A method and system for SSD abnormal power loss protection for enterprise-level storage

By establishing a closed topology table for SSD power loss and a cross-device collaborative write mechanism, the problems of high single-point failure risk and long recovery time in SSD abnormal power loss protection are solved, achieving efficient data protection and rapid recovery, and improving the data reliability and service availability of enterprise-level storage.

CN123114380APending Publication Date: 2026-11-17SHENZHEN PENGRUNNING TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611489926.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-23
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing SSD abnormal power loss protection methods lack cross-device collaborative protection mechanisms, resulting in a high risk of single point of failure. Furthermore, the data replay logic during power loss recovery is complex, and the mapping table reconstruction is time-consuming, making it difficult to guarantee data integrity and rapid service recovery.

Method used

By establishing an SSD power-down closure topology table, the identifier, fault domain, logical address range, closed storage area, closed write parameters, and built-in capacitor parameters of each SSD are collected to form a write closure record, determine the persistence boundary, divide the local closed content and the proxy closed content, and perform cross-device collaborative writing during abnormal power loss to generate a power-down closure record. After power is restored, data migration and address mapping updates are performed.

Benefits of technology

It improves data protection efficiency during abnormal power outages, avoids the risk of data loss due to energy depletion, simplifies the recovery process, shortens business interruption time, and enhances the data reliability and service availability of enterprise-level storage solutions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123114380A_ABST
    Figure CN123114380A_ABST
Patent Text Reader

Abstract

This invention discloses a method and system for SSD power failure protection for enterprise-level storage, relating to the field of SSD power failure protection technology. The method includes: encapsulating the write record to be protected to form closed content; segmenting the closed content into local and proxy closed content based on its data volume, write parameters, and built-in capacitor parameters; performing domain-based writing on the local and proxy closed content according to the SSD power failure closed topology table to generate a power failure closed record; reading the power failure closed record after power-on; establishing a recovery mapping between logical block addresses and closed storage area locations; updating the address mapping after data migration based on the recovery mapping; and generating the abnormal power failure protection result. This invention simplifies the recovery process by establishing a power failure closed record, shortens business interruption time, and improves the data reliability of enterprise-level storage solutions under abnormal power failure scenarios.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of SSD power loss protection technology, and in particular to a method and system for abnormal power loss protection of SSDs for enterprise-level storage. Background Technology

[0002] With the rapid development of big data, cloud computing, and artificial intelligence technologies, enterprise-level storage systems have placed unprecedentedly stringent demands on data processing performance, capacity, and reliability. Solid-state drives (SSDs), with their advantages of low latency, high throughput, and shock resistance, have gradually replaced traditional hard disk drives (HDDs) and become the core storage medium in enterprise-level data centers. In enterprise-level application scenarios, data consistency and persistence are the cornerstones of system design. To cope with the risk of abnormal power outages, existing technologies typically utilize the onboard backup power supply (such as tantalum capacitors or supercapacitors) of the SSD. Upon detecting a power outage signal, this power supply forces the "dirty data" in the DRAM cache to be written back to the non-volatile NAND flash memory chips as quickly as possible, thereby ensuring that data is not lost. This power outage protection mechanism based on backup power has become the standard configuration for ensuring the data integrity of enterprise-level SSDs, effectively supporting the continuity and stability of critical businesses such as financial transactions and online transaction processing.

