TWI891531B - Storage device management method and data storage system - Google Patents

Abstract

A storage device management method and a data storage system are disclosed. The method includes: detecting an access event; determining a target management strategy from a plurality of candidate management strategies according to a type of the access event; if the target management strategy is a first management strategy, performing a parallel access operation on at least two physical storage devices among the physical storage devices according to the access event; and if the target management strategy is a second management strategy, performing a single access operation on a single physical storage device among the physical storage devices according to the access event.

Description

Storage device management method and data storage system

The present invention relates to a storage device management method and a data storage system.

To increase storage device capacity and/or access speed, multiple storage devices can be combined into a Redundant Array of Independent Disks (RAID group). Multiple storage devices in the same RAID group can be accessed in parallel (e.g., data can be written or read simultaneously). However, in reality, after using multiple storage devices in a RAID group for a period of time, different storage devices in the same RAID group may experience uneven usage. For example, some storage devices in the same independent redundancy array group may be accessed too frequently, which can severely reduce the service life of these storage devices (or even the entire independent redundancy array group). However, if independent redundancy array groups are not used to manage these storage devices, data access performance will also be reduced.

In view of this, the present invention provides a storage device management method and a data storage system to improve the above problems.

An embodiment of the present invention provides a storage device management method for a host system, wherein the host system is connected to multiple physical storage devices, and the storage device management method includes: detecting an access event; determining a target management policy from multiple candidate management policies based on the type of the access event; if the target management policy is the first management policy among the multiple candidate management policies, performing parallel access operations on at least two of the multiple physical storage devices based on the access event; and if the target management policy is the second management policy among the multiple candidate management policies, performing a separate access operation on a single physical storage device among the multiple physical storage devices based on the access event.

An embodiment of the present invention further provides a data storage system comprising a host system and a plurality of physical storage devices. The plurality of physical storage devices are connected to the host system. The host system is configured to: detect an access event; determine a target management policy from a plurality of candidate management policies based on the type of the access event; if the target management policy is the first management policy among the plurality of candidate management policies, perform parallel access operations on at least two of the plurality of physical storage devices based on the access event; and if the target management policy is the second management policy among the plurality of candidate management policies, perform a separate access operation on a single physical storage device among the plurality of physical storage devices based on the access event.

Based on the above, after an access event is detected, a target management policy can be determined from a plurality of candidate management policies based on the type of the access event. Then, if the target management policy is the first management policy among the plurality of candidate management policies, then based on the access event, a parallel access operation can be performed on at least two of the plurality of physical storage devices to improve data access performance. However, if the target management policy is the second management policy among the plurality of candidate management policies, then based on the access event, a single access operation can be performed on a single physical storage device among the plurality of physical storage devices to adjust for differences in usage levels among the physical storage devices. In this way, the service life of the physical storage device can be effectively maintained or even extended while maximizing data access performance.

FIG1 is a schematic diagram of a data storage system according to an embodiment of the present invention. Referring to FIG1 , the data storage system 10 includes a host system 11 and physical storage devices 12(1) to 12(n). The host system 11 can store data in at least one of the physical storage devices 12(1) to 12(n), or read data from at least one of the physical storage devices 12(1) to 12(n). For example, the host system 11 is any system that can actually cooperate with the physical storage devices 12(1)-12(n) to store data, such as a smart phone, a tablet computer, a laptop computer, a desktop computer, an industrial computer, a server, a game console, or a computer system installed in a specific carrier (such as a vehicle, an airplane, or a ship). In addition, the physical storage devices 12(1)-12(n) can include various non-volatile physical storage devices such as a flash drive, a memory card, a solid state drive (SSD), a secure digital (SD) card, a compact flash (CF) card, or an embedded storage device. In one embodiment, the physical storage devices 12(1)-12(n) can be pluggably connected to the host system 11. In addition, the total number of physical storage devices 12(1)-12(n) can be any number greater than 1, and the present invention is not limited thereto.

The host system 11 includes a processor 111 and a connection interface 112. The processor 111 can be used to control the operation of the host system 11 in whole or in part. For example, the processor 111 may include a central processing unit (CPU), a graphics processing unit (GPU), a neural network processing unit (NPU), or other programmable general-purpose or special-purpose microprocessors, digital signal processors (DSPs), programmable controllers, application-specific integrated circuits (ASICs), programmable logic devices (PLDs), or other similar devices or combinations of these devices. In addition, the number of processors 111 can be one or more, and the present invention is not limited thereto. It should be noted that in the following embodiments, the operating description of the processor 111 may be equivalent to the operating description of the host system 11.

