WO2019021415A1

WO2019021415A1 - Storage system and data storing control method

Info

Publication number: WO2019021415A1
Application number: PCT/JP2017/027193
Authority: WO
Inventors: 誠司鳥越; 繁雄本間; 浩幸三留
Original assignee: 株式会社日立製作所; 株式会社日立情報通信エンジニアリング
Priority date: 2017-07-27
Filing date: 2017-07-27
Publication date: 2019-01-31

Abstract

A storage system 3 has a plurality of flash memory packages (FMPKs) 20 and a CPU 11. The plurality of FMPKs 20 include an FMPK 20 having a flash memory with a short write endurance and an FMPK 20 having a flash memory with a long write endurance. The CPU 11 determines whether an interface for a protocol of a write request is an object ID interface used when storing data with a low update probability or an SCSI interface used when storing data with a high update probability. If the interface is the object ID interface, the CPU 11 allocates, as a storing area to which write data is to be written, a storing area of the FMPK 20 of the flash memory with the short write endurance. If the interface is the SCSI interface, the CPU 11 allocates a storing area of the FMPK 20 of the flash memory with the long write endurance.

Description

Storage system and data storage control method

The present invention relates to a storage system or the like that stores write data corresponding to a write request.

Generally, a storage system has a plurality of storage devices for storing data, and a disk controller for controlling the plurality of storage devices. A disk controller provides a host computer with a large capacity data storage space.

In recent years, as a storage device mounted on a storage system, a non-volatile memory device (for example, an SSD: Solid State Drive) having a non-volatile memory (for example, an FM: flash memory) has come to be used .

For example, in the NAND type flash memory, the number of times data can be written is limited due to the configuration thereof, and therefore, in a non-volatile memory device having the flash memory, there is a write endurance. The write life is different depending on the type of nonvolatile memory to be used.

In storage systems, multiple types of non-volatile memory devices with different write lives may be provided. Non-volatile memory devices vary in cost depending on their performance, such as write lifetime. Therefore, in a storage system, controlling how data is stored in a non-volatile memory device is important in reducing the bit cost in the storage system.

For example, as a control technique for storing data, text data is stored in a flash memory having a low read error rate, and image data is stored in a flash memory having a high read error rate in a storage provided with a plurality of nonvolatile memories. A technique to do this is known (see, for example, Patent Document 1).

Storage systems include block storage that handles data in units of predetermined blocks with a host, and object storage that handles data in units of objects.

U.S. Pat. No. 8,874,835

Since the processing of the disk controller differs depending on whether the storage system is block storage or object storage, the storage system is generally used as any storage. However, it is also conceivable to use the storage system as block storage and object storage.

Here, assuming that the storage system is used as block storage and object storage, in order to operate as both storages, the disk controller performs I / O processing of data in block units and data in object units. It is necessary to correspond to a plurality of protocols for performing the above, and to have a function of managing data in units of blocks and data in units of objects.

In addition, when the storage system is provided with a plurality of types of nonvolatile memory devices having different write lives, which storage control of block unit data and object unit data is performed in order to reduce bit cost It will be very important how to do it.

The present invention has been made in view of the above circumstances, and an object thereof is to provide a technology capable of appropriately storing data stored according to a plurality of protocols and reducing bit cost.

In order to achieve the above object, a storage system according to one aspect includes a plurality of non-volatile memory devices each having a non-volatile memory, and a processor unit that is one or more processors managing the plurality of non-volatile memory devices. Have. The plurality of non-volatile memory devices of the storage system includes a first non-volatile memory device having a first non-volatile memory having a short initial write life and a second non-volatile memory having a high initial write life. And a non-volatile memory device.

The processor unit is a first protocol interface which is highly likely to be used when storing data with a low update possibility, the interface of the protocol to which the write request for the predetermined write data received from the host apparatus is compatible. It is determined whether the second protocol interface is highly likely to be used when storing data that is likely to be updated. When the processor unit allocates a storage area to which write data corresponding to the write request is to be written, when the interface of the protocol corresponding to the write request is the first protocol interface, the processor unit A second non-volatile memory device is allocated when a storage area is allocated and a storage area to which write data corresponding to the write request is to be written is allocated when the interface of the protocol corresponding to the write request is the second protocol interface. Allocating the storage area of and writing the write data to the allocated storage area.

According to the present invention, data stored according to a plurality of protocols can be stored properly, and bit cost can be reduced.

