KR20070061087A - Obeject based file system and method for inputting and outputting - Google Patents

Abstract

An object-based file I/O system and a method thereof are provided to offer performance and extension of the object-based file I/O system by storing a file to a file system based on an object storage device and reading a stored object from the file system in the object unit. An object I/O manager(205) performs file data I/O by connecting to the object storage device(102), which stores the file data by assigning the file data in a logical object unit. A metadata manager(203) stores and searches metadata by connecting to a metadata server(101) storing the metadata of the file data. A file access processor(202) transceives the metadata of the file data corresponding to an I/O request of a client(200) with the metadata manager. If the client requests the file data input, an object cache manager(204) extracts and transmits the metadata of the file data to the file access processor, directly receives the file data from the client, and transmits the file data to the object I/O manager by loading the file data to a memory in a page unit.

Description

Object based file system and method for inputting and outputting}

1 is an embodiment of a system component constituting an object-based storage environment.

2 is a diagram illustrating the components of a system configuring an object-based file input / output system according to a preferred embodiment of the present invention.

FIG. 3 is a conceptual diagram illustrating storage information and relationships of an object cache and a file inode in an object-based file input / output system according to an exemplary embodiment of the present invention.

4 is a conceptual diagram illustrating storage information and relationships of an object cache and an object input / output request in an object-based file input / output system according to a preferred embodiment of the present invention.

5 is a flowchart illustrating an input request of object-based file data in an object-based file input / output system according to a preferred embodiment of the present invention.

6 is a flowchart illustrating an output request of object-based file data in an object-based file input / output system according to a preferred embodiment of the present invention.

The present invention relates to a file system based on an object storage device, and more particularly, to load file data stored in the unit of an object on an object storage device in a unit of memory according to an I / O request of a client. The present invention relates to a system and a method for inputting and outputting file data requested by the client in object units.

The rapid development of the Internet environment and the evolution to network computing environments require storage systems that can provide high performance processing of large amounts of data, expansion of storage devices, and secure sharing of platform independent data. Traditional network attached storage (NAS) and storage area networks (SANs), however, are showing limitations in meeting these requirements in terms of performance, scalability, and data security. Object-based storage technology is recognized as a technology that can provide high performance, high scalability, data security, and platform-independent data sharing to solve the limitations of NAS and SAN.

The file system based on such an object storage device has a client-server structure distributed over a network, and is composed of a client file system, a metadata server, and an object storage device. The client file system provides a file access interface to the client application, and the metadata server manages the metadata for the files stored in the object storage device and is responsible for authentication and access control for the client. The object storage device stores and manages data and attribute information belonging to the file object.

This object-based file system model separates the part that performs logical data management of files and directories from the existing block device-based file system and the physical part of storage management such as allocating and mapping data blocks to storage devices. The physical storage management part should be executed in the object storage device, and the existing block-based interface should be changed to the object-based interface.

Therefore, unlike the existing block device-based file system, the object-based file system must process the file data stored in the object storage device in units of objects, and since the objects can be distributed and stored in multiple units, the storage path of the file data There is also a problem that requires processing.

An object of the present invention is to provide performance and extension of an object-based file system by providing a method of storing files in an object unit and a method of reading stored objects in a file system based on an object storage device. .

SUMMARY OF THE INVENTION The present invention has been made in an effort to provide a computer-readable recording medium having recorded thereon a program for executing an object-based file input / output method on a computer.

In one embodiment of the object-based file input and output system according to the present invention for solving the technical problem, connected to an object storage device for allocating and storing file data in logical object units to perform the input / output of the file data An object input / output manager; A meta data manager connected to a meta data server storing meta data of the file data to store and retrieve the meta data; A file access processor for transmitting and receiving metadata of file data corresponding to an input / output request of a client to and from the metadata manager; And when the client requests input of predetermined file data, extracts meta data of the file data and sends the file data to the file access processing unit, receives the file data directly from the client, loads the file data in a page unit, and loads the object data into a memory unit. Includes an object cache manager to send to.