[0003] Nevertheless, existing SSD abnormal power loss protection methods still have room for improvement. First, there is a lack of cross-device collaborative protection mechanisms. Once the SSD's local capacitors are depleted, data that has not yet been written will face the risk of permanent loss. There is also a lack of emergency channels to urgently migrate critical data to other healthy SSDs outside the same fault domain, resulting in a high risk of single point of failure. In addition, during the power loss recovery phase, there are often problems such as complex data replay logic and long time-consuming mapping table reconstruction, making it difficult to achieve rapid service recovery while ensuring data integrity. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an SSD abnormal power loss protection method for enterprise-level storage to solve the problem of high risk of single point of failure.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for protecting enterprise-level storage systems from abnormal power loss in SSDs, comprising: Collect the identifier, fault domain, logical address range, closed memory area, closed write parameters, emergency transfer relationship and built-in capacitor parameters of each SSD, and establish an SSD power failure closed topology table; After receiving a host write request and identifying the source SSD based on the logical address range, a write closure sequence number is assigned, and the logical block address, data cache location, write status, and host confirmation status are recorded to form a write closure record. Sequential verification is performed on write closure records whose host confirmation status is confirmed to determine the persistence boundary; After receiving an abnormal power failure signal, the source SSD uses the persistent boundary as the starting point for interception and performs boundary interception on the write closure record that the host has confirmed as confirmed, forming the write record to be protected. The record to be protected is closed and encapsulated to form closed content; The closure is segmented based on the data volume of the closure content, the closure write parameters, and the built-in capacitor parameters to form local closure content and proxy closure content. Based on the SSD power-down closure topology table, the local closure content and the proxy closure content are written in separate domains to generate power-down closure records; After power is restored, the power failure closure record is read, a recovery mapping between the logical block address and the location of the closed storage area is established, the local closed content and the proxy closed content are read according to the recovery mapping, the address mapping is updated after data migration, and the abnormal power failure protection result is generated.

[0007] As a preferred embodiment of the SSD abnormal power loss protection method for enterprise-level storage described in this invention, the step of establishing the SSD power loss closed topology table refers to using the identifier of each SSD as the association basis, binding each item with the fault domain, logical address range, closed storage area, closed write parameters, emergency transmission relationship and built-in capacitor parameters corresponding to each SSD to form a topology record corresponding to each SSD, and summarizing all topology records to form the SSD power loss closed topology table.

[0008] As a preferred embodiment of the SSD abnormal power loss protection method for enterprise-level storage described in this invention, wherein: the formation of the write closure record specifically includes: After receiving a host write request, read the logical block address carried in the host write request, and match the logical block address with the logical address range corresponding to each SSD in the SSD power-down closed topology table. Then, determine the SSD corresponding to the logical address range where the logical block address is located as the source SSD. Assign sequentially increasing and non-repeating write closure sequence numbers according to the order in which host write requests are received, write the write data carried by the host write request to the buffer and record the data buffer position; During the writing process of the source SSD, the write status is recorded according to whether the written data has been persisted, and the host confirmation status is recorded according to whether the host write request has returned a write success confirmation. Using the write closure sequence number as the association basis, the logical block address, data cache location, write status, and host confirmation status are bound one by one to form a write closure record.

[0009] As a preferred embodiment of the SSD abnormal power loss protection method for enterprise-level storage described in this invention, wherein: determining the persistence boundary specifically includes: Read the write closure sequence number, write status, and host confirmation status from the write closure record, filter the write closure records with the host confirmation status "confirmed", and sort the write closure records with the host confirmation status "confirmed" in ascending order of write closure sequence number; Starting from the smallest write closure sequence number, check whether adjacent write closure sequence numbers are consecutive, and check whether the write state corresponding to each write closure sequence number is persisted; Continue sequential verification when adjacent write closure sequence numbers are consecutive and the write status is persisted; stop sequential verification when adjacent write closure sequence numbers are not consecutive and the write status is not persisted; and determine the last write closure sequence number that passes consecutive verification as the persistence boundary.

[0010] As a preferred embodiment of the SSD abnormal power loss protection method for enterprise-level storage described in this invention, wherein: the formation of the write record to be protected specifically includes: Upon receiving an abnormal power failure signal, read the write closure sequence number corresponding to the persistence boundary, and read the write closure sequence number, logical block address, data cache location, write status, and host confirmation status from the write closure record; Filter the write closure records whose host confirmation status is confirmed and whose write closure sequence number is greater than the write closure sequence number corresponding to the persistence boundary, and sort the filtered write closure records in ascending order of write closure sequence number; Using the first write closure sequence number after the persistent boundary as the starting position for truncation, the logical block address, data cache location, write status, and host confirmation status corresponding to each write closure sequence number are truncated sequentially, while maintaining the order between the write closure sequences. The truncated write closure records are then aggregated to form the write record to be protected.

