KR101412885B1 - Method and storage system for controlling materialization in a storage medium using mode setting and delimiter - Google Patents

Abstract

Generating a plurality of write commands for writing data to a storage medium; Inserting at least one group start delimiter indicating the start of at least one group for the plurality of write commands between the plurality of write commands and at least one group end delimiter indicating the end of the at least one group ; Identifying at least one group from the plurality of write commands using the at least one group start delimiter and the at least one group end delimiter; Determining a persistence order or persistence atomicity for the plurality of write commands based on the identified at least one group; And processing the plurality of write commands based on the determination result.

Description

METHOD AND STORAGE SYSTEM FOR CONTROLLING MATERIALIZATION IN A STORAGE MEDIUM USING MODE SETTING AND DELIMITER BACKGROUND OF THE INVENTION 1. Field of the Invention [

The present invention relates to a method and a storage system for controlling persistence in a storage medium using mode setting and delimiter.

A higher abstraction layer, such as the file system of the host system or the database management system (DBMS), requires write requests to the storage device to be guaranteed and orderly. However, generic interface technologies provide only limited tools (eg, write-through methods, buffer flush commands, immediate persistence methods such as the FUA option) optimized for the structural characteristics of the hard disk.

Thus, storage systems perform synchronization between write requests using an immediate persistence method between two write requests that need to control the persistence order. However, the synchronization operation through immediate persistence causes a very large overhead, which has a decisive influence on the performance degradation of the write operation to the storage device.

Embodiments of the present invention provide for a configuration in which write data is persistent without synchronization through immediate persistence in the host system and host interface by directing persistent ordering control and persistent atomicity requests to the storage device Can be directly controlled.

In addition, embodiments of the present invention can enforce a persistence order between write data and control a series of data to be persistent with atomicity.

According to one embodiment, a method of controlling persistence in a storage medium includes generating a plurality of write commands for writing data to a storage medium; Inserting at least one group start delimiter indicating the start of at least one group for the plurality of write commands between the plurality of write commands and at least one group end delimiter indicating the end of the at least one group ; Identifying at least one group from the plurality of write commands using the at least one group start delimiter and the at least one group end delimiter; Determining a persistence order or persistence atomicity for the plurality of write commands based on the identified at least one group; And processing the plurality of write commands based on the determination result.

Wherein the step of inserting the at least one group start delimiter and the at least one group end delimiter comprises the steps of writing at least one of write commands included in the at least one group of the plurality of write commands, Inserting the group start delimiter into a first write command; And inserting the group end delimiter into a last write command included in the same group as the first write command.

Wherein the determining persistent atomicity for the plurality of write commands comprises determining at least one group start identifier among at least one group that begins with the at least one group start delimiter and ends with the at least one group end delimiter. Determining whether a group identifier for distinguishing the group from another group is included.

Wherein the processing of the plurality of write commands comprises starting the atomic persistence mode by the at least one group start delimiter among the plurality of write commands, The information contained in the write commands from the time when the atomic persistence mode is set until the atomic persistence mode is released, Wherein the atomic persistence mode is a mode in which persistence in the storage medium is made integral with respect to the write commands.

Inserting a mode start delimiter for setting a sequential persistence mode in a first write command among write commands in which persistence in the storage medium must be performed in order; And inserting a mode termination delimiter for releasing the sequential persistence mode into the last write command among the write commands that must be performed in order of the persistence, wherein the sequential persistence mode further comprises: The write data may be written in the write mode.

The step of determining the persistence order for the write data may include determining whether a mode start delimiter indicating the start of the sequential persistence mode and a mode end delimiter indicating the end of the sequential persistence mode are inserted .

Wherein the step of processing the plurality of write commands includes a step of writing the write commands after the sequential persistence mode is set by the mode start delimiter among the plurality of write commands and before the sequential persistence mode is released by the mode term identifier, And sequentially writing the write data corresponding to the write data to the storage medium in order of transmission through the host interface.

A method of controlling persistence in a storage medium, in accordance with an embodiment, includes generating a plurality of write commands for writing data to a storage medium; Inserting a mode start delimiter indicating the start of the sequential persistence mode and a mode end delimiter indicating an end of the sequential persistence mode; Determining a persistence order for the plurality of write commands using the mode start delimiter and the mode end delimiter; And processing the plurality of write commands based on the determination result, wherein the sequential persistence mode may be a mode for setting a persistence order between write data corresponding to a plurality of write commands in the storage medium .

Inserting the mode start delimiter and the mode termination delimiter comprises: inserting a mode start delimiter for setting a sequential persistence mode in a first write command among write commands to be persistent in the storage medium; And inserting a mode termination delimiter for releasing the sequential persistence mode into the last write command among the write commands which must be executed in order of persistence.

Wherein the determining of the persistence order for the plurality of write commands includes determining whether a mode start delimiter for setting the sequential persistence mode and a mode end delimiter for releasing the sequential persistence mode are inserted into the plurality of write commands And a step of judging whether or not the received signal is transmitted.

Wherein the step of processing the plurality of write commands includes a step of writing the write commands after the sequential persistence mode is set by the mode start delimiter among the plurality of write commands and before the sequential persistence mode is released by the mode term identifier, And sequentially writing the write data corresponding to the write data to the storage medium in order of transmission through the host interface.

A storage system that controls persistence in a storage medium according to one embodiment includes a plurality of write commands for writing data to a storage medium in a storage device from a host system and at least one group of write commands A host interface for receiving at least one group start delimiter indicating start and at least one group end delimiter indicating an end of said at least one group; Identifying at least one group from the plurality of write commands using the at least one group start delimiter and the at least one group end delimiter, and identifying at least one group from among the plurality of write commands based on the identified at least one group A controller for determining the persistence order or the persistence atomicity and processing the plurality of write commands based on the determination result; And a storage medium for perpetuating the plurality of write commands according to the processing result.

Wherein the group start delimiter is inserted into a first write command among write commands included in the at least one group among the plurality of write commands and to be permanently persistent in the storage medium, Th write command and the last write command included in the same group as the first write command.

Wherein the controller includes group delimiters for distinguishing the at least one group from the other groups among at least one group initiated by the at least one group start delimiter and terminated by the at least one group term delimiter Or not.

Wherein the controller writes the write command from the at least one group start delimiter to the at least one group termination delimiter after the at least one group delimiter is set by the at least one group delimiter, The information contained in the write commands from the time when the atomic persistence mode is set to before the atomic persistence mode is released is made to be in a non-persistent state And the atomic persistence mode may be a mode in which persistence in the storage medium is made integral with respect to the write commands.

A mode start delimiter for setting a sequential persistence mode is inserted into a first write command among write commands that are to be permanently persistent in the storage medium and a mode end delimiter for releasing the sequential persistence mode, , And the sequential persistent mode may be a mode for setting a persistence order between write data corresponding to a plurality of write commands in the storage medium.

