CN118642849A - A data processing method, device and equipment based on dual-port solid state hard disk - Google Patents

Abstract

According to the data processing method based on the dual-port solid state disk, the multi-core central processing unit receives the command, if the command is a name space command, the command is forwarded to the data management processor, the data management processor determines the behavior type corresponding to the command, the command is stored in different cache queues according to the behavior type corresponding to the command, the data management processor calculates the load index of each cache queue according to a preset load balancing algorithm, determines the queue priority of each cache queue according to the load index, and the data management processor selects a target cache queue from the cache queues according to the queue priority to process the target command in the target cache queue.

Description

A data processing method, device and equipment based on dual-port solid state hard disk

Technical Field

The present invention relates to the field of computer technology, and in particular to a data processing method, device and equipment based on a dual-port solid state hard disk.

Background Art

Dual-port NVMe SSD (Non-Volatile Memory Express Solid State Drive) is a high-performance, high-reliability solid-state drive technology with dual-channel data transmission. With the increasing demand for storage performance and data protection in data centers, cloud computing, and high-performance computing, dual-port NVMe SSD, as an innovative storage solution, is gradually gaining attention and application.

Currently, servers are mostly used by multiple users. When NVMe SSD runs in a server environment, a physical device needs to be divided into multiple logical namespaces. Each namespace can be regarded as an independent storage space. This division can achieve isolated storage of different data, avoid interference between data, and improve the flexibility and manageability of the system. Different controllers can manage different namespaces, achieving more fine-grained data management and more efficient data processing.

However, NVMe SSD devices are usually multi-core systems. During the multi-core processing, inter-core concurrency in the firmware system will frequently occur, which will greatly affect the working efficiency of the NVMe SSD device.

Summary of the invention

The purpose of the embodiments of the present invention is to provide a data processing method, device and apparatus based on a dual-port solid-state drive, so as to realize setting a data manager in the dual-port solid-state drive to specifically process commands for the namespace, and to ensure the orderly processing of each command cache queue through the load index, thereby releasing part of the computing power of the multi-core central processing unit of the solid-state drive. The specific technical solution is as follows:

In a first aspect of the present invention, a data processing method based on a dual-port solid-state hard disk is provided, wherein the dual-port solid-state hard disk is provided with a multi-core central processing unit and a data management processor, and the method comprises:

receiving a command through the multi-core central processing unit, and if the command is a namespace command, forwarding the command to the data management processor;

Determining, by the data management processor, a behavior type corresponding to the command, and storing the command in different cache queues according to the behavior type corresponding to the command;

Calculating the load index of each cache queue according to a preset load balancing algorithm by the data management processor, and determining the queue priority of each cache queue according to the load index;

The data management processor selects a target cache queue from the cache queue according to the queue priority, and processes the target command in the target cache queue.

Optionally, calculating the load index of each cache queue according to a preset load balancing algorithm by the data management processor, and determining the queue priority of each cache queue according to the load index, includes:

Obtaining queue load information corresponding to each cache queue, wherein the queue load information includes at least: queue weight, command response time, and current load;

Calculating the total queue load of the cache queue according to the current load of each cache queue, and determining the average load of the cache queue according to the total queue load;

Calculating the load index of each cache queue according to a preset load balancing algorithm;

Determining a queue priority of each of the cache queues according to the load index;

The formula corresponding to the preset load balancing algorithm is:

i is the total number of the cache queues, e is a constant, LI_i is the load index of the i-th cache queue, CI_i is the current load of the i-th cache queue, w_i is the queue weight corresponding to the i-th cache queue, w_avg is the average load of the cache queue, RT_i is the command response time corresponding to the i-th cache queue, α and β are adjustment parameters, and α and β are used to balance the load and response time.

Optionally, the namespace command includes: a create command, a delete command, a write command, and a read command, and the receiving of the command by the multi-core central processor and forwarding the command to the data management processor if the command is a namespace command includes:

Receiving a command through the multi-core central processing unit, and if the command is one of the namespace commands, verifying the command to obtain a verification result;

If the verification result is that the verification fails, intercepting the command;

If the verification result is that the verification passes, the command is forwarded to the data management processor.

Optionally, receiving a command through the multi-core central processing unit and forwarding the command to the data management processor if the command is a namespace command includes:

The command is received by a command processing module running in a multi-core central processing unit, and the command is sent to a solid state drive control module running in the multi-core central processing unit;

Applying for a buffer through the solid state drive control module, and transferring the command to the buffer;

If the command is a namespace command, the command is forwarded to the data management processor through the solid state drive control module.

Optionally, processing the target command in the target cache queue includes:

If the target command is a creation command, obtaining creation information carried in the target command;

According to the creation information, the storage space in the namespace resource pool in the dual-port solid-state hard disk is allocated to obtain a target storage space corresponding to the creation information; wherein the namespace information at least includes: storage capacity and a starting logical address;

Sending creation result information corresponding to the target storage space to the solid-state hard disk control module running in the multi-core central processing unit, wherein the creation result information includes a unique identifier of the target space and a creation success identifier.

Optionally, after sending the target storage space information corresponding to the target storage space to the solid state drive control module running in the multi-core central processor, the method further includes:

The solid state drive control module running in the multi-core central processing unit obtains command data matching the creation result information from the buffer, and outputs the target space unique identifier and the creation success identifier for the command data.

Optionally, the method further includes:

Receiving a power-off command through the multi-core central processing unit, and sending the power-off command to the data management processor;

After receiving the power-off command, the data management processor stops receiving new commands and stores the data managed by the data management processor in the flash memory of the dual-port solid state drive.