The connection interface 112 is connected to the processor 111. The connection interface 112 can be used to transmit signals (including data or instructions) to at least one of the physical storage devices 12(1)-12(n) or receive signals from at least one of the physical storage devices 12(1)-12(n). For example, the connection interface 112 can comply with one or more connection interface standards such as Serial Advanced Technology Attachment (SATA), Parallel Advanced Technology Attachment (PATA), Peripheral Component Interconnect Express (PCI Express), and/or Universal Serial Bus (USB). It should be noted that the number of the connection interface 112 can be one or more, and the present invention is not limited thereto.

In one embodiment, the host system 11 further includes any practically required hardware devices, such as a power management circuit, a network interface card, a keyboard (or touchpad), a screen, a microphone and/or a speaker, etc. The present invention does not limit the number and type of additional hardware devices that the host system 11 may include.

The physical storage devices 12(1)-12(n) each include one or more memory modules. Each memory module is used to store data. For example, each memory module may include one or more memory cell arrays. The memory cells in the memory cell array may be used to store data in a non-volatile manner. For example, each memory module may include a single-level cell (SLC) NAND flash memory module (i.e., a flash memory module in which one memory cell can store one bit), a multi-level cell (MLC) NAND flash memory module (i.e., a flash memory module in which one memory cell can store two bits), a triple-level cell (TLC) NAND flash memory module (i.e., a flash memory module in which one memory cell can store three bits), a quad-level cell (MLC) NAND flash memory module (i.e., a flash memory module in which one memory cell can store three bits), or a quad-level cell (MLC) NAND flash memory module. A QLC (Quadrature Least Connection) NAND-type flash memory module (i.e., a flash memory module in which one memory cell can store 4 bits), other types of flash memory modules, or non-volatile memory modules having the same or similar characteristics.

In one embodiment, the physical storage devices 12(1) to 12(n) each include one or more memory controllers and one or more connection interfaces. For example, the memory controller in the physical storage device 12(i) is used to control and/or manage the memory modules in the physical storage device 12(i). The connection interface in the physical storage device 12(i) is used to communicate (e.g., exchange data and/or signals) with the host system 11 (e.g., connection interface 112). In addition, the memory controller in the physical storage device 12(i) can access the memory modules in the physical storage device 12(i) based on instructions (e.g., write instructions, read instructions, or other instructions) from the host system 11.

In one embodiment, the processor 111 may detect a pending access event. For example, the access event is used to update, read, or delete data belonging to at least one logical unit. For example, the logical unit may include a logical block address (LBA) or other logical management unit.

In one embodiment, the access event may be used to update data belonging to a certain logic unit (also referred to as the first logic unit) from certain data (also referred to as the first data) to other data (also referred to as the second data). In one embodiment, the access event may be used to read data belonging to a certain logic unit (also referred to as the second logic unit). In one embodiment, the access event may be used to delete data belonging to a certain logic unit (also referred to as the third logic unit).

In one embodiment, the processor 111 may detect the type of the access event. Then, the processor 111 may determine (e.g., select) a management policy (also referred to as a target management policy) from a plurality of management policies (also referred to as candidate management policies) based on the type of the access event. Each candidate management policy includes management rules and/or access rules corresponding to at least one of the physical storage devices 12(1)-12(n). The processor 111 may manage and/or access at least one of the physical storage devices 12(1)-12(n) based on the determined (e.g., selected) management policy (i.e., the target management policy).

In one embodiment, based on the type of the access event, the processor 111 may determine a management policy (also referred to as a first management policy) among the multiple candidate management policies as a target management policy. In one embodiment, if the target management policy is the first management policy, then based on the access event, the processor 111 may perform a parallel access operation on at least two of the physical storage devices 12(1) to 12(n). For example, in the parallel access operation, the processor 111 may, based on the target management policy (i.e., the first management policy), instruct at least two of the physical storage devices 12(1) to 12(n) to perform a parallel data write operation to store the data updated (e.g., stored) as indicated by the access event.

In one embodiment, based on the type of the access event, the processor 111 may determine another management policy (also referred to as a second management policy) from the plurality of candidate management policies as a target management policy. The first management policy is different from the second management policy. In one embodiment, if the target management policy is the second management policy, then based on the access event, the processor 111 may perform a single access operation (also referred to as a non-parallel access operation) on a single physical storage device among the physical storage devices 12(1)-12(n). For example, in the single access operation, the processor 111 may, based on the target management policy (i.e., the second management policy), instruct a single physical storage device among the physical storage devices 12(1)~12(n) to perform a single (or independent) data write operation to store the data updated (e.g., stored) indicated by the access event.