[0011] As a preferred embodiment of the SSD abnormal power loss protection method for enterprise-level storage described in this invention, the formation of closed content specifically includes: Read the write closure sequence number, logical block address, data buffer location, write status, and host confirmation status from the write record to be protected, and sort the write records to be protected in ascending order of write closure sequence number; Read the written data according to the data cache location corresponding to each write closure sequence number, and verify the correspondence between the written data and the logical block address, write status and host confirmation status; Each write closure sequence number is sequentially bound to its corresponding logical block address, write status, host confirmation status, and write data. These are then concatenated continuously according to the order of the write closure sequence numbers. The total number of bytes after concatenation is counted as the data volume of the closure content, thus forming the closure content.

[0012] As a preferred embodiment of the SSD abnormal power loss protection method for enterprise-level storage described in this invention, the formation of local closed content and proxy closed content specifically includes: Read the amount of data in the closed content, and read the closed write parameters and built-in capacitor parameters of the source SSD from the SSD power-down closed topology table; The available capacitor energy that the source SSD can provide during an abnormal power outage is calculated based on the built-in capacitor capacity, current terminal voltage, minimum closing operating voltage and energy conversion efficiency in the built-in capacitor parameters. Calculate the amount of closed write consumption required for the source SSD to complete the unit data write and write verification based on the write time, write energy consumption and write verification time in the closed write parameters; The available capacitor energy is converted into the closed write consumption to obtain the amount of data that the source SSD can complete the closed write. Based on the cumulative number of bytes corresponding to each write closure sequence number in the cumulative closed content from smallest to largest, closed content whose cumulative byte count does not exceed the amount of data that the source SSD can complete the closure write is classified as local closed content, and closed content whose cumulative byte count exceeds the amount of data that the source SSD can complete the closure write is classified as proxy closed content.

[0013] As a preferred embodiment of the SSD abnormal power loss protection method for enterprise-level storage described in this invention, the generation of a power loss closure record specifically includes: Read the SSD identifier, fault domain, closed storage area and emergency transmission relationship corresponding to the source SSD in the SSD power failure closed topology table, and write the local closed content to the closed storage area of ​​the source SSD in ascending order of write closure sequence number; Based on the emergency transfer relationship, select the SSDs that have established an emergency transfer relationship with the source SSD, exclude the SSDs with the same fault domain as the source SSD, arrange the remaining SSDs according to the character order of the SSD identifier, determine the SSD at the top of the list as the proxy SSD, and transfer the proxy closed content to the proxy SSD and write it to the closed storage area of ​​the proxy SSD in ascending order of the write closure sequence number. After completing the domain-specific writing of local and proxy closed content, read the logical block address corresponding to each write closure sequence number in the local and proxy closed content, and read the SSD identifier and closed storage area location that saves the content corresponding to each write closure sequence number. Bind the write closure sequence number, logical block address, SSD identifier and closed storage area location item by item, and summarize them according to the write closure sequence number to form a power failure closure record.

[0014] As a preferred embodiment of the SSD abnormal power loss protection method for enterprise-level storage described in this invention, the generation of abnormal power loss protection results specifically includes: After power is restored, the write closure sequence number, logical block address, SSD identifier, and closed storage area location in the power failure closure record are read in ascending order according to the write closure sequence number. Using the logical block address as an index, the logical block address, SSD identifier, and closed storage area location are bound item by item to establish a recovery mapping between the logical block address and the closed storage area location. Based on the SSD identifier in the recovery map, locate the source SSD that stores the local closed content and the proxy SSD that stores the proxy closed content. Based on the closed storage area location in the recovery map, read the local closed content and the proxy closed content respectively, and verify whether the write closure sequence number and logical block address in the local closed content and the proxy closed content are consistent with the write closure sequence number and logical block address in the power failure closure record. After verification, the source SSD corresponding to each logical block address is identified based on the logical address range in the SSD power-down closure topology table. The write data in the local closure content and the proxy closure content are migrated back to the source SSD in ascending order of write closure sequence number. After completing the data migration, update the address mapping corresponding to the logical block address, summarize the write closure sequence number and logical block address of the completed data migration and address mapping update, and generate the abnormal power failure protection result.