In a second aspect of the present invention, a data processing device based on a dual-port solid-state hard disk is provided, wherein the dual-port solid-state hard disk is provided with a multi-core central processor and a data management processor, including:

A command acquisition module, configured to receive a command through the multi-core central processing unit, and if the command is a namespace command, forward the command to the data management processor;

A command allocation module, used to determine the behavior type corresponding to the command through the data management processor, and store the command in different cache queues according to the behavior type corresponding to the command;

A load balancing module, configured to calculate a load index of each cache queue according to a preset load balancing algorithm through the data management processor, and determine a queue priority of each cache queue according to the load index;

The command processing module selects a target cache queue from the cache queue according to the queue priority through the data management processor, and processes the target command in the target cache queue.

In the third aspect of the implementation of the present invention, there is also provided an electronic device, comprising a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the data processing method based on the dual-port solid-state hard drive as described in any one of the above items is implemented.

In a fourth aspect of the implementation of the present invention, a computer-readable storage medium is further provided, on which a computer program is stored. When the computer program is executed by a processor, the data processing method based on a dual-port solid-state hard drive as described in any one of the above items is implemented.

The data processing method based on the dual-port solid-state hard disk provided by the embodiment of the present invention receives commands through a multi-core central processing unit. If the command is a namespace command, the command is forwarded to a data management processor, and the behavior type corresponding to the command is determined by the data management processor. The command is stored in different cache queues according to the behavior type corresponding to the command. The load index of each cache queue is calculated by the data management processor according to a preset load balancing algorithm, and the queue priority of each cache queue is determined according to the load index. The target cache queue is selected from the cache queue according to the queue priority by the data management processor, and the target command in the target cache queue is processed. The embodiment of the present invention sets a data manager in the dual-port solid-state hard disk to specifically process commands for the namespace, and ensures the orderly processing of each command cache queue through the load index, thereby freeing up part of the computing power of the multi-core central processing unit of the solid-state hard disk.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below.

FIG1 is a flowchart of a data processing method based on a dual-port solid state drive provided by some embodiments of the present invention;

FIG2 is a flowchart of another data processing method based on a dual-port solid state drive provided by some embodiments of the present invention;

FIG3 is a dual-port logic framework diagram provided by some embodiments of the present invention;

FIG4 is a command interaction summary diagram provided by some embodiments of the present invention;

5 is an overview diagram of processing commands of a data management processor under concurrent conditions provided by some embodiments of the present invention;

6 is a flowchart of a command process for creating a namespace provided by some embodiments of the present invention;

FIG. 7 is a schematic diagram of the structure of a data processing device based on a dual-port solid state drive provided by some embodiments of the present invention.

DETAILED DESCRIPTION

To make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. However, it will be appreciated by those skilled in the art that in the embodiments of the present invention, many technical details are proposed in order to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed in the present application can also be implemented. The division of the following embodiments is for the convenience of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined and referenced with each other without contradiction.

1 , a flowchart of a data processing method based on a dual-port solid-state hard disk provided by some embodiments of the present invention is shown. The dual-port solid-state hard disk is provided with a multi-core central processor and a data management processor. The method may include:

Step 101: Receive a command through the multi-core central processing unit, and if the command is a namespace command, forward the command to the data management processor.

A dual-port SSD is a SSD with two independent data transmission ports, mainly used in enterprise-level storage solutions that require high reliability and redundancy. A dual-port SSD can be connected to two different controllers or hosts. In an SSD, a namespace is a logical storage unit, which is a way for a dual-port SSD to manage data storage and data access. The namespace function is usually a feature of enterprise-level SSDs or specific models of NVMe SSDs. By creating multiple namespaces, a dual-port SSD can provide storage spaces of different sizes for different hosts or controllers, which helps optimize the allocation of storage resources; different namespaces can provide performance isolation to ensure that activities in one namespace do not affect the performance of other namespaces; namespaces can also achieve data isolation, improve data security, and prevent unauthorized access.

In a solid-state drive, a multi-core CPU is the brain of the solid-state drive, responsible for managing the reading, writing, erasing of data and communication with the host system. A multi-core CPU means that multiple processing cores are integrated inside the controller, which can process tasks in parallel, thereby improving the performance and efficiency of the solid-state drive.

Although using namespaces in dual-port SSDs can provide performance isolation and data isolation, it increases the complexity of storage management and requires corresponding management tools and knowledge to effectively manage these logical storage units. It may also affect the overall performance of the dual-port SSD, especially when creating multiple small namespaces. In addition, since dual-port SSDs are mostly equipped with multi-core CPUs, inter-core concurrency in the firmware system will frequently occur during the processing of multiple hosts and multiple cores, greatly affecting the working efficiency of the device.

In some embodiments of the present invention, a multi-core central processing unit and a data management processor are provided in a dual-port solid-state hard disk. A command is received by the multi-core central processing unit. If the command is a namespace command, the command is forwarded to the data management processor, and the namespace command is processed separately by the data management processor. In the embodiment of the present invention, the data management processor is a hardware module unit, which is independent of the multi-core central processing unit. The namespace data is managed by the data management processor, and the multi-core central processing unit can continue to be used in the processing of other modules and processes.

In some embodiments of the present invention, the namespace command includes: a create command, a delete command, a write command, and a read command. Step 101 may also include the following sub-steps:

Sub-step 11: receiving a command through the multi-core central processing unit, verifying the command and obtaining a verification result.

For the received command, the system first performs verification, which includes checking the format of the command, the validity of the parameters, and the authority verification, etc. The purpose of the verification is to ensure the correctness and security of the command. In the embodiment of the present invention, the command is verified by a multi-core central processing unit and the verification result is obtained.