The controller may determine whether a mode start delimiter indicating the start of the sequential persistence mode and a mode end delimiter indicating the end of the sequential persistence mode are inserted between the plurality of write commands.

Wherein the controller writes the write data corresponding to the write commands until the sequential persistence mode is released by the mode termination delimiter after the sequential persistence mode is set by the mode start delimiter among the plurality of write commands , And sequentially persistent to the storage medium in order of transmission through the host interface.

The storage device can notify the host system through the host interface that the storage device is a device capable of controlling the persistent configuration related to the persistence order or the persistent atomicity.

The host system may transmit information to the storage device through the host interface to enable the storage device capable of the persistent configuration control to use the persistent configuration control function.

A storage system that controls persistence in a storage medium according to one embodiment includes a plurality of write commands for writing data to a storage medium in a storage device from a host system and a plurality of write commands for indicating the start of a sequential persistence mode A host interface for receiving a mode start delimiter and a mode end delimiter indicating the end of the sequential persistence mode; A controller for determining a persistence order for the plurality of write commands using the mode start delimiter and the mode end delimiter and processing the plurality of write commands based on the determination result; And a storage medium for holding the plurality of write commands according to a result of the processing, wherein the sequential persistence mode may be a mode for setting a persistence order between write data corresponding to a plurality of write commands in the storage medium .

Wherein the mode start delimiter is inserted into a first write command among write commands to be persistent in the storage medium and the mode end delimiter is inserted into the last write command among the write commands that must be executed in order .

The controller may determine whether a mode start delimiter for setting the sequential persistence mode and a mode term delimiter for releasing the sequential persistence mode are inserted into the plurality of write commands.

Wherein the controller writes the write data corresponding to the write commands until the sequential persistence mode is released by the mode termination delimiter after the sequential persistence mode is set by the mode start delimiter among the plurality of write commands And sequentially persistent to the storage medium.

According to an embodiment, accurate and efficient persistence can be performed by directly transmitting a persistence order or atomicity control request to a storage device using a mode setting and a delimiter.

In addition, according to one embodiment, a persistence unit composed of commands or command groups is defined using a delimiter, and atomic persistence is enforced per persistence unit, so that persistence intention for a plurality of write requests can be more effectively transmitted to the storage apparatus.

In addition, according to one embodiment, the structure of the file system, the buffer cache, and the I / O scheduler can be further simplified and improved by eliminating the task for writing synchronization on the host system by directly transmitting the order control request of the persistence to the storage device. The cost of development, maintenance, and maintenance of the host system can be greatly reduced.

1 is a conceptual diagram illustrating a hierarchical structure of a computer system including a storage device.
2 is a diagram illustrating a process in which a write request of a file system is synchronously performed in a storage device.
3 is a diagram illustrating a process in which a write request of a file system is performed asynchronously in a storage device.
4 is a flowchart illustrating a method for controlling persistence in a storage medium according to one embodiment.
5 is a flowchart illustrating a method for controlling persistence in a storage medium according to another embodiment.
FIG. 6 illustrates a method of representing an atomic persistence request according to a method of controlling persistence in a storage medium according to an exemplary embodiment of the present invention. Referring to FIG.
7 is a flowchart illustrating a method for controlling persistence in a storage medium according to an embodiment.
FIG. 8 is a diagram illustrating a method for expressing a persistence order and a persistent atomicity request according to a method for controlling persistence in a storage medium according to an embodiment.
9 is a diagram illustrating a method for transferring a sequential persistence request for a part of write commands according to a method of controlling persistence in a storage medium according to an exemplary embodiment.
10 is a diagram illustrating a method for transferring an atomic persistence request according to a method of controlling persistence in a storage medium according to an exemplary embodiment of the present invention.
11 is a diagram illustrating a method of using a sequential persistence mode and an atomic persistence mode according to a method of controlling persistence in a storage medium according to an exemplary embodiment of the present invention.
12 is a diagram illustrating a result of applying a method of controlling persistence in a storage medium according to an embodiment to a data structure of a Serial ATA (SATA) Set Features command.
13 is a block diagram of a storage system that controls persistence in a storage medium according to one embodiment.
14 is a block diagram of a storage system that controls persistence in a storage medium according to another embodiment.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to or limited by the embodiments. In addition, the same reference numerals shown in the drawings denote the same members.

Storage devices are used in a variety of environments ranging from traditional systems such as PCs and server systems to more recently used portable devices (e.g., smart phones, tablet PCs, PMPs, MP3Ps, digital cameras, etc.). Storage devices connected using a standardized interface may be accessed directly by the user, but it is more common to be accessed through an abstraction such as a file system.

The file system should fetch the files and directories that the user created and saved under any circumstances. In order to do so, the contents of the recorded file data and the directory entry are important, but it becomes more important that the metadata such as the inode is correctly fetched in the recorded state. The consistency of the file system to ensure that the recorded data is retrieved exactly as it is recorded must be integrity and durability in order to maintain consistency.

In particular, it is necessary to maintain consistency in a crash situation such as an unexpected power failure at an unexpected time, unexpected unmount of a user, or a sudden system failure.

As the use of mobile devices operating with batteries is increasingly used and the mounting / dismounting of devices in the form of removable storage becomes more frequent, A crash situation may occur.

The file system can maintain consistency again through the recovery process of reading the metadata recorded in the storage device and making its state known. Therefore, after a crash situation, the file system needs a process to recover its state.

Because of this, the file system wants the write data to be persisted to the storage device in the requested order in order to maintain consistency. This is because when the write data is persisted, ordering can be restored to a consistent state again.

However, if the data that is persisted in the storage at the time of recovery is different from the intended state of the file system and the recovery process fails, for example, if the metadata is corrupted and the entire structure of the file system is lost, When the file data is restored but the metadata pointed to by the metadata is lost, various problem situations such as the case where the file data remains but the metadata is wrong may occur.

A hard disk drive (HDD) is a device capable of random access to a logical block address (LBA), and has a very low price per unit capacity. It is most widely used today as a storage device in a general computer system have. For this reason, layers for storage access, including the current host interface standard protocol and the file system of the host system, have been developed assuming that the hard disk is used as a storage device implicitly.

However, due to the structural characteristics of the hard disk, when the write requests transmitted through the storage interface are processed in order, it may cause severe performance degradation. This is because the head-arm-assembly (HAA) must be mechanically operated to move the head to a position where the data physically exists to access certain data on the storage medium in the hard disk.

Therefore, in order to overcome the structural limitation of the hard disk, a write buffer is provided in the storage device to solve some problems.