[0015] Secondly, this invention provides an SSD abnormal power loss protection system for enterprise-level storage, comprising, The acquisition module collects the identifier, fault domain, logical address range, closed memory area, closed write parameters, emergency transfer relationship and built-in capacitor parameters of each SSD, and establishes an SSD power failure closed topology table; The recording module receives a host write request and identifies the source SSD based on the logical address range. It then assigns a write closure sequence number, records the logical block address, data cache location, write status, and host confirmation status, and forms a write closure record. The verification module performs sequential verification on write closure records whose status is confirmed by the host to determine the persistence boundary. The interception module intercepts the write closed records that the host has confirmed as confirmed after receiving an abnormal power failure signal from the source SSD, taking the persistent boundary as the interception starting point, and forms the write records to be protected. The encapsulation module performs closed encapsulation based on the write record to be protected, forming closed content; The segmentation module performs closure segmentation based on the data volume of the closure content, closure write parameters, and built-in capacitor parameters, forming local closure content and proxy closure content. The write module performs domain-based writing of local closed content and proxy closed content according to the SSD power-down closed topology table, generating power-down closed records; The module is established, and after power-on, the power-down closing record is read. A recovery mapping between the logical block address and the location of the closed storage area is established. Based on the recovery mapping, the local closed content and the proxy closed content are read, the data is back-transferred, and the address mapping is updated to generate the abnormal power-down protection result.

[0016] The beneficial effects of this invention are as follows: By determining the persistence boundary, the range of critical data that needs to be protected during abnormal power outages is locked, effectively improving protection efficiency. Furthermore, based on the real-time energy status of the SSD, the data to be protected is intelligently divided into local closed content and proxy closed content. By utilizing a cross-device collaborative mechanism, data exceeding the local capacitor's carrying capacity is urgently migrated to other healthy SSDs, thereby breaking through the physical limitations of a single device's capacitor energy and mitigating the risk of data loss due to energy depletion to a certain extent. Finally, by generating a power outage closure record, a "data map" is provided for the recovery process, enabling direct location and rapid data migration after power restoration. This simplifies the recovery process, shortens business interruption time, and improves the data reliability and service availability of enterprise-level storage solutions in abnormal power outage scenarios. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of an SSD abnormal power loss protection method for enterprise-level storage.

[0019] Figure 2 A flowchart for creating a closed topology table for SSD power-down.

[0020] Figure 3 A flowchart for forming local closed content and proxy closed content.

[0021] Figure 4 A flowchart for generating a power-down closure record. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for protecting enterprise-level storage SSDs from abnormal power loss, including the following steps: S1: Collect the identifier, fault domain, logical address range, closed storage area, closed write parameters, emergency transfer relationship and built-in capacitor parameters of each SSD, and establish an SSD power-down closed topology table.

[0026] S1.1: It should be noted that, in sequence, an identifier query command is sent to the management interface of each SSD, and the unique serial number returned by each SSD is received. The unique serial number is then used as the identifier of each SSD. A slot topology query command is sent to the backplane management interface to obtain the power input number and backplane number corresponding to the backplane slot where each SSD is located. SSDs with the same power input number and backplane number are assigned to the same fault domain. A logical address range query command is sent to the logical address configuration interface to receive the start value and end value of the logical block address corresponding to each SSD. The logical block address range is then assigned to the logical address configuration interface. The continuous address range between the starting address value and the ending logical block address value is defined as the logical address range; a reserved storage configuration query command is sent to the management interface of each SSD to receive the starting physical location and writable capacity of the reserved non-volatile storage space, and the storage range jointly defined by the starting physical location and the writable capacity is defined as the closed storage area; a write granularity query command is sent to the management interface of each SSD to obtain the minimum write granularity L, and the byte position is numbered continuously starting from 0. The byte value corresponding to each byte position is set to the remainder of the byte position number divided by 256, and so on. Arrange all byte values ​​sequentially according to byte position number to generate test data of length L. Use the starting physical position of the closed storage area as the test write position. Record the write start time when the test data write command is issued, and record the write completion time when the test data write completion confirmation is received. Read back the test data from the test write position, and check the read back test data and the written test data byte by byte according to byte position number. Record the write verification completion time when all byte values ​​are consistent. From the write start time to the write verification completion time, synchronously collect the write voltage value and write current value corresponding to each sampling time according to the shortest sampling period supported by the voltage sampling interface. The shortest sampling period is set to 100 microseconds. Calculate the difference between the write completion time and the write start time to obtain the write time consumption. Multiply the time difference between each adjacent sampling time with the write voltage value and write current value corresponding to the previous sampling time to obtain the energy consumption of each sampling interval. Add up the energy consumption of all sampling intervals to obtain the write energy consumption. Subtract the write completion time from the write verification completion time to obtain the write verification time consumption. Sum the write time consumption, write energy consumption, and write verification time consumption to form the closed write parameters.