Sub-step 12: If the verification result is that the verification fails, intercept the command.

If the verification result shows that the command fails the verification, the system will intercept the command and prevent it from being executed. The reasons for the failure of the verification may include incorrect command format, missing required parameters, or insufficient user permissions.

Sub-step 13: If the verification result is verification passed and the command is one of the namespace commands, forward the command to the data management processor.

If the verification result shows that the command has passed the verification, the system will continue to process the command. For commands that have passed the verification, if the command is one of the create command, delete command, write command, and read command, the system will forward it to the data management processor. After receiving the command, the data management processor will perform the corresponding operation. For example, the create command will cause a new namespace to be created, the delete command will cause the namespace to be removed, the write command will write data to the specified namespace, and the read command will read data from the namespace.

In some embodiments of the present invention, step 101 may further include the following sub-steps:

Sub-step 21: receiving the command through a command processing module running in the multi-core central processing unit, and sending the command to a solid state drive control module running in the multi-core central processing unit.

The multi-core CPU in the dual-port SSD is equivalent to the brain of the SSD, in which different modules are running to process different types of commands. The command processing module runs on the multi-core CPU and is responsible for receiving commands from users, applications or other system components. The received commands first need to be parsed to determine their type and the operations to be performed. The parsed commands are distributed to the corresponding processing modules according to their type and operation.

In an embodiment of the present invention, a command processing module and a solid-state drive control module are running on a multi-core central processing unit, and commands related to the solid-state drive will be sent to the solid-state drive control module through the command processing module running in the multi-core central processing unit. Specifically, the solid-state drive control module may refer to an NVMe module. The NVMe module refers to a software component that implements the NVMe protocol, which allows the operating system to communicate with the NVMe device. The main functions of the NVMe module include device identification, command processing, etc.; in the present invention, the NVMe module processes commands from the operating system or application program, and converts these commands into the format specified by the NVMe protocol.

Sub-step 22: applying for a buffer through the solid state drive control module, and transferring the command to the buffer.

In a storage system, when an operating system or an application needs to send a command to a solid-state drive device, a buffer is usually requested through a solid-state drive control module, and the command and its related data are transferred to the buffer.

In some embodiments of the present invention, the buffer is applied for and the command is prepared by the following steps: the application or operating system kernel prepares the commands to be sent to the NVMe device, which may include operations such as reading, writing, and erasing; the NVMe module will apply for a buffer after receiving the command, and the system will allocate a piece of memory as a buffer, which is used to temporarily store commands and related data, such as data blocks for read and write operations; the NVMe module transfers the prepared commands and their data to the allocated buffer, and the commands and data can be accessed and processed by the NVMe driver. In some embodiments of the present invention, the NVMe module will apply for a context and a buffer after receiving the command. The context usually includes the status of the command, the execution progress, the error information, etc., which helps the module to track the execution of the command and perform error handling when necessary. After the command is executed, the NVMe module will update the status information in the context and feedback the execution result to the host that initiated the command.

Sub-step 23: If the command is a namespace command, forward the command to the data management processor through the solid state drive control module.

In a storage system, a namespace refers to a set of logical block addresses that are mapped to physical flash memory or other storage media. Namespace commands usually involve the management of these logical blocks, such as creation, deletion, query status, etc. When the NVMe module receives namespace commands, it needs to forward these commands to the data management processor for corresponding processing.

In the related art, after the NVMe module receives the command from the command processing module, the NVMe module also needs to submit the command in the buffer to the I/O queue of the NVMe device, and then the NVMe device takes out the command from the I/O queue and executes it. However, most solid-state hard disk devices are multi-core systems. In the processing of multiple hosts and multi-core systems, inter-core concurrency in the firmware system will frequently occur, which greatly affects the working efficiency of the device. In some embodiments of the present invention, the NVMe module forwards the command to the data management processor, and the data manager processes the command for the namespace separately. Therefore, the module running in the multi-core central processor no longer needs to consider the processing of named data and the synchronization of namespace data between multiple cores.

Step 102: Determine the behavior type corresponding to the command by the data management processor, and store the command into different cache queues according to the behavior type corresponding to the command.

Among them, commands are used to perform specific operations or tasks, and the behavior type of commands refers to the task type corresponding to the command. Cache queue is a technology used in computer systems to improve data access speed. It combines the concepts of cache and queue. The queue follows the first-in-first-out strategy to decide which data should be kept in the cache.

In some embodiments of the present invention, the cache queue is used to manage the relevant commands for the namespace to ensure that they are executed in the correct order and time. After obtaining the command forwarded by the multi-core central processor, the type corresponding to the command is identified and judged, and the behavior type corresponding to the command is determined by the data management processor, and the command is stored in different cache queues according to the behavior type corresponding to the command.

In a specific implementation, the behavior type corresponding to the command may include all command types for the namespace, such as read commands, write commands, create namespace commands, delete namespace commands, etc. After receiving multiple commands, the commands will be stored in the cache queue, and different behaviors will be assigned to different cache queues according to the command parameters. For example, management commands such as create namespace and delete namespace are placed in a cache queue with a higher priority, while read and write commands are placed in a cache queue with a lower priority.

Step 103: The data management processor calculates the load index of each cache queue according to a preset load balancing algorithm, and determines the queue priority of each cache queue according to the load index.

Among them, load balancing algorithm refers to a balancing algorithm used to distribute workload among multiple computing resources to optimize resource usage, maximize throughput, reduce response time, and avoid overloading of any single resource.

After receiving multiple commands, the data management processor will store the commands in different cache queues. For example, management commands such as creating and deleting namespaces are placed in a cache queue with a higher priority, while read and write commands are placed in a cache queue with a lower priority, which can ensure the processing efficiency of high-priority commands. However, if the cache-independent commands with higher priority are always processed, the commands in other priority cache queues cannot be processed.