The object cache manager may receive the file data through the object input / output manager and load the file data in a page unit and transmit the file data to the client when the client requests output of the file data stored on the object storage device. do.

The metadata management unit may record the object cache header list of the object data page, the object cache header list of the object property page, the number of objects, the object storage device identifier array, the object identifier array, and the object size array as metadata of the file data. Characterized in that.

The object cache manager may write an object data page, an object storage device identifier, an object identifier, an object data page index, an object data page size, an object data page state, and metadata information to an memory as an object cache header to store the file data. It is characterized by loading in the memory.

The object cache manager may transmit the object input / output request list, the object input / output request type, the object cache header list, the object storage information, and the object input / output status related to the file data input / output request of the client to the object input / output manager. And performing input and output on the file data.

According to an embodiment of the present invention, there is provided an object-based file input / output method, comprising: an object input / output step of performing file input / output by storing file data in logical object units; A metadata management step of storing and retrieving metadata of the file data; A file access processing step of transmitting and receiving metadata of file data corresponding to an input / output request of a client to and from the metadata manager; And when the client requests input of predetermined file data, extracts and stores metadata of the file data, receives the file data directly from the client, loads the file data into a memory in page units, and stores the file data in a logical object unit. Object cache management step of assigning and storing.

The present invention also includes a computer-readable recording medium having recorded thereon a program for executing the object-based file input / output method according to the present invention.

Hereinafter, with reference to the accompanying drawings will be described in detail a preferred embodiment of the present invention.

1 is an embodiment of a system component constituting an object-based storage environment. Referring to FIG. 1, an object based file system includes a client file system 100, a metadata server 101, and an object storage device 102.

The client file system 100 receives and processes a file access request from the client application 200.

The metadata server 101 stores and manages metadata about all files in the file system.

The object storage device 102 stores and manages data and attribute information of a file. Accordingly, the object-based file system stores and manages metadata and data of the file separately in the metadata server 101 and the object storage device 102, respectively.

In the object-based file system, file data is allocated and stored in logical object units. In this case, a plurality of objects may be created according to the configuration environment of the object storage devices 102 constituting the object-based file system, and these objects are distributed and stored in the object storage device 102.

Thus, a file can consist of several objects. These objects are distinguished by unique object identifiers per object storage device 102. The identifier of the object storage device is used together with the object identifier to distinguish the objects that constitute the file.

The object is made by the metadata server 101 when the file is created and the identifier information for the created object is managed in the file inode 400 which maintains the metadata information of the file.

2 is a diagram illustrating the components of a system configuring an object-based file input / output system according to a preferred embodiment of the present invention. Referring to FIG. 2, the client file system 100 includes a file access interface 201, a file access processor 202, a file metadata manager 203, an object cache manager 204, and an object input / output manager 205. It is configured together.

The file access interface 201 provides a generic interface for accessing the object based file system in the client application 200.

The file access request passed to the file access processor 202 processes the file access request through the file metadata manager 203 or the object cache manager 204 in the file access processor 202.

The file metadata manager 203 is in charge of meta data processing and management of the file through the metadata server 101.

The object cache manager 204 supports caching of file data objects on a page basis.

The object input / output manager 205 may perform object input to or from the object storage device 102.

FIG. 3 is a conceptual diagram illustrating storage information and relationships of an object cache and a file inode in an object-based file input / output system according to an exemplary embodiment of the present invention. Referring to FIG. 3, an object data header, an object storage device identifier, an object identifier, an object data page index, an object data page size, an object data page state, file inode information, and the like are recorded in the object cache header. The object cache header list of the page, the object property page header list, the number of objects, an array of object storage identifiers, and an array of object sizes are recorded.

The object cache 300 is composed of object data pages that cache file data in units of pages and object property pages that store object properties of a file. Pages included in the object cache are managed by the object cache header 301 structure.

