KR20180016660A - Data processing apparatus in storage and controlling method thereof - Google Patents

Abstract

The present invention relates to a data processing apparatus in storage, capable of enhancing the performance and durability of the storage. According to the present invention, the apparatus comprises: an operating system (OS) determining a logical block address of data to be written in generating a write request of an application program part to transmit the determined logical block address and the data to be written together with a write command, and using a transferred file identification (ID) number to search a logical block address of data to be permanently stored to generate and transmit a range flush command including the searched logical block address when a synchronization request is generated; and an apparatus control unit to store the data to be written transmitted with the write command in a buffer memory unit when the write command is transmitted from the OS, and to store data corresponding to the logical block address transmitted with the range flush command among data of the buffer memory unit in a position of the logical block address of a non-volatile storage unit when the range flush command is transmitted.

Description

[0001] DATA PROCESSING APPARATUS IN STORAGE AND CONTROLLING METHOD THEREOF [0002]

The present invention relates to a data processing apparatus in a storage and a control method thereof.

An operation control unit of a host that manages a file using operating systems in which a file system and a device driver are built uses a FLUSH command of a storage interface Even when the supply is interrupted, the changed part of the stored data can be safely stored.

The flush command is a command for transferring data from the host to the storage and controlling the data storage operation of the storage. The command is for storing all the data in the volatile buffer memory portion existing in the storage in the nonvolatile storage portion of the storage .

In this case, the host generally refers to a part composed of an operation control unit, a central processing unit (CPU), a main memory and various input / output (I / O) interfaces. The storage includes a hard disk, a solid state disk or solid state drive, a flash memory, or a non-volatile memory.

A computer system composed of a host and storage may be a mobile computer, such as a personal computer, a server or a smart phone.

When an operating system requests an instruction to permanently store data, such as an explicit synchronization command, after a certain period of time has elapsed since the data was changed by the user's operation, the operating system of the host transmits a flush command to the storage, Writes all data present in the buffer memory unit to the nonvolatile storage unit in accordance with the transmitted flush command so that the data is permanently stored.

However, in a current storage structure, when a write operation is performed by a flush command triggered by a specific user program, not only the data belonging to the program but also all data existing in the buffer memory are written to the nonvolatile storage unit.

As described above, the operation of immediately writing data to the nonvolatile storage unit that does not require immediate persistence reduces the buffering effect and generates unnecessary write traffic.

In addition, due to the increase in traffic, the data processing time of the storage is delayed, and unnecessary write operation results in degradation of storage life.

Korean Registered Patent Publication No. 10-1450157 (Registered Date: Oct. 06, 2014, entitled "Nonvolatile Journaling Integrated Buffer Cache and Cooperative Data Management Device and Method for Nonvolatile Storage)

SUMMARY OF THE INVENTION Accordingly, the present invention has been made to solve the above-mentioned problems occurring in the prior art, and it is an object of the present invention to provide a method and apparatus for synchronizing data between a host and a storage system.

A data processing apparatus in a storage according to an aspect of the present invention determines a logical block address of data to be written when a write request of an application program unit is generated and transmits a logical block address and data to be written together with a write command, An operating system for searching for a logical block address of data to be permanently stored by using the transferred file identification number, and generating and transmitting a range flush command having a retrieved logical block address, and when the write command is transmitted from the operating system, Wherein the data to be written together is stored in a buffer memory unit, and when the range flush command is transmitted, data corresponding to the logical block address transmitted together with the range flush command among the data in the buffer memory unit is written into the logical block Above address The device comprises a control unit for recording.

The logical block address of the data to be permanently stored may be a logical block address of all data belonging to a specific file, a logical block address of only changed data in a specific file, or a logical block address of changed data to be permanently stored regardless of a specific file .

The operating system may determine changed data using a dirty bit.

According to another aspect of the present invention, there is provided a method of controlling a data processing apparatus in a storage, the method comprising: transmitting data and file identification information to be written along with a write request to an operating system; Determining a logical block address of data to be written and then transmitting the data to be written together with the write command and the determined logical block address to the device control unit, the device control unit recording the transferred data to be written in the buffer memory unit, Searching for a logical block address of data to be permanently stored using the file identification information transmitted from the application program unit when a synchronization request is generated or a permanent storage request is generated according to a predetermined time, The logical block address for the data Generating a range flush command, and transmitting the generated range flush command to the device control unit; and the device control unit transmits data corresponding to a logical block address of data to be permanently stored among the data temporarily stored in the buffer memory unit, And storing it permanently at the block address.

The logical block address of the data to be permanently stored may be a logical block address of all data belonging to a specific file, a logical block address of only changed data in a specific file, or a logical block address of changed data to be permanently stored regardless of a specific file .

The operating system may determine changed data using a dirty bit.