In some embodiments of the present invention, the load index of each cache queue is calculated by a data management processor according to a preset load balancing algorithm, and the queue priority of each cache queue is determined according to the load index. In a specific implementation of the present invention, the load balancing algorithm can take into account the weight, processing efficiency and load of each cache queue. Therefore, the load balancing algorithm ensures that high-priority commands are processed first, while lower-priority commands are processed in order, so as to avoid timeouts caused by commands in some cache queues not being processed for a long time while other cache queues are idle.

In some embodiments of the present invention, step 103 includes the following sub-steps:

Sub-step 31: Obtain queue load information corresponding to each cache queue, wherein the queue load information at least includes: queue weight, command response time and current load.

Among them, the queue weight is a preset value used to indicate the importance or priority of the queue. In the load balancing algorithm, the weight can be used to determine the workload assigned to each queue; the command response time refers to the time it takes from the command entering the queue to the command execution completion and returning the result. Some advanced queue management systems may provide a query interface to query real-time or historical response time data; the current load refers to the number of commands currently waiting to be processed in the queue, or the number of commands being processed by the queue.

In some embodiments of the present invention, the queue load information corresponding to each cache queue is obtained, and the queue load information at least includes: queue weight, command response time and current load. In a specific implementation, the weight of each queue corresponding to the namespace command can be set in the system configuration. The weights of the queues corresponding to different types of namespace commands can be the same or different, and the present invention does not limit this. Command response time refers to the total time taken for the namespace command to enter the corresponding queue to complete the execution of the namespace command and return the execution result; for example, the total time taken for the command to create a namespace from entering the queue to completing the namespace creation and returning the namespace creation completion result. In some embodiments of the present invention, the current load is the number of commands currently waiting to be processed in each queue corresponding to the namespace command, and it can also be the number of commands being processed in each queue corresponding to the namespace command; for example, when the current load is the number of commands currently waiting to be processed in the queue corresponding to the namespace command, if the number of commands currently waiting to be processed in a cache queue is 10, then the current load of the queue is determined to be 10; or, when the current load is the number of commands being processed in the queue, if the number of namespace commands being processed in a cache queue is 2, then the current load of the queue is determined to be 2.

Sub-step 32: Calculate the total queue load of the cache queue according to the current load of each cache queue, and determine the average load of the cache queue according to the total queue load.

After the current load of each cache queue is determined, the total queue load of the cache queue may be calculated according to the current load of each cache queue, and the average load of the cache queue may be determined according to the total queue load.

In a specific implementation, the current load of each cache may be summed to obtain the total queue load of the cache queue, and then the average load of the cache queue may be determined by an average value algorithm according to the total queue load and the total number of queues.

Sub-step 33: Calculate the load index of each cache queue according to a preset load balancing algorithm.

The preset load balancing algorithm takes into account the weight, processing efficiency and load of each cache queue. In the specific implementation, the processing efficiency refers to the command response time, and the load refers to the current load of each cache queue and the average load of all cache queues.

Sub-step 34: Determine the queue priority of each cache queue according to the load index.

A load index can be calculated for each cache queue through a preset load balancing algorithm, and then the processing priority is assigned based on this index.

The formula corresponding to the preset load balancing algorithm is:

i is the total number of the cache queues, e is a constant, LI_i is the load index of the i-th cache queue, CI_i is the current load of the i-th cache queue, w_i is the queue weight corresponding to the i-th cache queue, w_avg is the average load of the cache queue, RT_i is the command response time corresponding to the i-th cache queue, α and β are adjustment parameters, and α and β are used to balance the load and response time.

Step 104 : Selecting a target cache queue from the cache queues according to the queue priority by the data management processor, and processing the target command in the target cache queue.

The embodiment of the present invention calculates a load index for each cache queue through a load balancing algorithm and allocates processing priority according to the load index. Therefore, after the queue priority is determined, the target cache queue is selected from the cache queue according to the queue priority, and the target command in the target cache queue is processed.

In some embodiments of the present invention, step 104 may further include the following sub-steps:

Sub-step 41: If the target command is a creation command, obtain creation information carried in the target command.

To ensure data independence, the host needs to create a new namespace for the current controller. The data, capacity, and attributes between different namespaces can be different. First, the host sends a create namespace command. The hosts of the two ports can each send this command. Specifically, it can be sent to the NVMe module through the PCIe (PerIP2403000heral Component Interconnect Express, a high-speed serial computer expansion bus standard) port of the dual-port solid-state drive. After receiving the command, the NVMe module recognizes that this command is a create namespace command based on the command parameters. The NVMe module will apply for context and buffer, transfer the command at the same time, and send the command to the data management processor. After receiving the command, the data manager first stores the create namespace command in the cache queue. When executing the command, the data management processor takes out the create namespace command from the cache queue as the target command, and obtains the creation information carried in the target command to prepare for the creation of the namespace.

Sub-step 42: Allocate storage space in the namespace resource pool in the dual-port solid-state hard disk according to the creation information to obtain a target storage space corresponding to the creation information; wherein the namespace information includes at least: storage capacity and a starting logical address.

Storage capacity refers to the size of the storage space that needs to be allocated, and the starting logical address refers to the starting point of the storage space allocation, that is, from which logical address the storage space is allocated.

In some embodiments of the present invention, before allocation, it is confirmed whether there is enough unallocated space in the resource pool to meet the required storage capacity, and it is checked whether the starting logical address in the resource pool matches the requested starting logical address. When it is determined that both the storage capacity and the logical address meet the requirements, the storage space in the namespace resource pool in the dual-port solid-state drive is allocated according to the creation information to obtain the target storage space corresponding to the creation information.