In one embodiment, the type of the access event may reflect whether the access event is a sequential write event or a random write event. The sequential write event is used to update data in a plurality of consecutive logic units. In contrast, the random write event is used to update data in at least one non-consecutive logic unit.

In one embodiment, the amount of data updated indicated by a sequential write event (also referred to as the first data amount) is necessarily greater than the amount of data updated indicated by a random write event (also referred to as the second data amount). For example, the first data amount may be 16KB, and the second data amount may be 4KB. Furthermore, both the first data amount and the second data amount can be adjusted based on practical needs and are not limited by the present invention.

In one embodiment, processor 111 may compare the amount of data updated indicated by the access event with a threshold value. Processor 111 may then determine the type of access event based on the comparison result. For example, if the comparison result indicates that the amount of data updated indicated by the access event is greater than or equal to the threshold value, processor 111 may determine that the access event is a sequential write event. However, if the comparison result indicates that the amount of data updated indicated by the access event is less than the threshold value, processor 111 may determine that the access event is a random write event.

In one embodiment, if the access event is a sequential write event, the processor 111 may determine the first management policy as the target management policy. However, if the access event is a random write event, the processor 111 may determine the second management policy as the target management policy. In one embodiment, the processor 111 may also determine another management policy from the plurality of candidate management policies as the target management policy based on whether the type of the access event is a specific type, and manage or access at least one of the physical storage devices 12(1)-12(n) based on the target management policy, but the present invention is not limited thereto.

In one embodiment, if the target management policy is the first management policy, the processor 111 may bind the at least two physical storage devices into the same independent RAID group. The processor 111 may then perform parallel access operations on the at least two physical storage devices belonging to the independent RAID group based on the access event.

In one embodiment, if the target management policy is the second management policy, the processor 111 may release the binding previously applied to at least two of the physical storage devices 12(1)-12(n). Thereafter, the processor 111 may perform a separate access operation on the released single physical storage device according to the access event.

In one embodiment, the single physical storage device is the physical storage device with the lowest or relatively low usage among the physical storage devices 12(1)-12(n). In one embodiment, the single physical storage device is the physical storage device with a usage level lower than a preset usage level among the physical storage devices 12(1)-12(n).

In one embodiment, the processor 111 may obtain usage information from the physical storage devices 12(1) to 12(n). The usage information may reflect the usage levels of the physical storage devices 12(1) to 12(n). For example, the usage information obtained from the physical storage device 12(i) may reflect the usage level of the physical storage device 12(i). Then, the processor 111 may set one of the physical storage devices 12(1) to 12(n) as the single physical storage device based on the usage information. Thereafter, the processor 111 may perform the single access operation on the single physical storage device based on the access event.

In one embodiment, the usage information includes count information. The count information may reflect the number of times a target operation has been performed on each of the physical storage devices 12(1)-12(n). For example, the usage information obtained from the physical storage device 12(i) may reflect the number of times a target operation has been performed on the physical storage device 12(i). For example, the target operation includes at least one of a read operation, a write operation, an erase operation, a decode operation, and a reread operation.

Taking physical storage device 12(i) as an example, a read operation on physical storage device 12(i) is used to read data from physical storage device 12(i). A write operation on physical storage device 12(i) is used to store data in physical storage device 12(i). An erase operation on physical storage device 12(i) is used to erase data in physical storage device 12(i). A decode operation on physical storage device 12(i) is used to decode data read from physical storage device 12(i). The reread operation for the physical storage device 12(i) is used to reread data from the physical storage device 12(i) using an adjusted read voltage when a decoding failure occurs, in an attempt to reduce error bits in the read data.

In one embodiment, the processor 111 may evaluate the usage level of each of the physical storage devices 12(1)-12(n) based on the usage level information obtained from the physical storage devices 12(1)-12(n). For example, if the count information obtained from the physical storage device 12(j) is higher than the count information obtained from the physical storage device 12(k), the processor 111 may determine that the usage level of the physical storage device 12(j) is higher than the usage level of the physical storage device 12(k).

In one embodiment, the processor 111 may compare the usage information obtained from the physical storage devices 12(1)-12(n). Then, based on the comparison result, the processor 111 may set the physical storage device with the lowest or relatively low usage among the physical storage devices 12(1)-12(n) as the single physical storage device.

In one embodiment, the processor 111 may set a selected single physical storage device (i.e., the least used or relatively less used physical storage device among the physical storage devices 12(1)~12(n)) as a primary storage device, and set the remaining physical storage devices among the physical storage devices 12(1)~12(n) as secondary storage devices.