According to this aspect, the storage operation of only data requiring immediate synchronization is performed in the nonvolatile storage unit, thereby increasing the efficiency of buffering and scheduling, thereby reducing the amount of data to be subjected to the write operation and reducing the response time, thereby improving the processing efficiency.

1 is a schematic block diagram of a data processing apparatus in a storage according to an embodiment of the present invention.
2 is a data flow diagram for a data processing apparatus in a storage according to an embodiment of the present invention.
3 is a conceptual diagram illustrating an operation of storing data into a storage by operation of a data processing apparatus in the storage according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, but it should be understood that there may be other elements in between do. On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

An apparatus and method for controlling data in a storage according to an embodiment of the present invention will now be described with reference to the accompanying drawings.

First, a data processing apparatus in a storage according to an embodiment of the present invention will be described in detail with reference to FIG.

The data processing apparatus in the storage according to an embodiment of the present invention shown in FIG. 1 includes a host 100 and a storage 200 connected to the host 100.

The host 100 is a system having an operation control unit 110, a main memory 120, and various input / output interfaces (not shown).

In the host 100, the operations of the main memory 120 and the input / output interfaces are controlled by the operation control unit 110.

The operation control unit 110 includes an application program unit 111 for performing a specific task and an operating system 112 for performing operations such as hardware corresponding to commands transmitted from the application program unit 111. [

The main memory 120 is a fast storage area for temporarily storing data accessed when a data processing apparatus is a computer system. The main memory 120 is used for caching or buffering the data of the storage 200 and is a high speed volatile memory capable of random access such as a random access memory (RAM) Or non-volatile memory.

The storage 200 is a storage medium for permanently storing data stored in the main memory 120 of the host 100 according to a command of the host 100 even when the power is turned off. The storage 200 is a nonvolatile memory such as a hard disk, an SSD, It is a volatile storage medium.

The storage 200 includes a device control unit 210 for controlling the operation of the storage 200 and a storage unit 220 for storing the data permanently.

The device control unit 210 reads data from the storage unit 220 according to a command transmitted from the host 100 and transmits the data to the host 100 or transmits data transmitted from the host 100 to the storage unit 220 Write to the location.

The storage unit 220 includes a buffer memory unit 221 connected to the device control unit 210 and a nonvolatile storage unit 222 connected to the device control unit 210.

The buffer memory unit 221 is a storage space for temporarily storing data to be transferred under the control of the operation control unit 110 of the host 100. The nonvolatile storage unit 222 is a physical storage space in which data is permanently stored to be.

The operation of the data processing apparatus having such a structure will be described with reference to FIG.

First, the application program unit 111 of the operation control unit 110 reads data (i.e., data to be written) to be written with a write system call to the operating system 112 according to a user operation or a predetermined operation pattern The file information is transmitted (S11).

When the write request is transmitted from the application program unit 111, the operating system 112 reads the file information and the data transmitted together with the write request (S12).

In this case, the file information transmitted from the application program unit 111 includes an identification number (ID) of a file to which the transmitted data is to be written, that is, a file identification number, a location of data to be written stored in the main memory 120, An offset indicating a start position of an address to be written, and a length indicating a size of an address to which the transmitted data is to be written.

Accordingly, the operating system 112 determines the corresponding file by using the file identification number of the transferred file information, and determines the address of the storage 200 to which the transmitted data is to be written using meta data for the determined file And determines a logical block address (LBA) (S13).

Next, the operating system 112 transmits the data to be written (i.e., the transferred data) and the logical block address to the device control unit 210 of the storage 200 through a write command (S14).

At this time, a write command is generated according to a protocol defined by an interface defined between the storage 200 and the host 100, and the interface may be a SATA (AT attachment), a SAS (single attachment station) , a small computer system interface (NVMe), a non-volatile memory (NVM) express, and a peripheral component interconnect (PCI) express.

The device control unit 210 of the storage 200 writes the data to be written in the buffer memory unit 221 according to the write command (S15). At this time, the recording position of the buffer memory unit 221 is determined in accordance with the state of the buffer memory unit 221 determined by the apparatus control unit 210. [

In this way, when the write command is transmitted from the operating system 112, the device controller 210 first writes the data in the high-speed buffer memory 221 in order to improve the write performance of the storage 200.

At this time, the device controller 210 manages the address information (Logical Block Address) of the data recorded in the buffer memory unit 221. The data stored in the buffer memory unit 221 under the control of the device control unit 210 is written to the nonvolatile storage unit 222 according to a policy for controlling the performance of the device control unit 210 to maximize the input / And written to the nonvolatile storage 222 can be made at any time.

A write operation is performed on the data in the buffer memory unit 221 of the storage 200 by the write command between the host 100 and the storage 200 and is stored in the storage 200. [

If a predetermined time has elapsed in the state where file data is stored in the buffer memory 221 of the storage 200 or if an explicit synchronization request is generated from the application program unit 111 at step S16, 112 transmits the Range FLUSH command including the information of the logical block address of the data to be permanently stored to the device control unit 210 of the storage 200.