Sub-step 43: Sending the creation result information corresponding to the target storage space to the solid-state hard disk control module running in the multi-core central processor, wherein the creation result information includes a unique identifier of the target space and a creation success identifier.

Regardless of whether the operation of creating the namespace is successful or not, the allocation result needs to be returned, and the creation result information corresponding to the target storage space is sent to the solid-state drive control module running in the multi-core central processing unit.

In some embodiments of the present invention, processing the target command in the target cache queue also includes: judging the legality of the target command, and when the command is legal, the data management processor sends the target command to the subsequent module for processing. When the command is illegal, the data management processor will send the command and the illegal status to the solid-state drive control module, and the solid-state drive control module will perform subsequent processing. Specifically, the data management processor is implemented through hardware and is a component responsible for managing the namespace and storage resources on the solid-state drive device. After receiving the command, the data management processor performs operations on the namespace. After the command is executed, the data management processor sends the target command to the subsequent module for processing, and returns the result to the NVMe module.

In some embodiments of the present invention, quotas and restrictions can be imposed on namespaces. By introducing a new quota system, administrators are allowed to set a storage usage cap for each namespace to prevent a certain namespace from over-consuming storage resources. In other embodiments of the present invention, through a snapshot mechanism or a cloning mechanism, administrators are allowed to quickly create read-only or writable copies of a namespace. These copies can be used for data backup, testing, or development environments without affecting the performance or stability of the original namespace.

In some embodiments of the present invention, after sub-step 43, the following sub-steps are further included:

Sub-step 51: obtaining command data matching the creation result information from the buffer through the solid-state hard disk control module running in the multi-core central processing unit, and outputting the target space unique identifier and the creation success identifier for the command data.

As can be seen from the above sub-step 22, the NVMe module will apply for a buffer when obtaining a command. Specifically, when applying for a buffer, it will also apply for a corresponding context, and transfer the command and related data to the buffer. The context includes the status, execution progress, and error information of the command. Therefore, the data management processor sends the creation result information to the NVMe module. After receiving the message, the NVMe module will find the command according to the context, obtain the specific command data transferred in sub-step 12 from the buffer, and then the NVMe module returns the completed message and the allocated namespace unique identifier to the host.

In some embodiments of the present invention, it also includes:

Receiving a power-off command through the multi-core central processing unit, and sending the power-off command to the data management processor;

After receiving the power-off command, the data management processor stops receiving new commands and stores the data managed by the data management processor in the flash memory of the dual-port solid state drive.

Among them, the power-off instruction is a control signal or command used to instruct an electronic device or system to prepare to turn off the power. In a storage system, the power-off instruction is usually used to safely shut down a solid-state drive or other storage device to ensure data integrity and prevent data loss. The storage device or its controller receives a power-off instruction that may come from an operating system, hardware management interface, or other system component, and the device stops receiving new read and write requests to prevent data inconsistencies during the shutdown process.

In some embodiments of the present invention, when the device encounters a power-off scenario, a power-off command is received through a multi-core central processor. Specifically, the NVMe module will send a power-off message to the data management processor. After receiving the power-off message, the data management processor will no longer receive and process new command messages, and will send the data that it manages and has not been written to the disk to the NFC (NandFlash Controller, Flash Memory Control) module. In addition, the NVMe module will also send the power-off message to the log management module, algorithm module and other modules. These modules will each enter the power-off process after receiving the power-off message, and then send the data to the NFC module. The NFC module will persist the data in the flash memory of the solid-state drive, write all unwritten data to the flash memory, ensure the persistence of the data, and prevent data from being lost in the event of a sudden power outage through the processing steps for the power-off command, thereby ensuring the safe storage of data and the normal shutdown of the system.

The data processing method based on the dual-port solid-state hard disk provided by the embodiment of the present invention receives commands through a multi-core central processing unit. If the command is a namespace command, the command is forwarded to a data management processor, and the behavior type corresponding to the command is determined by the data management processor. The command is stored in different cache queues according to the behavior type corresponding to the command. The load index of each cache queue is calculated by the data management processor according to a preset load balancing algorithm, and the queue priority of each cache queue is determined according to the load index. The target cache queue is selected from the cache queue according to the queue priority by the data management processor, and the target command in the target cache queue is processed. The embodiment of the present invention sets a data manager in the dual-port solid-state hard disk to specifically process commands for the namespace, and ensures the orderly processing of each command cache queue through the load index, thereby freeing up part of the computing power of the multi-core central processing unit of the solid-state hard disk.

2, a second flow chart of the data processing method based on a dual-port solid state drive according to an embodiment of the present invention is shown, which specifically includes:

Step 201: receiving the command through a command processing module running in a multi-core central processing unit, and sending the command to a solid state drive control module running in the multi-core central processing unit.

In some embodiments of the present invention, a dual-port solid-state drive refers to an NVMe SSD device, and a multi-core system is used in the NVMe SSD device, which mainly runs in a server environment. The server supports the creation of multiple controllers and multiple namespaces. The server is a dual-port server and is adapted to the NVMe SSD device. Specifically, when the server is turned on, the NVMe SSD device starts at the same time. Because it runs in a dual-port environment, the NVMe SSD device can support 2 PCIe ports and each is instantiated into a separate NVMe subsystem (NVMe subsystem), and the 2 ports each correspond to an independent host. Each host includes at least one controller and one namespace. As shown in Figure 3, it is a dual-port logical framework diagram provided by some embodiments of the present invention, and the two ports correspond to their respective hosts, and each port corresponds to multiple controllers and multiple namespaces. As shown in Figure 4, it is a command interaction summary diagram provided by some embodiments of the present invention, which receives commands through a command processing module running in a multi-core central processor, and sends commands to a solid-state drive control module running in a multi-core central processor.