FIG. 1 is a diagram for explaining the outline of the first embodiment. FIG. 2 is an overall configuration diagram of a computer system according to the first embodiment. FIG. 3 is a block diagram of the FMPK according to the first embodiment. FIG. 4 is a configuration diagram of a data control table according to the first embodiment. FIG. 5 is a configuration diagram of a RAID group management table according to the first embodiment. FIG. 6 is a configuration diagram of a drive management table according to the first embodiment. FIG. 7 is a configuration diagram of a write destination management table according to the first embodiment. FIG. 8 is a configuration diagram of an object data management table according to the first embodiment. FIG. 9 is a configuration diagram of an object data write destination management table according to the first embodiment. FIG. 10 is a diagram for explaining the mapping between logical pages and physical pages in the FMPK. FIG. 11 is a configuration diagram of the mapping table according to the first embodiment. FIG. 12 is a flowchart of input / output processing by the disk controller according to the first embodiment. FIG. 13 is a flowchart of data write processing by the FMPK controller according to the first embodiment. FIG. 14 is a configuration diagram of a RAID group management table according to the second embodiment.

Several embodiments will be described with reference to the drawings. Note that the embodiments described below do not limit the invention according to the claims, and all of the elements described in the embodiments and the combinations thereof are limited to the essential means of the invention. Absent.

In the following description, information may be described by the expression “AAA table”, but the information may be expressed by any data structure. That is, the "AAA table" can be called "AAA information" to indicate that the information does not depend on the data structure.

Further, in the following description, the “processor unit” includes one or more processors. At least one processor is typically a microprocessor, such as a CPU (Central Processing Unit). Each of the one or more processors may be single core or multi-core. The processor may include hardware circuitry that performs some or all of the processing.

Further, the processing performed by the functional unit configured by the processor executing a program may be processing performed by the processor or processing performed by a computer including the processor.

Further, part or all of the processing performed by the processor may be performed by a hardware circuit. The program executed by the processor may be installed from a program source. The program source may be a program distribution server or storage medium (eg, portable storage medium).

Also, in the following description, “RAID” is an abbreviation of Redundant Array of Independent (or Inexpensive) Disks. A RAID group is composed of a plurality of physical devices (typically physical devices of the same type), and stores data according to the RAID level associated with the RAID group. RAID groups may be referred to as parity groups. The parity group may be, for example, a RAID group that stores parity.

First, an outline of the first embodiment will be described.

FIG. 1 is a diagram for explaining the outline of the first embodiment.

In the computer system 1, a host computer (referred to as a host) 2 transmits a write request (Write request) to the storage system 3 via a SAN (Storage Area Network) 4. As a write request transmitted by the host 2, a write request according to an interface (first protocol interface) of HTTP (Hypertext Transfer Protocol), which is a protocol for writing data in object units, and data in block units are performed. There is a write request according to SCSI (Small Computer System Interface: second protocol interface) which is an interface of the protocol.

The write request according to HTTP includes, for example, a command (for example, Put) requesting a write, a unique ID (object ID) of an object to be written, and data to be written (write data). In addition, for a write request according to SCSI, for example, a command (for example, Write) for requesting a write, a LUN (Logical Unit Number) indicating a write destination volume of the storage system 3, and a logic indicating an address of a write destination in the volume It includes a block address (LBA: Logical Block Address) and data to be written (write data).

In the storage system 3, when the disk controller 10 receives a write request via the SAN 4 ((1) in FIG. 1), the write request is referred to the application layer header of the Open Systems Interconnection (OSI) reference model in the write request. It is determined whether the protocol interface (I / F) of the protocol is a protocol I / F (ID I / F) according to HTTP or a protocol I / F (SCSI I / F) according to SCSI ((2 in FIG. 1). )).

As a result, when the protocol I / F is the ID I / F, it can be determined that the write data is the object data ((3) in FIG. 1), so that the disk controller 10 has a write endurance. FMPK (flash memory package) 20 including a NAND flash memory chip (FM chip) having a small amount of memory, for example, an FMPK 20A including a QLC (Quad Level Cell) FM chip capable of storing 4-bit data in one memory cell. A storage area is assigned as a storage area for storing write data, and the write data is stored in this storage area ((4) in FIG. 1).

On the other hand, if the protocol I / F is SCSI I / F, it can be determined that the write data is block unit data (block data) ((5) in FIG. 1), so the disk controller 10 writes When it is necessary to allocate the previous storage area, an FMPK 20 including a NAND flash memory chip with a long write life, here, for example, an MLC (Multiple Level Cell) method capable of storing three or more values in one memory cell. The memory area of the FMPK 20 C including the FM chip of F, and the FMPK 20 C including the FM chip of the Triple Level Cell (TLC) method capable of storing 3-bit data in one memory cell is allocated as a storage area for storing write data. Write data to storage area It is paid (in FIG. 1 (6)).