In one embodiment, the processor 111 may perform the synchronous access operation on the primary storage device and the secondary storage device in response to a sequential write event to improve data access performance. In addition, in one embodiment, the processor 111 may perform the individual access operation on the primary storage device in response to a random write event to adjust for differences in usage among the physical storage devices 12(1)-12(n) by increasing usage of the primary storage device (i.e., the physical storage device with the lowest or relatively low usage).

FIG2 is a schematic diagram illustrating accessing at least two physical storage devices based on a first management policy according to an embodiment of the present invention. Referring to FIG2 , in one embodiment, it is assumed that the detected access event is a sequential write event 201. Based on the sequential write event 201, the processor 111 may select a first management policy from a plurality of candidate management policies as a target management policy. Then, the processor 111 may bind the physical storage devices 21(1) to 21(m) into a redundant array of independent hard disks (RAID) group 21 based on the target management policy (i.e., the first management policy). For example, the physical storage devices 21(1) to 21(m) may be at least two of the physical storage devices 12(1) to 12(n) of FIG1 .

Then, the processor 111 can perform a parallel access operation on the bound physical storage devices 21(1)-21(m) according to the sequential write event 201. For example, in this parallel access operation, the processor 111 can distribute the data updated (e.g., stored) indicated by the sequential write event 201 to the physical storage devices 21(1)-21(m) through sequential write. In this way, the data write performance of the data storage system 10 can be effectively improved.

FIG3 is a schematic diagram illustrating accessing a single physical storage device based on a second management policy according to an embodiment of the present invention. Referring to FIG3 , in one embodiment, it is assumed that the detected access event is a random write event 301. Based on the random write event 301, the processor 111 may select the second management policy from a plurality of candidate management policies as the target management policy. The processor 111 may then release the bindings previously assigned to some physical storage devices based on the target management policy (i.e., the second management policy). The processor 111 may then select the least utilized or relatively least utilized physical storage device from the released physical storage devices as the physical storage device 31. For example, the processor 111 may unbind the physical storage devices 21(1)-21(m) that were originally bound to the independent hard disk redundant array group 21 in FIG2 . Then, the processor 111 may select the least used or relatively less used physical storage device 31 from the physical storage devices 21(1)-21(m).

Then, the processor 111 may perform a separate access operation on the physical storage device 31 according to the random write event 301. For example, in this separate access operation, the processor 111 may store the data updated (e.g., stored) indicated by the random write event 301 separately in the physical storage device 31 through random write. In one embodiment, by performing a random write on the least used or relatively low used physical storage device 31 when the random write event 301 is detected, a beneficial technical effect of performing wear leveling on the physical storage devices 12(1)-12(n) may be achieved. In this way, the overall service life of the physical storage devices 12(1)~12(n) can be effectively maintained or even extended without affecting the data access performance.

FIG4 is a flow chart of a storage device management method according to an embodiment of the present invention. Referring to FIG4 , in step S401, an access event is detected. In step S402, a target management policy is determined from a plurality of candidate management policies based on the type of the access event. If the target management policy is the first management policy among the plurality of candidate management policies, in step S403, based on the access event, parallel access operations are performed on at least two of the plurality of physical storage devices. In addition, if the target management policy is the second management policy among the plurality of candidate management policies, in step S404, based on the access event, a separate access operation is performed on a single physical storage device among the plurality of physical storage devices.

FIG5 is a flow chart of a storage device management method according to an embodiment of the present invention. Referring to FIG5 , in step S501, an access event is detected. In step S502, it is determined whether the access event is a sequential write event. If the access event is a sequential write event, in step S503, a first management strategy among multiple candidate management strategies is determined as a target management strategy. In step S504, at least two of the multiple physical storage devices are bound to the same independent hard disk redundant array group. In step S505, based on the access event, a parallel access operation is performed on the at least two physical storage devices.

On the other hand, if the access event is not a sequential write event, then in step S506, a second management policy from the multiple candidate management policies is determined as the target management policy. In step S507, at least two of the multiple physical storage devices are unbound (e.g., unbound from the same independent hard drive redundant array group). In step S508, a single access operation is performed on the unbound single physical storage device based on the access event.

However, the steps in Figures 4 and 5 have been described in detail above and will not be repeated here. It is worth noting that the steps in Figures 4 and 5 can be implemented as multiple code blocks or circuits, and the present invention is not limited thereto. Furthermore, the methods in Figures 4 and 5 can be used in conjunction with the above exemplary embodiments or independently, and the present invention is not limited thereto.