The first is that the write data is not written to the actual storage medium but stored on the write buffer, and then the completion response to the write command is given to the host system.

Second, it is possible to independently process the interface command processing and the storage medium access operation through the write buffer, so that the command queuing method can be applied.

Third, write data is first stored in the buffer and then reordered to write to the actual storage medium. The movement of the mechanically operating head-arm assembly is optimized to write data to the storage medium in a non-volatile manner Performance (persistence performance) can be improved.

The storage device wants to handle all writes asynchronously, if possible, to improve the performance of the common case. As a result of the use of the write buffer, the storage device can be accessed asynchronously as shown in FIG. 3, so that the performance of the storage system can be improved and the write request can be processed more efficiently. However, since the processing order of the write request can be reversed, the host system can not know how the write-requested data is actually persistent in the non-volatile storage medium. This increases the risk of file system inconsistencies in sudden crash situations.

Therefore, in order to keep the file system consistent even in rare cases such as a crash situation, it is necessary to accurately keep the order in which file system write requests are written to the storage medium.

Current host interface techniques provide a way to control the order in which write data is persisted on a storage medium by synchronously processing write requests as in FIG. 2, described below. However, unnecessary performance degradation may be caused by forcing synchronous writing so that the order in which write data is persisted in the storage medium is correct in order to ensure the consistency of the file system.

Therefore, the host system basically uses the asynchronous storage write method in order to efficiently process the write request to the storage device.

In the case of the asynchronous storage device writing method, reordering is performed in the buffer cache of the host system and the write buffer of the storage device, respectively, as shown in FIG. 3, which will be described later. In the case of the order reordering sequence occurring in the buffer cache, since it is an operation occurring in the host system, it is possible to confirm reordering through internal communication with the file system, and also provides a room for optimizing write request processing through internal communication .

However, the order reordering that occurs in the write buffer of the storage device has a different problem. The host system can only access the storage device through the interface of the storage device. Currently, the interface technology of the storage device does not provide a method of directly transmitting the write data persistence order to the storage device. Therefore, the host system does not know whether or not the storage device has rearranged the order of persistence for data in the write buffer.

The host system transfers the write command to the storage device, and upon receipt of the command completion, the host system can recognize that the processing of the write command is completed. However, the completion of the write command does not mean the completion of the data persistence.

In general, for a storage device that uses write buffers, it responds to completion of the write command if the data is successfully stored in the write buffer. Since it is an independent task that the storage device internally performs to persist data in the write buffer to the storage medium, the host system can not know whether the data of the write command that has been completed is persistent. In addition, when a plurality of write commands are transmitted, it is also impossible to know in what order the write data for each command is made persistent to the storage medium.

Therefore, the general host interface technology controls the persistence order of write data through the following three methods.

The first is a write-through method. This method disables the write buffer of the storage device, bypasses the write buffer, and persists the data directly to the storage medium when all write commands are processed. In this method, each write request is processed by a synchronous storage write method as shown in FIG. 2, which will be described later, and the completion response time for the write command becomes very long, which may cause performance degradation.

The second is a method of using a buffer flush command, a processing of a write command by an asynchronous storage write method, and a method of performing a synchronization through a buffer flush command when it is necessary to control a persistence order.

The buffer flush command is a command that operates synchronously, meaning that all write data stored voluntarily in the storage device is stored non-volatile on the storage medium. Upon receipt of the buffer flush command, the storage device perpetuates all data on the write buffer and responds to completion of the buffer flush command after the persistence operation is completed. This is handled synchronously with a kind of barrier synchronization method, so that other write commands can not be transferred to the storage while the buffer flush command is being processed. In addition, since the completion response time of the buffer flush command includes the time required for the persistence operation, it is a main factor for the performance degradation of the entire system.

Therefore, in the case of a storage device having a large write buffer size or a policy of leaving a lot of data in a volatile state in the write buffer, the response time of the buffer flush may be longer.

The third is to use the FUA option. The FUA has the meaning of storage medium forced access (Forced Unit Access) and is an option that can be individually applied to each command. In the case of a write command with the FUA option set, the write buffer is bypassed, and the write completion data must be persisted to the storage medium before the command completion response is sent. The FUA option only works for individual commands, and there is no rule for the persistence order between the FUA command and other FUA commands, or between FUA commands and non-FUA commands.

All three methods use synchronous storage writes, which can be referred to as 'W immediate materialization'.

If the host system needs sequential control of the persistence, it processes each of the plurality of write commands in a synchronous write manner, or writes the write command asynchronously, followed by a synchronous buffer flush command, .

As a result, what the host system wants is to ensure the order of persistence so that the write-requested data can be stored on the storage medium in the order desired by the file system. However, the interface of the existing storage device is optimized only for the structural characteristics of the hard disk, thus providing limited tools. Therefore, in order to control the order of persistence between write data, synchronization occurs when the host system becomes the subject by using the synchronous storage write method. However, there is a kind of durability bottleneck that causes excessive degradation of the performance of the storage system due to excessive synchronization between the host system and the storage device, which may cause a large performance degradation.

Therefore, there is a need to expand and modify the storage device interface so that the host system can handle the most efficiently while ensuring the desired accuracy of the persistent write-requested data to the storage medium.

Memory-based storage devices such as SSD (Soli State Drive) or NVRAM (Non-volatile memory) storage devices based on NAND flash memory have recently appeared and gradually replace the area of the hard disk. Such a memory - based storage device has little difference in performance between sequential access to continuous areas and random access to discrete areas.

In other words, memory-based storage devices can keep the order of data persistence requested by the file system as it is without degrading performance, as opposed to having to write back to the nonvolatile storage medium in order to improve performance on the hard disk . In other words, in an environment under a new storage medium, unlike a hard disk, it is easy to control the order in which the storage device permanently writes data to the storage medium, and the persistence order control request requested by the host system can be easily ensured.

Despite these advantages, existing storage system components and host interface technologies have historically evolved to be optimized for the structural characteristics of hard disks. Therefore, even in a storage device using a new storage medium, synchronous immediate persistence requests frequently occur, which may inevitably degrade performance. Therefore, it is necessary to improve the persistence operation more efficiently by reflecting the characteristics of the new media.

1 is a conceptual diagram illustrating a hierarchical structure of a computer system including a storage device.

1, a computer system may include a host system 110, a host interface 130, and a storage device 150.

The host system 110 generally operates based on an operating system (OS) and starts a task of writing and reading data to the storage device 150 by a direct request from the user or a request generated by an application driven by the user .

At this time, the operating system abstracts and displays the storage device 150 as a layer such as a file system or a database management system (DBMS). The storage device 150 may be, for example, a file or directory of a file system, It can store metadata such as the inode of the file system, as well as user data such as tables and records of the management system (DBMS).