Step 202: Apply for a buffer through the solid state drive control module, and transfer the command to the buffer.

In a storage system, when a command (such as read, write, erase, etc.) is received, it is usually necessary to temporarily store these commands in a buffer for further processing. In some embodiments of the present invention, the NVMe module applies to the system memory management unit for one or more buffers, which can be internal fast memories for temporary storage of data and commands. The NVMe module transfers the received command to the applied buffer, and stores the command data, address information, command type, etc. in the specified location of the buffer. If there are multiple commands waiting to be processed in the system, the NVMe module will queue these commands according to a certain priority or order to ensure that the commands are executed in the correct order to avoid data inconsistency or conflict.

Step 203: if the command is a namespace command, forward the command to the data management processor through the solid state drive control module;

In the related art, if the command is for a namespace, it is necessary to process the command in the multi-core CPU, such as creating a namespace, deleting a namespace, reading and writing a namespace, etc. In some embodiments of the present invention, if the command is determined to be a namespace command, the NVMe module will forward the command to the data management processor for processing, and then the multi-core CPU can no longer process the related commands, freeing up part of the computing power of the multi-core CPU.

Step 204: Determine the behavior type corresponding to the command by the data management processor, and store the command in different cache queues according to the behavior type corresponding to the command.

Step 205: The data management processor calculates the load index of each cache queue according to a preset load balancing algorithm, and determines the queue priority of each cache queue according to the load index.

The above 204-205 refer to the contents of steps 102 and 103 discussed in the previous section, and the present invention will not repeat them here.

Step 206: The data management processor selects a target cache queue from the cache queues according to the queue priority, processes the target command in the target cache queue, and if the target command is a create command, obtains the create information carried in the target command.

The data management processor selects the target cache queue from the cache queue and processes the target command in the target cache queue. If the target command is a create command, it is necessary to obtain the creation information carried in the target command. In the present invention, the central processing unit in the dual-port solid-state hard disk is a plurality of cortex cpu cores (processor cores), and the NVMe module runs on a plurality of cortex cpu cores. After processing the command, the NVMe module transfers the command and sends it to the data management processor. The plurality of cortex cpu cores are simultaneously sent to the data management processor. For the data management processor, there will be a high concurrency situation. As shown in FIG5, it is a general diagram of the processing commands of the data management processor under the concurrency provided by the embodiment of the present invention. The data management processor accepts commands and checks commands. If it is a create or delete command, it is stored in a cache queue with a high priority. If it is a read-write command, it is stored in a cache queue with a lower priority, thereby improving the overall performance and response speed, and ensuring the priority processing of key operations at the same time. According to the preset load balancing algorithm, the command is taken out from the cache queue and processed accordingly, and the processing result is returned to the modules running on each central processing unit core.

Step 207: Allocate storage space in the namespace resource pool in the dual-port solid state drive according to the creation information to obtain a target storage space corresponding to the creation information; wherein the namespace information includes at least: storage capacity and a starting logical address.

Specifically, the creation information carried by the command includes parameters such as size (storage capacity) and start lba (starting logical address). Starting from start lba, a storage space of size is allocated to the new namespace, ensuring that the allocated space is continuous and does not conflict with the allocated space in the resource pool. A unique space identifier is allocated to the new namespace, ensuring that the identifier is unique in the system and does not conflict with the existing namespace identifier. The allocation information of the new namespace, including the identifier, the starting logical address, and the allocated capacity, is recorded in the metadata of the resource pool. In addition, the available space information of the resource pool is updated to mark the allocated area.

Step 208: Send the creation result information corresponding to the target storage space to the solid-state hard disk control module running in the multi-core central processor, wherein the creation result information includes a unique identifier of the target space and a creation success identifier.

Step 209: The solid state drive control module running in the multi-core central processing unit obtains command data matching the creation result information from the buffer, and outputs the target space unique identifier and the creation success identifier for the command data.

In some embodiments of the present invention, regardless of whether the operation of creating a namespace is successful or not, the corresponding creation result information is sent to the NVMe module running in the multi-core central processor. As shown in Figure 6, it is a command processing flow chart for creating a namespace provided by some embodiments of the present invention. The host sends a creation instruction, the naming processing module receives the command, checks the command, and if the command is illegal and returns the operation code and ends the command, the command is legal, then the command is sent to the NVMe module, the NVMe module will determine the command type, if it is other commands, it will be handled by other modules, the present invention will not be explained here, if it is a namespace command, then the command needs to be sent to the data management processor, before sending, the N VMe module needs to apply for context and buffer and transfer the command to the buffer, after the transfer is completed, the N VMe module sends the command to the data management processor, the data management processor determines that it is a namespace creation command, then allocates resources from the namespace resource pool to the new namespace, and returns the creation result to the NVMe module; if it is a read-write command, it will enter the corresponding read-write process and return the result to the NVMe module. After the NVMe module gets the processing result from the data management processor, it will find this command from the previously applied context and buffer, and send a completion message for the command to the host. Specifically, if the namespace is created successfully, the detailed information of the new namespace (including the identifier, the starting logical address, and the allocated capacity) will be returned to the NVMe module. If the namespace creation fails (for example, insufficient resource pool space or invalid starting lba), an error message will be returned to the NVMe module.