The object cache header 301 includes an object data page 302 or an object property page 302 pointing to an area of system memory allocated for caching actual file data or object attribute data, and an object cache (referred to by the object cache header 301). Object storage identifier 303 and object identifier 304, the index 305 in the object in which the object data or attribute is stored, the size of the cached object data or property page 306 Maintains and manages state information 307 of the current object data or property page, and a field 308 that points to the inode of the file containing the object data or property page. Such information is used for managing the object cache 300 and input and output of the object.

On the other hand, the file inode 310 maintains the following information regarding the objects contained in the file. First, several object data pages and object property pages belonging to the same file may be cached in the object cache (311-312). They are then linked together and managed in file inodes. The file inode includes the number 313 of objects constituting the file, an identifier of object storage devices in which the objects are stored, an object identifier unique to each object storage device, and a size of each object (314-316). Therefore, when an object cache is allocated for input / output of file object data or object property, the allocated object cache is connected to the file inode.

4 is a conceptual diagram illustrating storage information and relationships of an object cache and an object input / output request in an object-based file input / output system according to an exemplary embodiment of the present invention. Referring to FIG. 4, information such as an object input / output request list, an object input / output request type, an object cache header list, object storage information, and an object input / output state are recorded in the object input / output request signal transmitted to the object input / output manager 205. .

The input / output request of the file data by the client application 200 is made through the object input / output request 400 to the object cache 300 and the object storage device 102 to store the object. Multiple object input / output requests 400 for the object storage device 102 may be requested at the same time, and may process several object data or property pages through one object input / output request 400.

Accordingly, the object I / O request 400 includes the next object I / O request 401 to the same object storage device, the type 402 of the object I / O request, the list 403 of object data or property pages included in the object I / O request, and the object I / O. It maintains and manages information 404 on the object storage device from which the request is made, and fields indicating the state 405 of the object input / output. In this case, each object input / output request 400 is made only for one object of the object storage device 102.

5 is a flowchart illustrating an input request of object-based file data in an object-based file input / output system according to a preferred embodiment of the present invention. Referring to FIG. 5, a method in which a client stores the file data in an object storage device according to an input request of file data will be described using the system of FIG. 2.

The file input request from the client application 200 first checks whether data to be input to the object cache 300 is already cached (500).

If data is not cached, an object data page for caching data is allocated to the object cache (501). In operation 502, information on an object in which data is to be stored and information on an object storage device are set in an object cache header indicating an object data page.

At this time, the object and the object storage device set the object and the object storage device identifier maintained in the file inode 400 in the object cache header 301 according to the relative position where the data is stored in the file.

When the object to be stored is determined, it is checked whether the object has enough space to store data (503). Thus, if the storage space is insufficient, the object storage device is requested (504) to pre-allocate storage space for storing data in the object.

It then copies the data in the client buffer space to the allocated object data page (505) and then indicates that the state of the object data page has changed and is caching the latest data (506).

If it is requested to store data synchronously (507), first create an object input request (508), connect the object data page to be stored to the generated object input request (509), and then input the object to the object storage device. Forward the request (510).

Otherwise, the cached pages that are changed here are made to the object storage device 102 when the system periodically stores the changed pages to the object storage device 102, and makes an object input request and connects the object data pages to be stored to the object storage device.

6 is a flowchart illustrating an output request of object-based file data in an object-based file input / output system according to a preferred embodiment of the present invention. Referring to FIG. 6, a method of outputting the file data stored in the object storage device in response to a request for outputting file data will be described using the system of FIG. 2.

Before reading the file data from the object storage device 102, it is first checked whether data is already cached in the object cache (600).

If it is not cached, the object cache reads data and allocates an object data page to cache (602). Otherwise, if the data is already cached, it is checked whether the cached data is the latest data (601).

If the latest data is not cached, in order to read the latest data from the object storage device, an object storage device storing the data to be read and an identifier of the object are set in the object cache header (603).