As described above, changing the type of FMPK 20 allocated to the storage area for storing the write data based on the protocol I / F is due to the difference in the usage status of the object data and the block data targeted by each protocol.

That is, object data is basically not likely to be updated, and even if an update occurs, the entire new data is written instead of overwriting the data. Therefore, the number of rewrites to the same storage area can be suppressed. Therefore, even if the object data is stored in the FMPK including the FM chip having a short writing lifetime, the decrease in the writing lifetime of the FMPK can be suppressed. From this, it can be said that the HTTP interface is a protocol interface that tends to be used when storing data with low updateability.

On the other hand, block data is likely to be frequently updated, and the number of rewrites to the storage area increases. For this reason, storing block data in the FMPK including the FM chip having a short writing life causes a decrease in the writing life of the FMPK, so the block data is stored in the FMPK including the FM chip having a long writing life. Is preferred. From this, SCSI can be said to be a protocol interface that is likely to be used when storing data that is likely to be updated.

Next, the first embodiment will be described in detail.

FIG. 2 is an overall configuration diagram of a computer system according to the first embodiment.

The computer system 1 includes a host 2 as an example of one or more upper-level devices, one or more storage systems 3, and a SAN 4 that connects the host 2 and the storage system 3. The SAN 4 may be another type of communication network.

The host 2 performs an I / O request to the storage system 3, that is, a read request to read data stored in the storage system 3 (Read request), and a write request to write data to the storage system 3 (Write request) Send to 3 Here, as a write request, the host 2 writes a write request according to the SCSI, which is an interface of a protocol for writing data in block units, and a write request according to an HTTP interface, which is a protocol for writing data in object units. Can be sent.

The storage system 3 has a disk controller 10 and a plurality of FMPKs 20.

The FMPK 20 is a storage device for storing write data from a host apparatus such as the host 2 and is a storage device (nonvolatile memory device) adopting a nonvolatile memory such as a flash memory as a storage medium. In the present embodiment, a plurality of types of FMPKs having different rewrite lifetimes and input / output performances exist in the plurality of FMPKs 20 depending on the type and performance of the non-volatile memory used.

The internal configuration of the FMPK 20 will be described later. The FMPK 20 is connected to the disk controller 10 by, for example, a transmission line (SAS link) conforming to the SAS (Serial Attached SCSI) standard, a transmission line (PCI link) conforming to the PCI (Peripheral Component Interconnect) standard, or the like.

The disk controller 10 includes a central processing unit (CPU) 11 as an example of a processor unit, a host interface (host I / F) 12, a disk interface (disk I / F) 13, a memory 14, and an internal bus 15. The internal bus 15 mutually connects the CPU 11, the host I / F 12, the disk I / F 13 and the memory 14. Although only one component of the disk controller 10 is shown in FIG. 2, at least one of these components is provided in the disk controller 10 in order to achieve high performance and high availability. May be Further, in the disk controller 10, each component may be interconnected via an internal switch instead of the internal bus 15.

The host I / F 12 converts the protocol between the communication protocol (for example, fiber channel) used in the data transfer path between the host 2 and the disk controller 10 and the communication protocol used inside the disk controller 10, host 2 and disk Processing of data transfer with the controller 10 is performed. In this embodiment, the host I / F 12 uses an object ID protocol (first protocol) used to store data in object units and a SCSI protocol (second protocol) used to store data in predetermined blocks. Communication can be performed according to the protocol.

The disk I / F 13 performs protocol conversion between a protocol (for example, SAS) used by the FMPK 20 and a communication protocol (for example, PCI-Express) used inside the disk controller 10, and data between the disk controller 10 and the FMPK 20. Perform transfer (read, write) processing.

The CPU 11 performs various controls of the storage system 3. The memory 14 may be configured by, for example, a volatile storage medium such as a DRAM or an SRAM, or may be configured by a non-volatile memory. The memory 14 stores programs executed by the CPU 11 and management information of the storage system 3 used by the CPU 11. The memory 14 is also used to temporarily store data (write data, read data) to be subjected to I / O to the FMPK 20. In the present embodiment, the memory 14 stores a program executed by the CPU 11 in order to execute processing to be described later by the disk controller 10. As management information stored in the memory 14, for example, a data control table 100 (see FIG. 4), a RAID group management table 200 (see FIG. 5), a drive management table 300 (see FIG. 6), and a write destination management table 400 described later. An object data management table 500 (see FIG. 8) and an object data write destination management table 600 (see FIG. 9) are included (see FIG. 7).

Next, the FMPK 20 will be described in detail.

FIG. 3 is a block diagram of the FMPK according to the first embodiment.