The host interface 130 shows the storage device 150 as a set of sectors 512B or blocks that are basic units of the storage space and stores the same as a logical block address space (LBA) Thereby making the host system 110 accessible.

A storage device 150 refers to an auxiliary storage device capable of storing (storing or writing) data in a meaningful manner and retrieving stored data as it is retrieved (read or read).

The storage device 150 may be connected to the host system 110 through a standardized interface such as ATA or SCSI and may receive the interface command generated by the host system 110, (Write) and fetch (read) operations.

Hereinafter, it is assumed that a file system is used in order to simplify the explanation and facilitate understanding.

The write request (1) of the file system 111 is transferred to the storage device 150 through several layers. The process of writing the data requested to be written completely (nonvolatilely) to the actual storage medium 155 It is defined as materialization or persistent write.

The request of the file system 111 is transferred to the storage medium 155 of the storage device 150 through various steps such as the host interface 130 and the write buffer 153 in the storage device 150, The following is a brief description of the process of being perpetuated.

First, when a write request (1) of the file system 111 occurs, the write request is stored in the buffer cache 113. The buffer cache 113 is intended to be used to cache a block accessed by the file system 111, in order to alleviate the bottleneck due to access to a relatively slow storage device. By using such a buffer cache 113, unnecessary input / output (I / O) operations of the storage device can be reduced and access to data blocks between different processes can be synchronized.

The write request and write data stored in the buffer cache 113 may be converted into an I / O request (2) to the device driver 115 according to a predetermined policy. The I / O scheduler 113 may merge and reorder write requests in the buffer cache 113 to increase the efficiency of access to the storage device 150.

The I / O request (2) transferred to the device driver 115 is converted into a command (3) of the host interface and transferred to the storage device 150. The write data transferred through the write command is stored in the storage device 150 And is written to the write buffer 153 which is a volatile memory.

The data in the write buffer 153 is persistent to the storage medium 155 by the storage controller 151. As in the case of the I / O scheduler of the host system, the persistent request (4) They can be merged or reversed in order by policy. In FIG. 1, each step conveys the request to the lower level, and the completion of the request can be known through the responses (⑤, ⑥, ⑦, ⑧) from the lower level.

The processing of such a file system write request is largely accomplished in synchronous with the use of the buffer cache 113 of the host system 110 and the write buffer 153 of the storage device 150, And writing methods.

Hereinafter, the synchronous storage write method will be described with reference to FIG. 2, and the asynchronous storage write method will be described with reference to FIG.

2 is a diagram illustrating a process in which a write request of a file system is synchronously performed in a storage device.

Referring to FIG. 2, according to the synchronous storage device writing method, it is seen that the process from the request of the file system to the access to the storage medium, and the process from the completion of the persistence of the storage medium to the completion of the writing of the file system are serialized .

In this case, the response time for receiving the write completion response to the write request becomes very long in the file system. In addition, the synchronous storage write method is very inefficient because one write request can not be processed until another write request is completed.

3 is a diagram illustrating a process in which a write request of a file system is performed asynchronously in a storage device.

Referring to FIG. 3, a process according to the asynchronous storage device writing method can be examined.

According to the asynchronous storage write method, the buffer cache sends a write complete response (8) to the file system immediately after receiving the write request (1). Accordingly, the file system can accept the write-completed data even though the actual write-requested data is not sent to the storage device, and can forward the next write request to the buffer cache.

The write data accumulated in the buffer cache of the host system is merged or reordered by the I / O scheduler, converted into an I / O request (2), and stored in the form of a command (3) of the host interface Device.

The storage device stores the write data in the write buffer in the storage device, then notifies the host system of the completion of the write I / O (7) in response to the completion of the write command (6). The storage controller then makes the data on the write buffer in the storage device persistent to the storage medium (4) and (5).

In such an asynchronous storage write method, it is possible to complete a request with a shorter response time in an upper layer through use of a write buffer, thereby obtaining better processing performance. Also, by using the write buffer in the host system and the storage device, the write request of the file system, the write command of the host interface, and the data persistence process of the storage device can be independently performed, and the entire storage system can be used more efficiently .

However, when the completion response to the upper layer is notified that the operation is not finished, the problem may occur in a crash situation such as power supply interruption. Furthermore, since the volatile write buffer is used, there is a problem that data may be lost. Therefore, even in the case of an asynchronous storage write method, synchronization between independent processes is indispensable.

The limitations of the general storage device interface technology are summarized as follows.

First, since the request of the host system is processed in units of interface commands, data transmission and processing are performed on a command-by-command basis. Therefore, it is impossible to efficiently control a configuration in which a plurality of data is persistent to a storage medium in response to a write request of the file system. Thus, additional synchronization in the host system is required and the overall write performance is degraded.

Next, in the case of the write command, although it is possible to directly control the data transfer order from the host system to the storage device through the interface, it does not provide a method of controlling the order in which data is persistent to the storage medium. Therefore, in order to control the order of persistence, the host system performs the synchronization which is the subject, which degrades the performance.

Accordingly, in one embodiment, in order to correctly transmit the persistence intention of the file system to a plurality of write requests, the write data is divided into a command unit or a group of commands through a mode setting of the storage device to define a materialization unit And write data belonging to each persistent unit can be made persistent with atomicity in writing to the storage medium.

In addition, in one embodiment, write data belonging to each persistent unit may be persisted with atomicity in writing to the storage medium, and sequentially stored in the order of being transmitted to the host interface through the storage device mode setting So that the order of persistence can be controlled.

4 is a flowchart illustrating a method for controlling persistence in a storage medium according to one embodiment. Referring to FIG. 4, the operation of the storage system according to the atomic persistence mode can be examined.

A host system according to one embodiment may generate 410 a plurality of write commands for writing data to a storage medium.

The host system may insert at least one group start delimiter and at least one group end delimiter between the plurality of write commands (420). The group start delimiter is a delimiter indicating the start of at least one group for a plurality of write commands. The group start delimiter may, for example, take the form of G _start as in 630 of FIG.

Also, the group end delimiter is an identifier indicating the end of at least one group for a plurality of write commands. The group end delimiter may, for example, have the form of G _End as in 630 of FIG.

At 420, the host system may include a group start delimiter in the first write command among the write commands that are included in at least one group of the plurality of write commands and must be permanently persistent in the storage medium. Then, the host system can insert the group end delimiter into the last write command included in the same group as the first write command. The atomic persistence mode can be set and released by inserting such group start delimiter and group end delimiter.

The storage device may identify at least one group from the plurality of write commands using at least one group start delimiter and at least one group end delimiter (430).

The storage device may determine persistence order or persistence atomicity for a plurality of write commands based on the identified at least one group (440).