[0027] Following the order of the SSDs' identifiers, each SSD is sequentially used as the SSD for sending test data. Each SSD other than the one for sending test data is sequentially used as the SSD for receiving test data. Test data is transferred from the sending SSD to the receiving SSD, and then read back from the receiving SSD. The transmission and retrieval of test data are completed within the response time limit specified in the interface protocol. When the read-back test data matches the sent test data byte-by-byte, the identifiers of the sending and receiving SSDs and the transmission direction are recorded. The response time limit specified in the interface protocol is set to 50 milliseconds. (The rest of the text is incomplete and likely refers to a summary process.) The system establishes emergency transmission relationships; it progressively reduces the power supply voltage of each SSD according to the minimum voltage step size allowed by the SSD management interface, with the minimum voltage step size set to 0.05V. Test data writing and byte-by-byte verification are repeatedly performed at each power supply voltage level. The power supply voltage at the last level that allows for complete test data writing and consistent byte-by-byte verification is determined as the minimum closing operating voltage. After charging the internal capacitor to its rated voltage, the external power supply is disconnected. Using the test data writing process as a discharge load, from the disconnection of the external power supply until the internal capacitor voltage drops to the minimum closing operating voltage, the internal capacitor voltage at each sampling moment is synchronously collected at the shortest sampling period of 100 microseconds. The output voltage value, built-in capacitor output current value, SSD input voltage value, and SSD input current value are used to calculate the discharge charge in each sampling interval. The time difference between each adjacent sampling moment is multiplied by the built-in capacitor output current value corresponding to the previous sampling moment to obtain the discharge charge. The discharge charge in all sampling intervals is summed to obtain the total discharge charge. The voltage difference between the rated voltage and the minimum closing operating voltage of the built-in capacitor is calculated, and the total discharge charge is divided by the voltage difference to obtain the built-in capacitor capacity. When acquiring data from each SSD, the voltage across the built-in capacitor is acquired through the voltage sampling interface to obtain the current terminal voltage. The time difference between each adjacent sampling moment is multiplied by the built-in capacitor output current value corresponding to the previous sampling moment to obtain the discharge charge. The built-in capacitor output energy is obtained by multiplying the capacitor output voltage value and the built-in capacitor output current value in each sampling interval. The built-in capacitor output energy is accumulated in all sampling intervals to obtain the built-in capacitor output energy. The SSD received energy in each sampling interval is obtained by multiplying the time difference between each adjacent sampling moment with the SSD input voltage value and SSD input current value corresponding to the previous sampling moment. The SSD received energy is accumulated in all sampling intervals to obtain the SSD received energy. The SSD received energy is then divided with the built-in capacitor output energy to obtain the energy conversion efficiency. The built-in capacitor capacitance, current terminal voltage, minimum closing operating voltage, and energy conversion efficiency together constitute the built-in capacitor parameters.