A synchronization request by the application program unit 111 is generated through a system call such as fsync or fdatasync and means that the data belonging to a specific file is immediately and permanently stored. And transmits the file identification number of the target file to the operating system 112.

 The operating system 112 reads the metadata of the file corresponding to the file identification number transmitted from the application program unit 111 when a synchronization request from the application program unit 111 occurs or a persistent storage request is generated according to a predetermined time A range flush command is generated together with the searched logical block address and transmitted to the device controller 210 of the storage 200 (S18) , S19).

The device control unit 210 reads the data stored temporarily in the buffer memory unit 221 from the operating system 112 of the host 100 when the logical block address of the data to be synchronized with the range flush command is transmitted from the operating system 112 of the host 100 The data corresponding to the logical block address transmitted is determined (S20), and the corresponding data is written to the nonvolatile storage unit 222 at the corresponding position and stored in the nonvolatile storage unit 222 (S21).

The logical block address of the data to be stored in the nonvolatile storage unit 222 that the operating system 112 sends together with the range flush command includes a method of sending all logical block addresses of all data belonging to a specific file, A method of transmitting only a logical block address of only changed data that has been determined by discriminating only changed data in a specific file, and a method of designating and sending logical block addresses of changed data to be permanently stored regardless of a specific file.

A method for the operating system 112 to determine changed data among data belonging to a specific file may use information about data of the storage 200 held in the operating system 112 and may use the data of the storage 200 A dirty bit or the like is a typical example. However, the present invention is not limited to this, and the operating system 112 may utilize other data structures that can identify changed data. In this case, the operating system 112 determines the logical block address of the changed data to be permanently stored in the nonvolatile storage unit 222 using the dirty bit, and transmits the logical block address for the determined changed data to the device control unit 210 .

The device control unit 210 of the storage 200 may store all the data temporarily stored in the buffer memory unit 221 in the nonvolatile storage unit 222 when a range flush command is transmitted instead of the conventional flush command, Only the data corresponding to the block address is stored in the non-volatile storage unit 222.

For example, as shown in FIG. 3, when the logical block address of '11' is transmitted from the operating system 112 of the host 100 together with the range flush command, the device control unit 210 reads' The nonvolatile storage unit 222 records the data C 'at the logical block address of' 11 'if there is data C' having the logical block address of '11'.

Accordingly, only the data required to be immediately synchronized is stored in the non-volatile storage unit 222 of the storage 200, thereby increasing the efficiency of buffering and scheduling, thereby reducing the amount of data to be subjected to the write operation and reducing the response time The processing efficiency is improved.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, It belongs to the scope of right.

100: host 110: operation control unit
120: main memory 111: application program unit
112: Operating system 200: Storage
210: device control unit 220:
221: buffer memory unit 222: nonvolatile storage unit

Claims (6)

The logical block address of the data to be written when the write request of the application program unit is generated is determined, and the logical block address and the data to be written together with the write command are transmitted together with the write command. When the synchronization request is generated, And generates and transmits a range flush command having the retrieved logical block address, and
Wherein when a write command is transmitted from the operating system, data to be written transmitted together with the write command is stored in a buffer memory unit, and when the range flush command is transmitted, the logical block address To a location of the logical block address of the non-volatile storage unit.
The method of claim 1,
The logical block address of the data to be permanently stored may be a logical block address of all data belonging to a specific file, a logical block address of only changed data in a specific file, or a logical block address of the changed data to be permanently stored irrespective of a specific file Data processing device. 3. The method of claim 2,
Wherein the operating system determines the changed data using a dirty bit.
The application program part transmits the data to be written together with the write request, the file identification information to the operating system,
Wherein the operating system determines a logical block address of data to be written using the file identification information when the write request is transmitted, and transmits data to be written together with a write command and a determined logical block address to a device controller,
Wherein the apparatus control unit records the transmitted data to be written in the buffer memory unit,
Wherein the operating system searches for a logical block address for data to be permanently stored using file identification information transmitted from an application program unit when a synchronization request is generated or a persistent storage request is generated according to a predetermined time,
The operating system generates a range flush command together with the logical block address of the retrieved permanently stored data and transmits the command to the device controller;
Wherein the apparatus control unit permanently stores data corresponding to a logical block address of data to be permanently stored in the buffer memory unit in the logical block address of the nonvolatile controller, / RTI >
5. The method of claim 4,
The logical block address of the data to be permanently stored may be a logical block address of all data belonging to a specific file, a logical block address of only changed data in a specific file, or a logical block address of the changed data to be permanently stored irrespective of a specific file A method of controlling a data processing apparatus.
The method of claim 5,
Wherein the operating system determines changed data by using a dirty bit.

Images