At this time, in order to determine the persistent atomicity, the storage device can determine whether a group identifier for distinguishing one group from another group among the groups is included. Here, one group is started by the group start delimiter and can be ended by the group end delimiter.

The group delimiter may have the form of, for example, G _Delim at 630 in FIG.

If the group end delimiter is included in a single group that is started by the group start delimiter and terminated by the group end delimiter, each group can be distinguished based on the group end delimiter.

The storage device may process a plurality of write commands based on the determination result (450).

The specific operation of the storage system according to the atomic persistence mode will be described in detail with reference to FIG.

5 is a flowchart illustrating a method for controlling persistence in a storage medium according to another embodiment. Referring to FIG. 5, operation of the storage system can be examined when the atomic persistence mode and the sequential persistence mode are mixed according to an embodiment.

The host system may generate a plurality of write commands for writing data to the storage medium (510).

The host system may insert at least one group start delimiter and at least one group end delimiter between the plurality of write commands (520).

Wherein at least one group start delimiter indicates the start of at least one group for a plurality of write commands. And at least one group termination delimiter may indicate termination of at least one group for a plurality of write commands.

In addition, the host system inserts the mode start delimiter into the first write command among the write commands that must be made persistent in the storage medium, inserts the mode end delimiter into the last write command among the write commands that must be executed in order, (530).

Here, the mode start delimiter is for setting a sequential persistence mode, and the mode end delimiter is for canceling a sequential persistence mode. The sequential persistence mode is a mode for setting a persistence order between write data corresponding to a plurality of write commands in a storage medium. While the sequential persistence mode is set, persistence of the write data may be performed in the order inputted through the host interface with respect to the write commands affected by the mode setting. The mode start delimiter may have the same form as, for example, O _{Start in} FIG. 9, and the mode end delimiter may be, for example, as shown in FIG. 9 O _End of the same.

The storage device may identify at least one group from the plurality of write commands using at least one group start delimiter and at least one group end delimiter (540).

The storage device may determine a persistence order or persistence atomicity for a plurality of write commands based on the identified at least one group (550).

The storage device can determine the persistence order according to whether a mode start delimiter and a mode end delimiter are inserted between a plurality of write commands.

If it is determined in step 550 that there is a persistence order for a plurality of write commands, the storage device stores the write command after the sequential persistence mode is set by the mode start delimiter among the plurality of write commands and before the sequential persistence mode is released by the mode end delimiter The write data corresponding to the write commands up to the write command can be sequentially persisted to the storage medium in order to be transmitted through the host interface (560).

The storage device may be configured to execute the write commands from at least one of the plurality of write commands after the atomic persistence mode is set by the group start delimiter to at least one group termination delimiter before the atomic persistence mode is released. (570) whether or not the suspended persistence has been completed. Herein, the atomic persistence mode may be a mode in which persistence in a storage medium is made integral with respect to write commands.

In step 570, if the storage device is stopped before completion of the persistence, the storage device makes all the information included in the write commands until the atomic persistence mode is released after the atomic persistence mode is set to the non-persistent state (580).

However, if the operation is interrupted after the persistence is completed at 570, the storage device may terminate the operation immediately without performing additional operations.

The specific operation of the storage system when the atomic persistence mode and the sequential persistence mode are mixed will be described with reference to FIG.

FIG. 6 illustrates a method of representing an atomic persistence request according to a method of controlling persistence in a storage medium according to an exemplary embodiment of the present invention. Referring to FIG.

Referring to FIG. 6, it can be seen that the write data is abstracted in units of persistence. Here, the persistence unit may be a write command or a group of write commands consisting of a plurality of write commands.

In order to represent a persistence unit for a plurality of write commands, such as 610, the host system prepares a group start delimiter (G) in front of a first write command (e.g., W ₀ ) that should be included in one persistence unit, _Start ) to insert the atomic persistence mode. In addition, the host system may release the atomic persistence mode by inserting a group end delimiter (G _End ) after the last write command (e.g., W ₂ ) that should be included in a persistence unit.

That is, while the atomic persistence mode is set, the write commands entered can be expressed as a group to convey the intention to be persistent with atomicity. For this purpose, in one embodiment, "setting the atomic persistence mode" and "releasing the atomic persistence mode" can be added as a sub-command of the mode change command of the existing storage device.

For example, the ATA interface may provide the Set Features command to cause the host system to change the operating mode of the storage device. Thus, in one embodiment, ATD interface can be implemented by adding / disabling the atomic persistence mode as a sub command of the Set Features command.

In addition, when a group delimiter (G _delim ) is input while the atomic persistence mode is set, a new atomic group can be started with the group delimiter as a boundary.

The group delimiter can also be implemented as a subcommand of the storage device mode change command.

In one embodiment, consecutive write commands in this manner may belong to the same persistence group to convey the atomicity requirement of persistence. Write commands within a persistence group can be persisted in any order and only atomicity of persistence can be guaranteed.

In one embodiment, W ₀ , W ₁ , W ₂ , ... The group start delimiter (G _Start ), the group end delimiter (G _End ), and the group delimiter (G _delim ) are inserted in the write commands 610 of W ₉ to set and release the atomic persistence mode Device.

Through this, the storage device can extract a logical atomic persistence unit such as 650.

7 is a flowchart illustrating a method for controlling persistence in a storage medium according to an embodiment. Referring to FIG. 7, the operation of the storage system according to the sequential persistence mode can be examined.

A host system according to one embodiment may generate 710 a plurality of write commands for writing data to a storage medium.

The host system may insert a mode start delimiter indicating the start of the sequential persistence mode and a mode end delimiter indicating the end of the sequential persistence mode between a plurality of write commands.

More specifically, the host system may insert a mode start delimiter for setting the sequential persistence mode in the first write command among the write commands in which persistence in the storage medium must be performed in order (720).

The host system may insert a mode termination delimiter for releasing the sequential persistence mode to the last write command among the write commands that must be performed in order of persistence (730).

Here, the sequential persistent mode is a mode for setting a persistence order between a plurality of write commands in the storage medium, which is set by inserting a mode start delimiter, and can be released by inserting a mode end delimiter.

The storage device may determine a persistence order for a plurality of write commands using a mode start delimiter and a mode end delimiter (740).

At this time, the storage device can determine the persistence order based on whether a mode start delimiter for setting a sequential persistence mode and a mode end delimiter for releasing a sequential persistence mode are inserted into a plurality of write commands.

That is, when the mode start delimiter and the mode end delimiter are inserted, the storage device can determine that there is a persistence order among the write commands.

The storage device may process a plurality of write commands based on the determination result of 740. [

That is, after the sequential persistence mode is set by the mode start delimiter among the plurality of write commands, and the write data corresponding to the write commands until the sequential persistence mode is released by the mode end delimiter, And sequentially persisted to the storage medium in order of transmission through the interface (750).