The data processing method based on the dual-port solid-state hard disk provided by the embodiment of the present invention receives commands through a multi-core central processing unit. If the command is a namespace command, the command is forwarded to a data management processor, and the behavior type corresponding to the command is determined by the data management processor. The command is stored in different cache queues according to the behavior type corresponding to the command. The load index of each cache queue is calculated by the data management processor according to a preset load balancing algorithm, and the queue priority of each cache queue is determined according to the load index. The target cache queue is selected from the cache queue according to the queue priority by the data management processor, and the target command in the target cache queue is processed. The embodiment of the present invention sets a data manager in the dual-port solid-state hard disk to specifically process commands for the namespace, and ensures the orderly processing of each command cache queue through the load index, thereby freeing up part of the computing power of the multi-core central processing unit of the solid-state hard disk.

7 , a schematic diagram of the structure of a data processing device based on a dual-port solid-state hard disk provided by an embodiment of the present invention is shown. As shown in FIG7 , the device may include:

The command acquisition module 701 is used to receive a command through the multi-core central processing unit, and if the command is a namespace command, forward the command to the data management processor.

The command allocation module 702 is used to determine the behavior type corresponding to the command through the data management processor, and store the command in different cache queues according to the behavior type corresponding to the command.

The load balancing module 703 is used to calculate the load index of each cache queue according to a preset load balancing algorithm through the data management processor, and determine the queue priority of each cache queue according to the load index.

The command processing module 704 selects a target cache queue from the cache queues according to the queue priority through the data management processor, and processes the target command in the target cache queue.

In an optional embodiment of the present invention, the load balancing module 703 includes:

The load information acquisition submodule is used to obtain queue load information corresponding to each of the cache queues, and the queue load information at least includes: queue weight, command response time and current load amount.

The load average calculation submodule is used to calculate the total queue load of the cache queue according to the current load of each cache queue, and determine the average load of the cache queue according to the total queue load.

The load index calculation submodule is used to calculate the load index of each cache queue according to a preset load balancing algorithm.

The priority determination submodule is used to determine the queue priority of each cache queue according to the load index.

The formula corresponding to the preset load balancing algorithm is:

i is the total number of the cache queues, e is a constant, LI_i is the load index of the i-th cache queue, CI_i is the current load of the i-th cache queue, w_i is the queue weight corresponding to the i-th cache queue, w_avg is the average load of the cache queue, RT_i is the command response time corresponding to the i-th cache queue, α and β are adjustment parameters, and α and β are used to balance the load and response time.

In an optional embodiment of the present invention, the command acquisition module 701 includes:

A first check submodule, configured to receive a command through the multi-core central processing unit, check the command, and obtain a check result;

A second verification submodule, configured to intercept the command if the verification result is that the verification fails;

The third check submodule is used to forward the command to the data management process if the check result is check passed and the command is one of the namespace commands.

In an optional embodiment of the present invention, the command acquisition module 701 further includes:

The command acquisition submodule is used to receive the command through the command processing module running in the multi-core central processing unit, and send the command to the solid state drive control module running in the multi-core central processing unit.

The buffer application submodule is used to apply for a buffer through the solid state drive control module and transfer the command to the buffer.

The command forwarding submodule is used to forward the command to the data management processor through the solid state drive control module if the command is a namespace command.

In an optional embodiment of the present invention, the command processing module 704 includes:

The information acquisition submodule is used to obtain the creation information carried in the target command if the target command is a creation command.

The space allocation submodule is used to allocate the storage space in the namespace resource pool in the dual-port solid-state hard disk according to the creation information to obtain the target storage space corresponding to the creation information; wherein the namespace information at least includes: storage capacity and starting logical address.

The result returning submodule is used to send the creation result information corresponding to the target storage space to the solid state hard disk control module running in the multi-core central processing unit, wherein the creation result information includes a unique identifier of the target space and a creation success identifier.

In an optional embodiment of the present invention, the result returning submodule includes:

The result returning subunit is used to obtain the command data matching the creation result information from the buffer through the solid state drive control module running in the multi-core central processing unit, and output the target space unique identifier and the creation success identifier for the command data.

In an optional embodiment of the present invention, the data processing device based on the dual-port solid state hard disk further includes:

The first power-off module is used to receive a power-off command through the multi-core central processing unit and send the power-off command to the data management processor.

The second power-off module is used to stop receiving new commands after receiving the power-off command through the data management processor, and store the data managed by the data management processor in the flash memory of the dual-port solid state drive.

The data processing method based on the dual-port solid-state hard disk provided by the embodiment of the present invention receives commands through a multi-core central processing unit. If the command is a namespace command, the command is forwarded to a data management processor, and the behavior type corresponding to the command is determined by the data management processor. The command is stored in different cache queues according to the behavior type corresponding to the command. The load index of each cache queue is calculated by the data management processor according to a preset load balancing algorithm, and the queue priority of each cache queue is determined according to the load index. The target cache queue is selected from the cache queue according to the queue priority by the data management processor, and the target command in the target cache queue is processed. The embodiment of the present invention sets a data manager in the dual-port solid-state hard disk to specifically process commands for the namespace, and ensures the orderly processing of each command cache queue through the load index, thereby freeing up part of the computing power of the multi-core central processing unit of the solid-state hard disk.

An embodiment of the present invention further provides an electronic device, which may include a processor, a memory, and a computer program stored in the memory and capable of running on the processor. When the computer program is executed by the processor, the data processing method based on the dual-port solid-state hard disk is implemented as described above.

An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the data processing method based on a dual-port solid-state hard disk is implemented as described above.

Some embodiments of the present invention further provide a computer program product, including a computer program, which implements the above data processing method based on a dual-port solid-state hard disk when executed by a processor.

