to the inode and index structure shown in FIG. 5, when the size of the index block and the data block is used as 256 kilobytes, 64K block indexes are included for each index block, thereby covering a data block area of 16 gigabytes. Thus, one inode 510 has a file size of up to 208 gigabytes through 13 indexes. Typically, this is sufficient, given that the two-hour MPEG-2 movie file is about 5 gigabytes in size. Most files in a continuous media file system have only one index block, and the contents of the index block of each file are cached by the main memory when the file is opened, thereby improving performance. It will be easy to see the file structure for the media.

It solves the problems of the existing disk operating system, the excessive number of disk accesses and the time required to copy data from memory to memory, satisfy the real-time transmission of multimedia data, which is a requirement of on-demand video requester, which is a continuous medium, The structure of a file system for continuous media for simultaneous service is shown in FIG. In FIG. 6, an application unit 610 for receiving a request for a service from a multimedia service subscriber and notifying the subscriber of the service, and a stream API eye interface 620 for interfacing user mode and Kennel mode applications. In addition, the stream control provided to the kernel to control the channel mode in the user mode is stored in the interface unit 630 and the stream control is read and temporarily stored in the disk medium data according to the interface unit's request, and the data is transferred to the network manager unit 670. A continuous medium file system 640 comprising a file system suitable for handling media, stream buffers 650 and 651 located in the continuous medium file system 640 for temporarily storing stream data, and a continuous medium disk for storing stream data of the continuous medium. 660, assign the channel of the network and network A load distribution and management and consists of continuous media data to the network manager unit for performing a function of transmitting to the network 670 and to a network connection unit 680 of the various devices. In the system of FIG. 6, in order to maximize the performance of the system, the continuous medium data of the disk is processed in the kernel mode, read from the disk into the buffer, and transmitted directly to the network. That is, in the case of programming in the user mode, at least two repetitive copies must be made. In the present invention, this is solved by directly transmitting the line through the line L13. In order to support a movie to many subscribers at the same time, the file system needs to minimize search time. In the present invention, the disk data block is extended to 256 kilobytes in order to continuously store the continuous media data and to reduce the number of disk accesses. All data blocks are accessed by a single indirection through the index block.

According to the file configuration method for continuous media and the file system for continuous media as described above, the maximum number of streams can be supported while satisfying the real-time transmission requirements of multimedia data streams, and the performance of the file system can be improved by shortening the data searching time. It is effective to improve.

Although the above-described invention has been described and limited by way of example with reference to the drawings, the same will be apparent to those skilled in the art that various changes and modifications can be made without departing from the spirit of the invention. For example, the size of the index block and the data block can be further expanded or changed to suit a large capacity according to a case.

Claims (12)

In the method for constructing a continuous medium file suitable for processing a large amount of data, the disk area for the continuous medium is formatted in the order of a boot block, a super block, an inode block, an index block, and a data block, and the size of the index block and the data block is changed. And expanding to a large capacity so that access to the data block is performed by a single indirect designation through the index block. The method of claim 1, wherein the size unit of the extended data block is about 256 kilobytes. 3. The method of claim 2, wherein the index data in the index block is prefetched into an area of main memory at the time of file opening to ensure the performance of direct specification. The method of claim 3, wherein the size unit of the index block is the same as the size of the data block. A file system for continuous media suitable for processing large data, comprising: a network manager connected to a network connected to a plurality of subscribers to manage channel allocation and load, and receive and transmit the desired continuous media data to the network; An application program unit which is connected to the network manager via an established interface and transmits a corresponding signal in response to the service request of the subscriber to the subscriber and applies a control signal to the interface to provide the continuous media data desired by the subscriber; And a continuous medium for providing the continuous medium data desired by the subscriber to the network manager by accessing a corresponding data block in the continuous medium disk by a single indirect designation method according to the control signal applied through the interface. Continuous media file system that is characterized by having a file system section. 6. The file system of claim 5, wherein the area of the disk for the continuous medium is formatted in the order of a boot block, a super block, an inode block, an index block, and a data block. And the size unit of the extended data block is about 256 kilobytes. 8. The file system of claim 7, wherein the index data in the index block is prefetched into the static memory at the time of file opening to ensure the performance of direct specification. 9. The file system for continuous media according to claim 8, wherein the size unit of the index block is equal to the size of the data block to have 64k block indexes per index block. The file system for a continuous medium according to claim 5, wherein the continuous file system unit has a strip buffer for temporarily storing data therein. 6. The file system of claim 5, wherein the set interface includes a stream control interface that performs a function for controlling the kennel mode in the user mode. 6. The file system for a continuum of claim 5, wherein the application program unit further comprises a stream api for interfacing a user mode and a kennel mode application.