[0028] S1.2: It should be noted that the SSDs are arranged in ascending order of their identifiers, and one SSD identifier is selected sequentially as the current association basis. Based on the current association basis, the fault domain to which the current SSD belongs is located in the fault domain division results; the start and end values ​​of the logical block address corresponding to the current SSD are located in the logical address range; the starting physical location and writable capacity corresponding to the current SSD are located in the closed storage area; the write time, write power consumption, and write verification time corresponding to the current SSD are located in the closed write parameters; the transmission records in the emergency transmission relationship where the identifier of the SSD sending test data matches the identifier of the current SSD are selected; and the current SSD is located in the built-in capacitor parameters. The corresponding built-in capacitor capacity, current terminal voltage, minimum closed operating voltage, and energy conversion efficiency are recorded. After verifying that the fault domain, logical address range, closed memory area, closed write parameters, emergency transfer relationship, and built-in capacitor parameters correspond to the current SSD's identifier, the topology record for the current SSD is recorded item by item in the order of SSD identifier, fault domain, logical address range, closed memory area, closed write parameters, emergency transfer relationship, and built-in capacitor parameters. The item-by-item location, correspondence verification, and item-by-item recording for each SSD are repeated to obtain the topology record for each SSD. All topology records are arranged in ascending order according to the identifier of each SSD to form the SSD power-down closed topology table.

[0029] S2: After receiving the host write request and identifying the source SSD based on the logical address range, it assigns a write closure sequence number, records the logical block address, data cache location, write status, and host confirmation status, and forms a write closure record.

[0030] S2.1: It should be noted that after receiving a write request from the host, the logical block address is read from the address field of the write request. The topology records in the SSD power-down closed topology table are read sequentially from smallest to largest according to the identifier of each SSD. The logical block address start value and logical block address end value corresponding to the logical address range are read from each topology record. The logical block address is compared with the logical block address start value and logical block address end value in turn. When the logical block address is greater than or equal to the logical block address start value and less than or equal to the logical block address end value, it is determined that the logical block address falls within the logical address range. The reading of subsequent topology records is stopped, and the SSD identifier is read from the topology record where the logical address range falls within the logical block address. The SSD corresponding to the SSD identifier is identified as the source SSD.

[0031] S2.2: It should be noted that host write requests are processed sequentially according to the order in which they are received. For the first host write request, the write closure sequence number is set to 1. For subsequent host write requests, the write closure sequence number corresponding to the previous host write request is read, and the write closure sequence number corresponding to the previous host write request is incremented by 1 to become the write closure sequence number corresponding to the current host write request, so that all write closure sequence numbers are continuously incremented and do not repeat. The write data and the amount of write data carried by the host write request are read. A continuous free cache area is searched in ascending order of cache address. When the available capacity of the continuous free cache area is greater than or equal to the amount of write data, the starting address of the continuous free cache area is determined as the data cache location. The write data is written to the cache in byte order starting from the data cache location. The write closure sequence number, data cache location, and amount of write data are bound item by item to complete the data cache location record.

[0032] Read the write data from the data cache location corresponding to the write closure sequence number, and send a persistent write command to the source SSD according to the logical block address corresponding to the write closure sequence number. When the persistent write command is issued, record the write status corresponding to the write closure sequence number as "writing in progress". After the source SSD completes the non-volatile storage write of the write data and returns a persistent completion flag, update the write status corresponding to the write closure sequence number from "writing in progress" to "persistent". Write closure sequence numbers that have not returned a persistent completion flag remain "writing in progress". Before returning a write success confirmation to the host, record the host confirmation status corresponding to the write closure sequence number as "unconfirmed". After returning a write success confirmation to the host, update the host confirmation status corresponding to the write closure sequence number from "unconfirmed" to "confirmed", and bind the write status and host confirmation status to the write closure sequence number respectively.

[0033] The logical block address, data cache location, write status, and host confirmation status corresponding to each write closure sequence number are read sequentially. Using the write closure sequence number as the record index, the write closure sequence number, logical block address, data cache location, write status, and host confirmation status are associated and saved in a fixed order to form a write closure record.

[0034] S3: Perform sequential verification on write closure records whose host confirmation status is confirmed to determine the persistence boundary.