The FMPK 20 includes an FMPK controller 30 and a plurality of FM chips 40. The FMPK controller 30 has a CPU 31 as an example of a processor, a disk I / F 32, an FM chip I / F 33, and a memory 34, which are interconnected via an internal bus 35.

The disk I / F 32 is an interface controller for performing communication between the FMPK 20 and the disk controller 10. The disk I / F 32 is connected to the disk I / F 13 of the disk controller 10 via a transmission line (SAS link or PCI link).

The FM chip I / F 33 is an interface controller for communicating between the FMPK controller 30 and the FM chip 40. The FM chip I / F 33 is connected to the FM chip 40 via a transmission line.

The CPU 31 performs processing and the like related to various commands received from the disk controller 10. The memory 34 stores programs executed by the CPU 31 and various management information. The memory 34 may be configured by a volatile storage medium such as a DRAM, an SRAM, or the like, and may be configured by a non-volatile memory. The memory 34 stores a logical / physical mapping table 500 (see FIG. 11), which will be described later.

The FM chip 40 is, for example, a nonvolatile semiconductor memory chip such as a NAND flash memory. Examples of the FM chip 40 include a QLC FM chip, a TLC FM chip, and a SLC (Single Level Cell) FM chip. In the present embodiment, in the same FMPK 20, the FM chip 40 of the same system is mounted. Generally, the writing life of the FM chip 40 at the initial stage is in the order of the SLC method, the MLC method, the TLC method, and the QLC method, in descending order. In the present embodiment, for example, the QLC FM chip 40 corresponds to the first non-volatile memory, and the MLC and TLC FM chips 40 correspond to the second non-volatile memory. Therefore, the FMPK 20 including the QLC FM chip 40 corresponds to the first non-volatile memory device, and the FMPK 20 including the MLC or TLC FM chip 40 corresponds to the second non-volatile memory device.

Next, the management information managed in the memory 14 of the disk controller 10 will be described in detail.

First, the data control table 100 will be described.

FIG. 4 is a configuration diagram of a data control table according to the first embodiment.

The data control table 100 stores a corresponding entry for each I / O request received from the host 2. The entries of the data control table 100 include fields of an entry number 101, a time stamp 102, a direction 103, an I / F 104, a command 105, and data 106.

The entry number 101 stores a number for identifying an entry. The time stamp 102 stores the date and time when the I / O request corresponding to the entry is received. The direction 103 stores the direction of the I / O request, that is, whether it is a write (Write) to the storage system 3 or a read (Read) from the storage system 3. The I / F 104 stores the type of interface in the I / O request. In the present embodiment, if the I / O request is an interface conforming to the object ID protocol, an ID is stored in the I / F 104, and if it is an interface conforming to the SCSI, SCSI is stored. The command 105 stores a command corresponding to the I / O request. For example, in the case of a write request using an interface according to the object ID protocol, Put is stored, and in the case of a read request using an interface according to the object ID protocol, Get is stored, and in a write request using an interface according to the SCSI protocol. If there is, Write is stored, and if it is a read request using an interface conforming to the SCSI protocol, Read is stored. The data 106 stores read data and write data corresponding to the I / O request.

Next, the RAID group management table 200 will be described.

FIG. 5 is a configuration diagram of a RAID group management table according to the first embodiment.

The RAID group management table 200 stores an entry for managing information on RAID groups managed by the storage system 3. The entries in the RAID group management table 200 are: write endurance 201, RAID group number 202, RAID configuration 203, drive number 204, flash memory type (FM Type) 205, use 206, RAID group It includes fields with the logical address 207.

The write life 201 stores information on the extent of the initial write life for the FMPKs that make up the RAID group corresponding to the entry. In this embodiment, “many” is set when the initial write life of the FMPK 20 configuring the RAID group is long, and “low” is set when the initial write life is short. The RAID group number 202 stores a number (RAID group number) for identifying the RAID group corresponding to the entry. The RAID configuration 203 stores the configuration of the RAID group corresponding to the entry. For example, one of RAIDs 0 to 6 is set as the configuration of the RAID group. The drive number 204 stores the number of the drive (here, the FMPK 20) that configures the RAID group corresponding to the entry. The flash memory type 205 stores the type of the FM chip 40 of the FMPK 20 that configures the RAID group corresponding to the entry. Examples of the type of the FM chip 40 include QLC, TLC, and SLC. The application 206 stores the application of data stored in the FMPK 20 of the RAID group corresponding to the entry. The application of data is, for example, for object data or block data. The RAID group logical address 207 stores a logical address (RAID group logical address) in the RAID group for the storage area of the drive corresponding to the entry.

Next, the drive management table 300 will be described.

FIG. 6 is a configuration diagram of a drive management table according to the first embodiment.