The specific operation of the storage system according to the sequential persistence mode will be described with reference to FIG.

FIG. 8 is a diagram illustrating a method for expressing a persistence order and a persistent atomicity control request according to a method for controlling persistence in a storage medium according to an embodiment.

Referring to FIG. 8, an extension of the interface technology for maintaining backwards compatibility with existing host interface technology and performing persistence in a storage medium according to an exemplary embodiment can be examined.

The storage device operation mode can be changed by adding / deleting the " sequential persistence mode "and" atomic persistence mode "to the mode change command of the existing storage device as shown in FIG. At this time, the sub-command can be selected so as not to overlap the identification code of the sub-command defined previously. For example, compatibility with the existing technology can be maintained by selecting a different code from the existing code cache enable / disable command.

The method of setting / releasing the sequential persistence mode, setting / releasing the atomic persistence mode, and specifying the group identifier can be variously implemented according to the embodiment.

In addition to the method of transferring to the subcommand of the storage device mode change command as shown in FIG. 8, there are five kinds of information (setting / release of the sequential persistence mode and setting of the atomic persistence mode / Release, and group delimiter) can be inserted and transmitted, or an independent command can be defined. In addition, it is also possible to include such information in an out-of-band spare data area of write data.

In addition, even if the host system and the storage system are physically present in the same system in the storage system, the storage system can control the persistence in the storage medium through arbitrary request transmission in a logically distinct environment It may be applied.

This is not a structure in which a physically separated storage device such as a general hard disk (HDD) and an SSD is connected via a standard interface such as SATA or SAS, but a storage device and a host system are physically connected to one system ), And even when the storage device receives a request through various request delivery methods (for example, function call or message transfer) inside the system.

In an embodiment, for convenience of explanation, a method of transferring to a sub command of a storage device mode change command will be described.

9 is a diagram illustrating a method for transferring a sequential persistence request for a part of write commands according to a method of controlling persistence in a storage medium according to an exemplary embodiment.

Like the atomic persistence request, the write data corresponding to the write commands entered while the sequential persistence mode is set can be sequentially persisted until the sequential persistence mode is released.

To this end, it is possible to add "sequential persistence mode setting" and "sequential persistence mode release" as subcommands of the storage device mode change command.

In one embodiment, an interface extension technique may be applied to a series of write commands to deliver sequential persistence requests for some write commands.

9, a sequential persistence mode is set by inserting a mode start delimiter (O _Start ) between write commands W ₀ and W ₁ , and a mode end delimiter (O _End ) is inserted between write commands W ₂ and W ₃ So that the sequential persistence mode can be released.

In addition, a sequential persistence mode is set by inserting a mode start delimiter (O _Start ) between write commands W ₄ and W ₅ , and a mode end delimiter (O _End ) is inserted between W ₈ and W ₉ to release the sequential persistence mode .

In this case, the storage device is made up of W ₂ -> W ₁ and W ₈ -> W ₇ -> W ₆ -> W ₅ , and the rest can be persisted in any order. In addition, W ₂ -> W ₁ and W ₈ -> W ₇ -> W ₆ -> W ₅ Write order columns can also be independently persisted without having to guarantee mutual order.

10 is a diagram illustrating a method for transferring an atomic persistence request according to a method of controlling persistence in a storage medium according to an exemplary embodiment of the present invention. Referring to FIG. 10, the operation of the storage system in the atomic persistence mode can be examined when the sequential persistence mode is not used.

The storage system can define a persistence unit that is atomically persistent by inserting / disabling atomic persistence mode and group delimiters between a series of write commands. In this case, the storage device must atomically persist each of the persistent groups W ₁ , W ₂ , W ₅ , W ₆ , W ₇ , W ₈ , and W ₉ , There is no order dependency between the two.

11 is a diagram illustrating a method of using a sequential persistence mode and an atomic persistence mode according to a method of controlling persistence in a storage medium according to an exemplary embodiment of the present invention. Referring to FIG. 11, the operation of the storage system can be examined when the sequential persistence mode and the atomic persistence mode are combined.

The sequential persistence mode and the atomic persistence mode according to an embodiment can be combined, and the effect of using the two modes in combination can be examined through Table 1 below.

If both the sequential persistence mode and the atomic persistence mode are set, the persistence sequence between the groups can also be controlled.

All write commands are entered with the sequential persistence mode set and atomically persistent for (W ₁ , W ₂ ), (W ₃ , W ₄ ), (W ₇ , W ₈ , W ₉ ) Required.

In this case, the storage device has an order dependency of (W _{1 ,} W ₂ ) -> W ₁ - (W ₇ , W ₈ , W ₉ ) -> W ₆ -> W ₅ -> (W ₃ , W ₄ ) You can keep and handle persistence. In this case, as in the case of FIG. 10, the write commands belonging to the same atomic persistence group may be persistent regardless of the order.

If a persistent configuration control request (for example, a persistent atomicity / order control request) from a host system is followed and persisted, the storage device always maintains a consistent persistence form on the file system . Thus, the file system can always restore the write commands to a consistent state, so it is not necessary to inform the host system of the result of the persistence control request. Just because the point of persistence is slowed down a bit, the storage device will perform the persistence process as intended by the host system.

12 is a diagram illustrating a result of applying a method of controlling persistence in a storage medium according to an embodiment to a data structure of a Serial ATA (SATA) Set Features command. Referring to FIG. 12 and the following Table 3, an embodiment of the sub-command code table of the SATA interface expansion technique can be examined.

It is aimed at Serial ATA, which is widely used in computer systems for general users such as PCs and laptops. It extends sub-command of Serial ATA's Set Features command to request sequential persistence, You can see that you can define persistent persistence groups.

12 shows a data structure of the Set Features command of the SATA interface. Here, if the command code EFh is written in the Byte 2 of the 0th dword, it is recognized as the Set Features command, and the host system can change the operation mode of the storage device according to the code of the sub-command written in Byte 3 of the 0th dword . The code of the sub-instruction that can control the persistent shape according to the embodiment is shown in [Table 2].

Since the code of the newly added subcommand does not overlap with the code of the subcommand used in the existing SATA interface shown in [Table 3], it is possible to express the function for controlling the persistent shape while maintaining compatibility with the existing interface .

13 is a block diagram of a storage system that controls persistence in a storage medium according to one embodiment. 13, a storage system 1300 according to one embodiment may include a host interface 1310, a controller 1330, and a storage medium 1350. The host interface 1310 may include a plurality of write commands for writing data to the storage medium from the host system and at least one group of write commands inserted between the plurality of write commands, At least one group start delimiter indicating the start of at least one group and at least one group end delimiter indicating the end of at least one group.