[0035] It should be noted that the write closure sequence number, write status, and host confirmation status are read from the write closure record. Write closure records with a confirmed host status are filtered out, and these records are arranged in ascending order of their write closure sequence numbers. The first write closure sequence number is determined as the minimum write closure sequence number. The write status corresponding to the minimum write closure sequence number is read, and its consistency is checked against the persisted status. When the write status corresponding to the minimum write closure sequence number is persisted, this minimum write closure sequence number is recorded as the last consecutively verified write closure sequence number. Starting from the minimum write closure sequence number, two adjacent write closure records are selected sequentially. The write closure sequence number following the last one is subtracted from the first one, resulting in... The system calculates the difference between the write closure sequence numbers and reads the write status corresponding to the next write closure sequence number. If the difference between the write closure sequence numbers is equal to 1 and the write status corresponding to the next write closure sequence number is already persisted, the system updates the next write closure sequence number to the last consecutively verified write closure sequence number and continues to select the next group of adjacent write closure records for verification. If the difference between the write closure sequence numbers is not equal to 1, the sequential verification stops. If the write status corresponding to the next write closure sequence number is not already persisted, the sequential verification stops. The last consecutively verified write closure sequence number before the sequential verification stops is determined as the persistence boundary. If the write status corresponding to the smallest write closure sequence number is not already persisted, the sequential verification stops, and the persistence boundary determined by the most recent sequential verification remains unchanged.

[0036] S4: After the source SSD receives an abnormal power failure signal, it uses the persistent boundary as the starting point to perform boundary truncation on the write closure record that the host has confirmed as confirmed, forming a write record to be protected.

[0037] S4.1: It should be noted that upon receiving an abnormal power failure signal, the write closure sequence number corresponding to the persistent boundary is read from the storage location of the persistent boundary, and this write closure sequence number is used as the comparison benchmark for subsequent boundary truncation. Following the storage order of the write closure records, the write closure sequence number and host confirmation status corresponding to each write closure record are read one by one. The host confirmation status is checked for consistency with the confirmed status, and the write closure sequence number is compared numerically with the write closure sequence number corresponding to the persistent boundary. When the host confirmation status is confirmed and the write closure sequence number is greater than the write closure sequence number corresponding to the persistent boundary, the corresponding write closure record is retained. The process involves several steps: First, if the host confirmation status is not confirmed, the corresponding write closure record is excluded. Second, if the write closure sequence number is less than or equal to the write closure sequence number corresponding to the persistence boundary, the corresponding write closure record is excluded. Third, the write closure sequence numbers of two adjacent filtered write closure records are compared sequentially. If the write closure sequence number that appears earlier is greater than the write closure sequence number that appears later, the order of the two adjacent write closure records is swapped. This process is repeated until each of the preceding write closure sequence numbers is less than the following write closure sequence number, resulting in a list of filtered write closure records arranged in ascending order of write closure sequence number.

[0038] S4.2: It should be noted that the filtered write closure records are read in ascending order of write closure sequence number. When a filtered write closure record exists, the first write closure sequence number is determined as the smallest write closure sequence number greater than the write closure sequence number corresponding to the persistence boundary, and this smallest write closure sequence number greater than the write closure sequence number corresponding to the persistence boundary is determined as the starting position for truncation after the boundary. Starting from the starting position for truncation after the boundary, the logical block address, data cache location, write status, and host confirmation status corresponding to each write closure sequence number are read sequentially in ascending order of write closure sequence number. The write closure sequence number, logical block address, data cache location, write status, and host confirmation status are saved item by item in a fixed order, while maintaining the order of write closure sequence numbers among the write closure records. After truncating each filtered write closure record, the truncated write closure records are summarized in ascending order of write closure sequence number to form the write record to be protected. When a filtered write closure record does not exist, a write record to be protected that does not contain a write closure record is formed.

[0039] S5: Close and encapsulate the record to be protected to form closed content.