As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

The algorithm and display provided herein are not inherently related to any particular computer, virtual device or other equipment. According to the above description, it is obvious that the structure required for constructing this type of device is. In addition, the present invention is not directed to any specific programming language. It should be understood that various programming languages can be utilized to realize the content of the present invention described herein, and the description of the above specific language is to disclose the best mode of the present invention.

In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the intention that the claimed invention requires more features than those expressly recited in each claim. Rather, as reflected in the claims below, inventive aspects lie in less than all of the features of the individual embodiments disclosed above. Therefore, the claims that follow the detailed description are hereby expressly incorporated into the detailed description, with each claim itself serving as a separate embodiment of the present invention.

Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and/or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components in the sorting device according to the present invention. The present invention may also be implemented as a device or apparatus program for executing part or all of the methods described herein. Such a program for implementing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc., does not indicate any order. These words may be interpreted as names.

Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices, apparatuses and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

It should be noted that the various data-related processes in the embodiments of the present application are all carried out in compliance with the corresponding data protection laws and policies of the country where the device is located, and with the authorization given by the owner of the corresponding device.

Claims (10)

1. A data processing method based on a dual-port solid-state hard disk, characterized in that the dual-port solid-state hard disk is provided with a multi-core central processing unit and a data management processor, the method comprising: receiving a command through the multi-core central processing unit, and forwarding the command to the data management processor if the command is a namespace command; Determining, by the data management processor, a behavior type corresponding to the command, and storing the command in different cache queues according to the behavior type corresponding to the command; Calculating the load index of each cache queue according to a preset load balancing algorithm by the data management processor, and determining the queue priority of each cache queue according to the load index; The data management processor selects a target cache queue from the cache queue according to the queue priority, and processes the target command in the target cache queue. 2. The method according to claim 1, characterized in that the data management processor calculates the load index of each cache queue according to a preset load balancing algorithm, and determines the queue priority of each cache queue according to the load index, comprising: Obtaining queue load information corresponding to each cache queue, wherein the queue load information includes at least: queue weight, command response time, and current load; Calculating the total queue load of the cache queue according to the current load of each cache queue, and determining the average load of the cache queue according to the total queue load; Calculating the load index of each cache queue according to a preset load balancing algorithm; Determining a queue priority of each of the cache queues according to the load index; The formula corresponding to the preset load balancing algorithm is: i is the total number of the cache queues, e is a constant, LI_i is the load index of the i-th cache queue, CI_i is the current load of the i-th cache queue, w_i is the queue weight corresponding to the i-th cache queue, w_avg is the average load of the cache queue, RT_i is the command response time corresponding to the i-th cache queue, α and β are adjustment parameters, and α and β are used to balance the load and response time. 3. The method according to any one of claims 1 to 2, wherein the namespace command includes: a create command, a delete command, a write command, and a read command, and the receiving of the command by the multi-core central processor, if the command is a namespace command, forwarding the command to the data management processor, comprises: Receiving a command through the multi-core central processing unit, verifying the command, and obtaining a verification result; If the verification result is that the verification fails, intercepting the command; If the verification result is verification passed and the command is one of the namespace commands, the command is forwarded to the data management process. 4. The method according to claim 1, wherein the receiving a command through the multi-core central processor and forwarding the command to the data management processor if the command is a namespace command comprises: The command is received by a command processing module running in a multi-core central processing unit, and the command is sent to a solid state drive control module running in the multi-core central processing unit; Applying for a buffer through the solid state drive control module, and transferring the command to the buffer; If the command is a namespace command, the command is forwarded to the data management processor through the solid state drive control module. 5. The method according to claim 4, characterized in that processing the target command in the target cache queue comprises: If the target command is a creation command, obtaining creation information carried in the target command; According to the creation information, the storage space in the namespace resource pool in the dual-port solid-state hard disk is allocated to obtain a target storage space corresponding to the creation information; wherein the namespace information at least includes: storage capacity and a starting logical address; The creation result information corresponding to the target storage space is sent to the solid state hard disk control module running in the multi-core central processing unit, wherein the creation result information includes a unique identifier of the target space and a creation success identifier. 6. The method according to claim 5, characterized in that after sending the target storage space information corresponding to the target storage space to the solid state drive control module running in the multi-core central processor, it comprises: The solid state drive control module running in the multi-core central processing unit obtains command data matching the creation result information from the buffer, and outputs the target space unique identifier and the creation success identifier for the command data. 7. The method according to any one of claims 1 to 2, characterized in that the method further comprises: Receiving a power-off command through the multi-core central processing unit, and sending the power-off command to the data management processor; After receiving the power-off command, the data management processor stops receiving new commands and stores the data managed by the data management processor in the flash memory of the dual-port solid state drive. 8. A data processing device based on a dual-port solid-state hard disk, characterized in that the dual-port solid-state hard disk is provided with a multi-core central processing unit and a data management processor, comprising: A command acquisition module, configured to receive a command through the multi-core central processing unit, and if the command is a namespace command, forward the command to the data management processor; A command allocation module, used to determine the behavior type corresponding to the command through the data management processor, and store the command in different cache queues according to the behavior type corresponding to the command; A load balancing module, configured to calculate a load index of each cache queue according to a preset load balancing algorithm through the data management processor, and determine a queue priority of each cache queue according to the load index; The command processing module selects a target cache queue from the cache queues according to the queue priority through the data management processor, and processes the target command in the target cache queue. 9. An electronic device, characterized in that it comprises a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the data processing method based on a dual-port solid-state hard drive as described in any one of claims 1 to 7 is implemented. 10. A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the data processing method based on a dual-port solid-state hard disk according to any one of claims 1 to 7 is implemented.