In this case, as in the object input, the object cache and the object storage device identifier maintained in the file inode 310 are set in the object cache header 301 according to the relative position of the data to be read.

After the object cache headers 301 are generated and connected (604-605) to the object output request, the object output request is forwarded to the corresponding object storage device (606).

The invention can also be embodied as computer readable code on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD_ROM, magnetic tape, floppy disks, and optical data storage, and may also include those implemented in the form of carrier waves (e.g., transmission over the Internet). . The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

As described above, optimal embodiments have been disclosed in the drawings and the specification. Although specific terms have been used herein, they are used only for the purpose of describing the present invention and are not intended to limit the scope of the present invention as defined in the claims or the claims. Therefore, those skilled in the art will understand that various modifications and equivalent other embodiments are possible from this. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

According to the present invention, the file data stored in the object unit on the object storage device is loaded into the memory in the unit of page according to the input / output request of the client to provide the input / output of the file data requested by the client in the unit of object, Through the file access interface, authentication and file metadata can be obtained from the metadata server and file objects can be accessed directly from the object storage device to improve system performance and provide extensions without sacrificing system performance. It can reduce the workload and share data independent of the file system's physical structure among client applications. In addition, you can set a separate security policy for logical objects stored in the object storage device.

The object is a logical storage unit for storing file data and may include a data access method, data property information, and a data security method. It also provides the ability to securely share objects across heterogeneous platforms by using an object access interface that is independent of the metadata for the object.

Claims (10)

An object input / output management unit connected to an object storage device which allocates and stores file data in logical object units and performs input / output of the file data; A meta data manager connected to a meta data server storing meta data of the file data to store and retrieve the meta data; A file access processor for transmitting and receiving metadata of file data corresponding to an input / output request of a client to and from the metadata manager; And When the client requests input of predetermined file data, the metadata of the file data is extracted and transmitted to the file access processing unit, the file data is directly received from the client, loaded into the memory on a page basis, and the object input / output management unit An object-based file input and output system comprising a transmitting object cache management unit. The method of claim 1, wherein the object cache management unit And when the client requests output of the file data stored on the object storage device, receives the file data through the object input / output management unit, loads the file data in a page unit, and transmits the file data to the client. The method of claim 1, And a file access unit for providing an interface for inducing the client to access object-based file data. The method of claim 1, wherein the metadata management unit Object-based, characterized in that the object cache header list of the object data page, the object cache header list of the object property page, the number of objects, the object storage identifier array, the object identifier array, and the object size array are recorded as metadata of the file data. File I / O System. The method of claim 1, wherein the object cache management unit The object data page, the object storage device identifier, the object identifier, the object data page index, the object data page size, the object data page state, and metadata information are written to the memory as an object cache header, and the file data is loaded into the memory. Object-based file I / O system. The method of claim 1, wherein the object cache management unit The object input / output request list, the object input / output request type, the object cache header list, the object storage information, and the object input / output status of the file data input / output request of the client are transmitted to the object input / output management unit to input / output the file data. Object-based file input and output system, characterized in that performing. An object input / output step of allocating file data in logical object units to perform input / output of the stored file data; A metadata management step of storing and retrieving metadata of the file data; A file access processing step of transmitting and receiving metadata of file data corresponding to an input / output request of a client to and from the metadata manager; And When the client requests input of predetermined file data, the metadata of the file data is extracted and stored, and the file data is directly received from the client and loaded into the memory in units of pages to store the file data in logical object units. Object-based file input and output method comprising the step of assigning and storing the object cache. The method of claim 7, wherein the object cache management step If the client requests the output of the pre-stored file data, the object-based file input and output method characterized in that the file data is loaded into the memory in units of pages and transmitted to the client. The method of claim 7, wherein And a file access step in which the client provides an interface for inducing access to object-based file data. The method according to any one of claims 7 to 9, A computer-readable recording medium having recorded thereon a program for performing the object-based file input / output method.

Images