Here, the group start delimiter may be included in at least one group of the plurality of write commands and inserted into the first write command among the write commands that must be permanently persistent in the storage medium.

Also, the group end delimiter may be inserted into the last write command included in the same group as the first write command.

The controller 1330 may identify at least one group from the plurality of write commands using at least one group start delimiter and at least one group end delimiter. The controller 1330 determines a persistence order or persistence atomicity for a plurality of write commands based on the identified at least one group, and can process a plurality of write commands based on the determination result.

The controller 1330 determines whether at least one group started by at least one group start delimiter and ending by at least one group end delimiter includes a group delimiter for distinguishing at least one group from another group Can be determined.

The controller 1330 determines whether or not at least one of the plurality of write commands includes at least one group start delimiter after the atomic persistence mode is set by the at least one group start delimiter, The information contained in the write commands from the time when the atomic persistence mode is set to the time when the atomic persistence mode is released can be made non-persistent.

Herein, the atomic persistence mode may be a mode in which persistence in a storage medium is made integral with respect to write commands.

The mode start delimiter for setting the sequential persistence mode may be inserted into the first write command among the write commands that must be made persistent in the storage medium in order. Also, the mode termination delimiter for releasing the sequential persistence mode can be inserted into the last write command among the write commands for which persistence is to be performed in order.

The sequential persistence mode may be a mode for setting a persistence order between write data corresponding to a plurality of write commands in the storage medium.

The controller 1330 can determine whether a mode start delimiter indicating the start of the sequential persistence mode and a mode end delimiter indicating the end of the sequential persistence mode are inserted between the plurality of write commands.

The controller 1330 writes the write data corresponding to the write commands until the sequential persistence mode is released by the mode end delimiter after the sequential persistence mode is set by the mode start delimiter among the plurality of write commands, It is possible to sequentially perform persistence on the storage medium in the order in which they are transmitted through the interface.

In the storage medium 1350, a plurality of write commands may be made persistent according to the processing result.

At this time, the storage device including the storage medium 1350 can notify the host system through the host interface 1310 that the storage device is a device capable of controlling the persistent configuration related to persistence order or persistent atomicity.

In addition, the host system can transmit information to the storage device via the host interface 1310, which enables the storage device capable of the persistent configuration control to use the persistent configuration control function.

14 is a block diagram of a storage system that controls persistence in a storage medium according to another embodiment.

14, a storage system 1400 according to another embodiment may include a host interface 1410, a controller 1430, and a storage medium 1450.

The host interface 1410 includes a plurality of write commands for writing data to the storage medium from the host system, a mode start delimiter indicating the start of the sequential persistence mode and a mode end delimiter indicating the end of the sequential persistence mode Lt; / RTI >

Here, the sequential persistence mode may be a mode for setting a persistence order between write data corresponding to a plurality of write commands in the storage medium. Also, the mode start delimiter may be inserted into the first write command among the write commands that must be made persistent in the storage medium in order. The mode termination delimiter can be inserted into the last write command among the write commands in which persistence must be performed in order.

The controller 1430 can determine a persistence order for a plurality of write commands using the mode start delimiter and the mode end delimiter, and process the plurality of write commands based on the determination result.

The controller 1430 can determine whether a mode start delimiter for setting the sequential persistence mode and a mode term delimiter for releasing the sequential persistence mode are inserted into a plurality of write commands.

The controller 1430 writes the write data corresponding to the write commands before the sequential persistence mode is released by the mode end delimiter after the sequential persistence mode is set by the mode start delimiter among the plurality of write commands, It is possible to sequentially perform persistence on the storage medium in the order in which they are transmitted through the interface.

A plurality of write commands may be persisted in the storage medium 1450 according to the processing result.

The above-described methods may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the medium may be those specially designed and constructed for the present invention or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

While the invention has been shown and described with reference to certain preferred embodiments thereof, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. This is possible.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined by the equivalents of the claims, as well as the claims.

110: host system
111: File system
113: Buffer Cache and I / O Scheduler
115: Device driver
130: Host interface
150: Storage device
151: Storage controller
153: write buffer
155: Storage medium

Claims (25)