In summary, the storage device management method and data storage system proposed in the embodiments of the present invention can determine how to manage and access at least one of multiple physical storage devices based on the type of access event to be executed (e.g., a sequential write event or a random write event). For example, for sequential write events, parallel access operations can be performed on at least two of the multiple physical storage devices to improve data access performance. Furthermore, for random write events, individual access operations can be performed on a single physical storage device among the multiple physical storage devices to accommodate differences in usage among the physical storage devices. This can effectively maintain or even extend the service life of multiple physical storage device regimes while maximizing data access performance.

Although the present invention has been disclosed above by way of embodiments, they are not intended to limit the present invention. Any person having ordinary skill in the art may make slight modifications and improvements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the scope of the attached patent application.

10: Data storage system 11: Host system 111: Processor 112: Connection interface 12(1)~12(n), 21(1)~21(m), 31: Physical storage device 201: Sequential write event 21: Independent hard drive redundant array group 301: Random write event S401~S404, S501~S508: Steps

Figure 1 is a schematic diagram of a data storage system according to an embodiment of the present invention. Figure 2 is a schematic diagram of accessing at least two physical storage devices based on a first management policy according to an embodiment of the present invention. Figure 3 is a schematic diagram of accessing a single physical storage device based on a second management policy according to an embodiment of the present invention. Figure 4 is a flow chart of a storage device management method according to an embodiment of the present invention. Figure 5 is a flow chart of a storage device management method according to an embodiment of the present invention.

S401~S404: Steps

Claims (14)

A storage device management method for a host system, wherein the host system is connected to multiple physical storage devices, includes: detecting an access event; determining a target management policy from multiple candidate management policies based on the type of the access event; if the target management policy is a first management policy among the multiple candidate management policies, performing parallel access operations on at least two of the multiple physical storage devices based on the access event; and if the target management policy is a second management policy among the multiple candidate management policies, performing a single access operation on a single physical storage device among the multiple physical storage devices based on the access event. The storage device management method of claim 1, wherein the step of determining the target management policy from the multiple candidate management policies based on the type of the access event comprises: If the access event is a sequential write event, determining the first management policy from the multiple candidate management policies as the target management policy; and If the access event is a random write event, determining the second management policy from the multiple candidate management policies as the target management policy. The storage device management method as described in claim 1, wherein the single physical storage device is the least utilized or relatively less utilized physical storage device among the multiple physical storage devices. The storage device management method of claim 1 further comprises: obtaining usage information from the plurality of physical storage devices, wherein the usage information reflects the usage levels of the plurality of physical storage devices; and setting one of the plurality of physical storage devices as the single physical storage device based on the usage information. The storage device management method as described in claim 4, wherein the usage information includes count information, and the count information reflects the number of times the target operation is executed on each of the multiple physical storage devices. The storage device management method as described in claim 5, wherein the target operation includes at least one of a read operation, a write operation, an erase operation, a decode operation, and a reread operation. The storage device management method of claim 1 further comprises: If the target management policy is the first management policy, binding the at least two physical storage devices into an independent hard disk redundant array group; and If the target management policy is the second management policy, unbinding the at least two physical storage devices. A data storage system comprises: a host system; and a plurality of physical storage devices connected to the host system, wherein the host system is configured to: detect an access event; determine a target management policy from a plurality of candidate management policies based on the type of the access event; if the target management policy is a first management policy among the plurality of candidate management policies, perform parallel access operations on at least two of the plurality of physical storage devices based on the access event; and if the target management policy is a second management policy among the plurality of candidate management policies, perform a single access operation on a single physical storage device among the plurality of physical storage devices based on the access event. The data storage system of claim 8, wherein the operation of determining the target management policy from the plurality of candidate management policies based on the type of the access event comprises: If the access event is a sequential write event, determining the first management policy from the plurality of candidate management policies as the target management policy; and If the access event is a random write event, determining the second management policy from the plurality of candidate management policies as the target management policy. The data storage system of claim 8, wherein the single physical storage device is the least utilized or relatively less utilized physical storage device among the multiple physical storage devices. The data storage system of claim 8, wherein the host system is further configured to: obtain usage information from the plurality of physical storage devices, wherein the usage information reflects usage levels of the plurality of physical storage devices; and configure one of the plurality of physical storage devices as the single physical storage device based on the usage information. The data storage system of claim 11, wherein the usage information includes count information, and the count information reflects the number of times the target operation is executed on each of the plurality of physical storage devices. The data storage system of claim 12, wherein the target operation includes at least one of a read operation, a write operation, an erase operation, a decode operation, and a reread operation. The data storage system of claim 8, wherein the host system is further configured to: If the target management policy is the first management policy, bind the at least two physical storage devices into an independent hard disk redundant array group; and If the target management policy is the second management policy, unbind the at least two physical storage devices.