The drive management table 300 is a table for managing the correspondence between the RAID group logical address in the FMPK 20 and the logical address of the FMPK 20. The entries of the drive management table 300 include fields of a drive number 301, a RAID group logical address 302, and an FMPK logical address 303.

The drive number 301 stores a number (drive number, FMPK number) for identifying the FMPK 20. The RAID group logical address 302 stores a logical address (RAID group logical address) in the RAID group. The FMPK logical address 303 stores the logical address (FMPK logical address) of the FMPK 20 corresponding to the RAID group logical address.

Next, the write destination management table 400 will be described.

FIG. 7 is a configuration diagram of a write destination management table according to the first embodiment.

The write destination management table 400 is a table for managing a write destination of write data to a certain virtual volume (for example, virtual volume according to Thin Provisioning: target volume). The entry of the write destination management table 400 includes fields of a target volume address 401, a RAID group number 402, and a RAID group logical address 403. The target volume address 401 stores each address in the target volume. The RAID group number 402 stores the number of the RAID group that stores the write data. The RAID group logical address 403 stores the logical address (RAID group logical address) of the RAID group in which the write data is stored. If no RAID group storage area is allocated to the target volume address, nothing is set in the RAID group number 402 and the RAID group logical address 403 of the entry corresponding to the target volume address. It has not been. In the write destination management table 400, write destinations related to a plurality of volumes may be managed. In this case, the entry may further include a field for storing identification information of volumes.

Here, by referring to the RAID group management table 200 using the RAID group number of the write destination management table 400 and the RAID group logical address, the drive number of the FMPK 20 in which the data of a certain address of the target volume is stored. The address of the FMPK 20 (FMPK logical address) in which data of a certain address of the target volume can be identified can be identified using the identified drive number and the RAID group logical address, and the drive management table 300 It can be identified by reference.

Next, the object data management table 500 will be described.

FIG. 8 is a configuration diagram of an object data management table according to the first embodiment.

The object data management table 500 is a table for managing objects stored in the storage system 3, and stores entries corresponding to each object. An entry of the object data management table 500 includes fields of an object ID 501, an object name 502, a length 503, and a type 504.

The object ID 501 stores information (object ID) for identifying an object corresponding to the entry. The object name 502 stores the name of the object corresponding to the entry. In the length 503, the length (data length) of the object corresponding to the entry is stored. The type 504 stores the type of object.

Next, the object data write destination management table 600 will be described.

FIG. 9 is a configuration diagram of an object data write destination management table according to the first embodiment.

The object data write destination management table 600 is a table for managing the write destination of object data (object data), and stores an entry corresponding to each object. The entries of the object data write destination management table 600 include fields of an object ID 601, a RAID group number 602, and a RAID group logical address 603.

The object ID 601 stores an object ID for identifying an object. The RAID group number 602 stores the number of the RAID group storing the object data. The RAID group logical address 603 stores the logical address (RAID group logical address) of the RAID group in which the object data is stored.

Here, the drive number of the FMPK 20 in which data of a certain object is stored is referred to the RAID group management table 200 using the RAID group number of the object data write destination management table 600 and the RAID group logical address. The address of the FMPK 20 (FMPK logical address) in which data of a certain object can be identified is stored by referring to the drive management table 300 using the identified drive number and the RAID group logical address. It can be identified.

Next, mapping between logical addresses and physical addresses in the FMPK 20 will be described.

FIG. 10 is a diagram for explaining the mapping between logical pages and physical pages in the FMPK.

In the disk controller 10, the logical recording area 50 in the FMPK 20 is managed by a logical address (FMPK logical address). In the FMPK 20, the logical storage area 50 is managed in units of logical pages 51 of a predetermined size. The logical page 51 is mapped to the physical page 53 in the physical block 52 of the FMPK 20.

The logical address of the FMPK 20 and the physical address are managed by the logical-physical mapping table 700.

FIG. 11 is a configuration diagram of a logical / physical mapping table according to the first embodiment.

The logical / physical mapping table 700 stores an entry indicating the correspondence between the logical address in the FMPK 20 and the physical address.

The entry stores fields of a logical address 701 and a physical address 702. The logical address 701 stores the logical address (FMPK logical address) in the FMPK 20. The physical address 702 stores the physical address corresponding to the FMPK logical address stored in the logical address 701 of the entry.

Next, the processing operation of the storage system 3 according to the present embodiment will be described.

FIG. 12 is a flowchart of input / output processing by the disk controller according to the first embodiment.

The input / output process is performed when the disk controller 10 receives an I / O request via the SAN 4.