Generating a plurality of write commands for writing data to a storage medium;
Inserting at least one group start delimiter indicating the start of at least one group for the plurality of write commands between the plurality of write commands and at least one group end delimiter indicating the end of the at least one group ;
Identifying at least one group from the plurality of write commands using the at least one group start delimiter and the at least one group end delimiter;
Determining a persistence order or persistence atomicity for the plurality of write commands based on the identified at least one group; And
Processing the plurality of write commands based on the determination result
Lt; / RTI >
The processing of the plurality of write commands
Wherein the at least one group performs at least one persistence of at least one write command included in the at least one group when the at least one group does not include a group identifier for distinguishing the at least one group from another group, Lt; / RTI > The method according to claim 1,
Wherein inserting the at least one group start delimiter and the at least one group end delimiter comprises:
Inserting the group start delimiter into a first write command among write commands included in the at least one group among the plurality of write commands and to be permanently persistent in the storage medium; And
Inserting the group end delimiter into a last write command included in the same group as the first write command
≪ / RTI > The method according to claim 1,
Wherein the step of determining persistent atomicity for the plurality of write commands comprises:
Whether or not a group delimiter for distinguishing the at least one group from another group is included among at least one group started by the at least one group start delimiter and terminated by the at least one group end delimiter Step to judge
≪ / RTI > The method according to claim 1,
The processing of the plurality of write commands
Wherein at least one of the plurality of write commands includes at least one group start delimiter and at least one group end delimiter for each write command from the time at which the atomic persistence mode is set by the at least one group start delimiter until the atomic persistence mode is released by the at least one group end delimiter, If the ongoing persistence is stopped before completion,
Making all the information included in the write commands until the atomic persistence mode is released after the atomic persistence mode is set to a state in which the information contained in the write commands is not persistent
Further comprising:
The atomic persistence mode
Wherein the persistence in the storage medium is consistent with the write commands in the storage medium. The method according to claim 1,
Inserting a mode start delimiter for setting a sequential persistence mode in a first write command among write commands in which persistence in the storage medium must be performed in order; And
Inserting a mode termination delimiter for releasing the sequential persistence mode into the last write command among the write commands that must be performed in order of persistence
Further comprising:
The sequential persistence mode
Wherein a persistence order is set between write data corresponding to a plurality of write commands in the storage medium. The method according to claim 1,
The step of determining the persistence order for the plurality of write commands
Determining whether a mode start delimiter indicating the start of the sequential persistence mode and a mode end delimiter indicating the end of the sequential persistence mode are inserted between the plurality of write commands
≪ / RTI > The method according to claim 6,
The processing of the plurality of write commands
Wherein the writing control unit is configured to write data corresponding to write commands from the plurality of write commands until the sequential persistence mode is set by the mode start delimiter and before the sequential persistence mode is released by the mode term identifier, Sequentially sequential to the storage medium in the order in which they are transmitted through the storage medium
≪ / RTI > Generating a plurality of write commands for writing data to a storage medium;
Inserting a mode start delimiter indicating the start of the sequential persistence mode and a mode end delimiter indicating the end of the sequential persistence mode between the plurality of write commands;
Determining a persistence order for the plurality of write commands using the mode start delimiter and the mode end delimiter; And
Processing the plurality of write commands based on the determination result
Lt; / RTI >
The sequential persistence mode
A mode for setting a persistence order between write data corresponding to a plurality of write commands in the storage medium,
The processing of the plurality of write comments
Wherein the write data corresponding to the write commands inputted while the sequential persistence mode is set among the plurality of write commands is sequentially persistent until the sequential persistence mode is released based on the determination result How to control persistence. 9. The method of claim 8,
The step of inserting the mode start delimiter and the mode end delimiter
Inserting a mode start delimiter for setting a sequential persistence mode in a first write command among write commands in which persistence in the storage medium must be performed in order; And
Inserting a mode termination delimiter for releasing the sequential persistence mode into the last write command among the write commands that must be performed in order of persistence
≪ / RTI > 9. The method of claim 8,
The step of determining the persistence order for the plurality of write commands
Determining whether a mode start delimiter for setting the sequential persistence mode and a mode end delimiter for releasing the sequential persistence mode are inserted into the plurality of write commands
≪ / RTI > 11. The method of claim 10,
The processing of the plurality of write commands
Wherein the writing control unit is configured to write data corresponding to write commands from the plurality of write commands until the sequential persistence mode is set by the mode start delimiter and before the sequential persistence mode is released by the mode term identifier, Sequentially sequential to the storage medium in the order in which they are transmitted through the storage medium
≪ / RTI > A computer-readable recording medium having recorded thereon a program for performing the method according to any one of claims 1 to 11. A plurality of write commands for writing data to a storage medium in the storage device from the host system and at least one group start delimiter indicating the start of at least one group inserted between the plurality of write commands, A host interface for receiving at least one group termination delimiter indicating termination;
Identifying at least one group from the plurality of write commands using the at least one group start delimiter and the at least one group end delimiter, and identifying at least one group from among the plurality of write commands based on the identified at least one group A controller for determining the persistence order or the persistence atomicity and processing the plurality of write commands based on the determination result; And
And a storage device for holding the plurality of write commands in accordance with the processing result
Lt; / RTI >
The controller comprising:
Wherein the at least one group performs at least one persistence of at least one write command included in the at least one group when the at least one group does not include a group identifier for distinguishing the at least one group from another group, A storage system that controls persistence in a storage system. 14. The method of claim 13,
The group start delimiter
Wherein the at least one of the plurality of write commands is embedded in a first write command among write commands that are included in at least one group and must be permanently persistent in the storage medium,
The group end identifier
Wherein the persistence in the storage medium inserted in the last write command included in the same group as the first write command is controlled. 14. The method of claim 13,
The controller
Whether or not a group delimiter for distinguishing the at least one group from another group is included among at least one group started by the at least one group start delimiter and terminated by the at least one group end delimiter A storage system for controlling persistence in a storage medium to be determined. 14. The method of claim 13,
The controller
Wherein at least one of the plurality of write commands includes at least one group start delimiter and at least one group end delimiter for each write command from the time at which the atomic persistence mode is set by the at least one group start delimiter until the atomic persistence mode is released by the at least one group end delimiter, If the ongoing persistence is stopped before completion,
The information included in the write commands from the time when the atomic persistence mode is set until the atomic persistence mode is released to the state where the atomic persistence mode is released,
The atomic persistence mode
Wherein the persistence in the storage medium is controlled in such a manner that persistence in the storage medium is made integral with respect to the write commands. 14. The method of claim 13,
The mode start delimiter for setting the sequential persistence mode is
Wherein persistence in the storage medium is inserted into a first write command of write commands that must be performed in order,
The mode termination identifier for releasing the sequential persistence mode is
The persistence is inserted into the last write command among the write commands to be executed in order,
The sequential persistence mode
Which is a mode for setting a persistence order between write data corresponding to a plurality of write commands in the storage medium. 14. The method of claim 13,
The controller
Wherein a mode start delimiter indicating a start of a sequential persistence mode and a mode end delimiter indicating an end of the sequential persistence mode are inserted between the plurality of write commands. 19. The method of claim 18,
The controller
Wherein the writing control unit is configured to write data corresponding to write commands from the plurality of write commands until the sequential persistence mode is set by the mode start delimiter and before the sequential persistence mode is released by the mode term identifier, Wherein the persistence control unit controls persistence in a storage medium that is sequentially persistent to the storage medium in order of transmission through the storage medium. 14. The method of claim 13,
The storage device
And notifies the host system via the host interface that the storage device is a device capable of controlling the persistent configuration related to the persistence order or the persistent atomicity. 21. The method of claim 20,
The host system
Wherein the persistent configuration control unit controls persistence in a storage medium that transmits information to the storage apparatus via the host interface, the information enabling the storage apparatus capable of persistent configuration control to use the persistent configuration control function. A plurality of write commands for writing data to a storage medium in a storage device from a host system and a mode start delimiter indicating a start of a sequential persistence mode and a mode end delimiter indicating an end of the sequential persistence mode, A receiving host interface;
A controller for determining a persistence order for the plurality of write commands using the mode start delimiter and the mode end delimiter and processing the plurality of write commands based on the determination result; And
And a storage medium for holding the plurality of write commands in accordance with the processing result
Lt; / RTI >
The sequential persistence mode
A mode for setting a persistence order between write data corresponding to a plurality of write commands in the storage medium,
The controller comprising:
Wherein the write data corresponding to the write commands inputted while the sequential persistence mode is set among the plurality of write commands is sequentially persistent until the sequential persistence mode is released based on the determination result A storage system that controls persistence. 23. The method of claim 22,
The mode start delimiter
Wherein persistence in the storage medium is inserted into a first write command of write commands that must be performed in order,
The mode term identifier
Wherein persistence in the storage medium is inserted in the last write command among the write commands in which the persistence is to be performed in order. 23. The method of claim 22,
The controller
Wherein the controller determines whether a mode start delimiter for setting the sequential persistence mode and a mode termination delimiter for releasing the sequential persistence mode are inserted into the plurality of write commands. 25. The method of claim 24,
The controller
Wherein the write command is written to the storage medium after the sequential persistence mode is set by the mode start delimiter among the plurality of write commands until the sequential persistence mode is released by the mode term identifier, Wherein the persistent storage in the storage medium is a persistent storage medium.

Images