When the CPU 11 of the disk controller 10 receives an I / O request via the host I / F 12, the CPU 11 determines whether the received I / O request is a write request (write request) (step S11).

As a result, if the received I / O request is not a write request (step S11: No), that is, if it is a read request, the CPU 11 executes read processing according to the I / O request (step S12), End input / output processing.

On the other hand, if the received I / O request is a write request (step S11: Yes), the CPU 11 determines whether the storage system 3 has an FMPK 20 having a different write life (step S13).

As a result, when there is no FMPK 20 having a different write life (step S 13: No), it means that the write life does not greatly change even if data is written to any FMPK 20. For example, the I / F of the write request When the storage area of the FMPK 20 for storing the write data is not allocated to the address of the volume corresponding to the write request and the storage area needs to be newly allocated, the CPU 11 allocates the storage area. The FMPK 20 number (FMPK number) is determined, the RAID group logical address to be assigned corresponding to the storage area of this FMPK 20 is determined, and the entry of the address of the volume corresponding to the write request of the write destination management table 400 is decided. FMPK number of MPK20 is a RAID group number belonging stores the RAID group logical address (step S14). Whether or not the storage area of the FMPK 20 for storing write data is allocated to the address of the volume corresponding to the write request is determined by the RAID group logical address 403 of the entry of the write destination management table 400 corresponding to the address of the volume. It can be determined by reference.

Further, for example, when the I / F of the write request is an ID, the CPU 11 registers the object data corresponding to the write request in the object data management table 500, and assigns the storage area to the FMPK 20. The RAID group logical address to be assigned corresponding to the storage area of is determined, the entry corresponding to this object data is added to the object data write destination management table 600, and the RAID group number to which the determined FMPK 20 belongs and the RAID The group logical address is stored (step S14).

Next, the CPU 11 specifies the FMPK logical address corresponding to the allocated RAID group logical address based on the drive management table 300, and writes the write data specifying the specified FMPK logical address to the determined FMPK 20. A request is issued (step S15), and the input / output processing ends. In the FMPK 20 that has received the write request from the CPU 11, the write data is to be written.

On the other hand, when there is an FMPK 20 having a different write life (step S13: Yes), the write life of the entire storage system 3 changes greatly depending on which FMPK 20 the data is written to. The write request is referred to, and it is confirmed what the I / F of the write request is (step S16). Here, the I / F of the write request can be confirmed by referring only to the information of the application layer header of the Open Systems Interconnection (OSI) reference model in the write request, and the write data itself It is not necessary to refer to the contents.

As a result, when the I / F of the write request is confirmed to be the interface of the object ID (step S16: ID), the write data corresponding to the write request may be data with low update possibility. Since this means high, the CPU 11 determines the number of the FMPK 20 storing the write data from the FMPK 20 having a short write life, and further assigns the RAID group logical address corresponding to the storage area of this FMPK 20. The object data entry corresponding to the write request is registered in the object data management table 500, an entry corresponding to the object data is added to the object data write destination management table 600, and the RAI of the determined FMPK 20 is added to this entry. Storing a group number, the RAID group the logical address, the (step S17). Here, the FMPK 20 having a short write life may be specified by referring to the write life 201 of the RAID management group management table 200 or may be specified based on the type of flash memory type 205.

Next, the CPU 11 issues a write request for write data to the determined FMPK 20 (step S18), and ends the input / output processing. In the FMPK 20 that has received a write request from the CPU 11, write data is written by a data write process described later. For this reason, write data having a low update possibility is appropriately stored in the FMPK 20 having a short write life.

On the other hand, when the I / F of the write request is confirmed to be an interface of SCSI (step S16: SCSI), the write data corresponding to the write request is highly likely to be data with high update possibility. In the case where the storage area for storing the write data is not allocated to the address of the volume corresponding to the write request, the CPU 11 stores the write data from among the FMPK 20 having a long write life. And the assigned RAID group logical address corresponding to the storage area of this FMPK 20, and the determined RAID group number of FMPK 20 in the entry of the address of the volume corresponding to the write request of the write destination management table 400. And RAID Storing a loop logical address (step S19). Here, the FMPK 20 having a long write life may be specified by referring to the write life 201 of the RAID management group management table 200, or may be specified based on the type of flash memory type 205.

Next, the CPU 11 specifies the FMPK logical address corresponding to the allocated RAID group logical address based on the drive management table 300, and writes the write data specifying the specified FMPK logical address to the determined FMPK 20. A request is issued (step S20), and the input / output processing ends. In the FMPK 20 that has received the write request from the CPU 11, the data write process described later is performed and the write data is written. For this reason, write data with high update possibility is appropriately stored in the FMPK 20 having a long write life.

When the FMPK 20 storing write data is already assigned to the address of the volume corresponding to the write request, the CPU 11 issues a write data write request to the FMPK 20 already assigned. It becomes.

Next, data write processing in the FMPK 20 will be described.

FIG. 13 is a flowchart of data write processing by the FMPK controller according to the first embodiment.

The data write process is executed when the CPU 31 of the FMPK controller 30 receives a write request from the disk controller 10.

When receiving the write request from the disk controller 10, the CPU 31 of the FMPK controller 30 calculates a logical page from the FMPK logical address assigned to the write target data in the write request (step S21). The logical page can be calculated based on the logical address and the size of the logical page set in advance.

Next, the CPU 31 refers to the logical / physical mapping table 700 to specify a physical page corresponding to the calculated logical page (step S22). Next, the CPU 31 stores write data corresponding to the write request for the identified physical page (step S23), and ends the data write process.

By this processing, the write data is properly stored in the FMPK 20 requested to be written by the disk controller 10.

According to the embodiment described above, based on the interface of the protocol of the write request:
Since the data characteristic of data with high update possibility or low data is judged and the FMPK 20 to be written is determined according to the characteristic, the data is appropriately stored according to the characteristics of each FMPK 20 Thus, the life of the storage system 3 can be extended, and the bit cost can be reduced.

Next, a storage system 3 according to the second embodiment will be described. The storage system 3 according to the second embodiment will be described in terms of differences from the storage system according to the first embodiment.

In the storage system 3 according to the second embodiment, the storage system 3 according to the first embodiment determines the FMPK 20 of the write data write destination according to the write life at the initial point, that is, the initial performance of the FMPK 20. The FMPK 20 to which the write data is to be written is determined based on the actual lifetime of each FMPK 20.

The storage system 3 according to the second embodiment includes a RAID group management table 210 (see FIG. 14) instead of the RAID group management table 200 shown in FIG.

FIG. 14 is a configuration diagram of a RAID group management table according to the second embodiment.

The RAID group management table 210 further includes a field of status (lifetime) 208 in the entry. The status (lifetime) 208 stores the lifetime of the FMPK 20 corresponding to the entry (available days: equivalent to remaining life information). The usable days of the state (lifetime) 208 are successively updated, for example, when the CPU 11 detects that writing to the FMPK 20 has occurred.

The CPU 11 according to the second embodiment refers to the value of the state (lifetime) 208 in step S17 shown in FIG. 12, and determines the number of the FMPK 20 for storing the write data from the FMPK 20 having a small value. Further, in step S19 in FIG. 12, the CPU 11 refers to the value of the state (lifetime) 208 and determines the number of the FMPK 20 for storing the write data out of the FMPK 20 having many values.

As described above, in the storage system 3 according to the second embodiment, the FMPK 20 as the storage destination of the write data is determined based on the lifetime updated according to the actual writing, so the storage system It is possible to determine the storage destination of data appropriately according to the actual state of 3.

The present invention is not limited to the embodiments described above, and can be appropriately modified and implemented without departing from the spirit of the present invention.

For example, in the first embodiment, the remaining life information is managed as shown in the RAID group management table 210, and when the protocol interface corresponding to the write request is the object ID interface, the write request is supported. When allocating a storage area for writing data, if there is not enough free space in the storage area of the FMPK 20 corresponding to the first nonvolatile memory device, the remaining life information in the FMPK 20 corresponding to the second nonvolatile memory device indicates When allocating the FMPK 20 with a low remaining life, and allocating the storage area for writing the data corresponding to the write request when the protocol interface is SCSI, it is sufficient for the storage area of the FMPK 20 corresponding to the second nonvolatile memory device Empty Without amount, it may be assigned FMPK20 remaining life often show remaining life information of the FMPK20 corresponding to the first non-volatile memory device.

In the above embodiment, the QLC FM chip 40 is used as the first nonvolatile memory, and the MLC and TLC FM chips 40 are used as the second nonvolatile memory. For example, the QLC and TLC FM chips are used. The first nonvolatile memory 40 may be used as the first nonvolatile memory 40, and the MLC and SLC FM chips 40 may be used as the second nonvolatile memory. In short, the FM chip 40 may be distinguished according to the initial writing life.

Further, in the above embodiment, a part or all of the processing performed by the CPU 11 or the CPU 31 may be performed by a hardware circuit. Also, the program in the above embodiment may be installed from a program source. The program source may be a program distribution server or storage medium (eg, portable storage medium).

1 ... computer system, 2 ... host, 3 ... storage system, 10 ... disk controller, 20 ... FMPK, 40 ... FM chip

Claims

A plurality of non-volatile memory devices, each having a non-volatile memory;
A processor unit, which is one or more processors for managing the plurality of nonvolatile memory devices;
The plurality of nonvolatile memory devices include a first nonvolatile memory device having a first nonvolatile memory having a short initial writing life and a second nonvolatile memory having a second non-volatile memory having a long initial writing life. Memory device and included
The processor unit is
The interface of the protocol to which the write request for the predetermined write data received from the upper level device is compatible is the first protocol interface which is highly likely to be used when storing data with low update possibility, or updated Determine if it is a second protocol interface that is likely to be used when storing data that is likely
When the storage area to which the write data corresponding to the write request is to be written is allocated when the interface of the protocol corresponding to the write request is the first protocol interface, the storage of the first non-volatile memory device An area is allocated, and when the interface of the protocol to which the write request corresponds is the second protocol interface, when the storage area to which the write data corresponding to the write request is allocated is allocated, Allocate storage area of non-volatile memory device,
A storage system that writes the write data to the allocated storage area.
The first protocol interface is an interface of a protocol used to store data in object units, and the second protocol interface is an interface of a protocol used to store data in predetermined blocks. A storage system according to item 1.
The first non-volatile memory is a non-volatile memory capable of writing a 4-bit value to one storage element, and the second non-volatile memory is a non-volatile capable of writing a 3-bit value or less to one storage element The storage system according to claim 1, which is a memory.
The processor unit is
The interface of the protocol to which the write request corresponds is the first protocol interface or the second protocol interface based on the information of the application layer header of the OSI (Open Systems Interconnection) reference model in the write request The storage system according to claim 1, which determines whether
The processor unit is
It is determined whether the interface of the protocol to which the write request corresponds is the first protocol interface or the second protocol interface without referring to the content of the write data corresponding to the write request. The storage system according to claim 4.
The processor unit is
Managing remaining life information on remaining write life in the non-volatile memory device based on a write status of data to the non-volatile memory device;
When the interface of the protocol to which the write request corresponds corresponds to the first protocol interface, when allocating a storage area to which the write data corresponding to the write request is to be allocated, the remaining lifetime information is used to The storage area of a non-volatile memory device with a small remaining of the write life is allocated, and the write data corresponding to the write request when the interface of the protocol to which the write request corresponds is the second protocol interface The storage system according to claim 1, wherein when allocating a storage area for writing, the storage area of a non-volatile memory device having a large remaining of the write life is allocated based on the remaining life information.
The processor unit is
Managing remaining life information on remaining write life in the non-volatile memory device based on a write status of data to the non-volatile memory device;
When the interface of the protocol to which the write request corresponds corresponds to the first protocol interface, when allocating a storage area for writing data corresponding to the write request, the storage area of the first non-volatile memory device is allocated If there is not enough free space, a non-volatile memory device with a small remaining life indicated by the remaining life information in the second non-volatile memory device is allocated, and in the case of the second protocol interface, the write request is made. When allocating a storage area for writing corresponding data, if there is not enough free space in the storage area of the second non-volatile memory device, the remaining life indicated by the remaining life information in the first non-volatile memory device is The method according to claim 1, wherein a large number of non-volatile memory devices are allocated. Storage system.
The first protocol interface is an interface of HTTP (Hypertext Transfer Protocol) used to store data in object units, and the second protocol interface is a SCSI used to store data in predetermined blocks. The storage system according to claim 2, which is a Small Computer System Interface.
Storage of write data corresponding to a write request by a storage system having a plurality of non-volatile memory devices each having a non-volatile memory, and a processor unit which is one or more processors managing the plurality of non-volatile memory devices A data storage control method for allocating and storing an area, comprising:
The plurality of nonvolatile memory devices include a first nonvolatile memory device having a first nonvolatile memory having a short initial writing life and a second nonvolatile memory having a second non-volatile memory having a long initial writing life. Memory device and included
The interface of the protocol to which the write request for the predetermined write data received from the upper level device is compatible is the first protocol interface which is highly likely to be used when storing data with low update possibility, or updated Determine if it is a second protocol interface that is likely to be used when storing data that is likely
A storage area of the first non-volatile memory when allocating a storage area for writing the write data corresponding to the write request, when the interface of the protocol corresponding to the write request is the first protocol interface Is assigned, and when the interface of the protocol to which the write request corresponds is the second protocol interface, the second non-volatile memory is assigned when a storage area for writing the write data corresponding to the write request is assigned. Allocate memory area of memory
A data storage control method for writing the write data in the allocated storage area.
The first protocol interface is an interface of a protocol used to store data in object units, and the second protocol interface is an interface of a protocol used to store data in predetermined blocks. The data storage control method according to Item 9.