[0040] It should be noted that the write records to be protected are read sequentially from smallest to largest according to the write closure sequence number. From each write record to be protected, the write closure sequence number, logical block address, data cache location, write status, and host confirmation status are read. Based on the data cache location, the write closure sequence number and the amount of data to be written are read from the corresponding data cache location. Starting from the data cache location, consecutive bytes equal to the amount of data to be written are read to obtain the written data. The write closure sequence number corresponding to the data cache location is checked for consistency with the write closure sequence number in the write record to be protected. If the two write closure sequence numbers are consistent, the write closure sequence number is used as the association basis, and the written data is matched item by item with the logical block address, write status, and host confirmation status corresponding to the same write closure sequence number. The process involves several steps: First, when the written data, logical block address, write status, and host confirmation status all correspond to the same write closure sequence number, verify the correspondence between the written data and the logical block address, write status, and host confirmation status. After verification, each write closure sequence number is bound sequentially to its corresponding logical block address, write status, host confirmation status, the amount of written data, and the written data in a fixed order. These are then concatenated continuously in ascending order of the write closure sequence number. The total number of bytes after concatenation is counted, and this total number is determined as the data volume of the closed content. Finally, the concatenated write closure sequence number, logical block address, write status, host confirmation status, the amount of written data, and the written data are stored together to form the closed content.

[0041] S6: Based on the data volume of the closed content, the closed write parameters, and the built-in capacitor parameters, the closed content is segmented to form local closed content and proxy closed content.

[0042] S6.1: It should be noted that the built-in capacitor parameters, including built-in capacitor capacity, current terminal voltage, minimum closing operating voltage, and energy conversion efficiency, are read. The current terminal voltage is compared with the minimum closing operating voltage. When the current terminal voltage is greater than the minimum closing operating voltage, the squares of the current terminal voltage and the minimum closing operating voltage are calculated respectively. The difference between the squares of the current terminal voltage and the minimum closing operating voltage is calculated to obtain the available voltage square difference. The product of the built-in capacitor capacity and the available voltage square difference is then divided by the value 2 to obtain the capacitor energy that the built-in capacitor can release during the period when the current terminal voltage drops to the minimum closing operating voltage. The capacitor energy is multiplied by the energy conversion efficiency to obtain the available capacitor energy that the source SSD can receive and use for closing writes during abnormal power loss. When the current terminal voltage is less than or equal to the minimum closing operating voltage, the available capacitor energy that the source SSD can provide during abnormal power loss is determined to be 0.

[0043] S6.2: It should be noted that the write time, write energy consumption, and write verification time in the closed write parameters are read, and the minimum write granularity L used when forming the closed write parameters is read; the write time and the write verification time are summed to obtain the closed operation time required to complete the data write and write verification corresponding to the minimum write granularity L; the write energy consumption and the closed operation time are compared to obtain the average closed power during the data write and write verification of the source SSD; the average closed power and the closed operation time are multiplied to obtain the energy consumption required to complete the data write and write verification corresponding to the minimum write granularity L; the energy consumption and the number of bytes corresponding to the minimum write granularity L are compared to obtain the closed write consumption required by the source SSD to complete the data write and write verification per byte.

[0044] The available capacitor energy is divided by the closed write consumption to obtain the theoretical number of bytes that the source SSD can support during an abnormal power outage. The fractional part of the theoretical byte number is rounded down to obtain the integer theoretical byte number. The integer theoretical byte number is divided by the number of bytes corresponding to the minimum write granularity L and the integer quotient is retained. The integer quotient is multiplied by the number of bytes corresponding to the minimum write granularity L to obtain the number of bytes that meet the minimum write granularity requirement of the source SSD. The number of bytes that meet the minimum write granularity requirement of the source SSD is determined as the amount of data that the source SSD can complete a closed write.

[0045] Read the closed content sequentially from smallest to largest according to the write closure sequence number. Initialize the cumulative byte count to 0. Calculate the current write closure sequence number, its corresponding logical block address, write status, host confirmation status, the amount of data written, and the number of bytes occupied by the written data to obtain the byte count corresponding to the current write closure sequence number. Add the byte count corresponding to the current write closure sequence number to the cumulative byte count to obtain the updated cumulative byte count. If the updated cumulative byte count is less than or equal to the amount of data that the source SSD can complete the closed write, include the current write closure sequence number, its corresponding logical block address, write status, host confirmation status, the amount of data written, and the written data in the local closed content. Use the updated cumulative byte count as the cumulative byte count when reading the next write closure sequence number, and continue reading the next write closure sequence number. If the updated cumulative byte count is greater than the amount of data that the source SSD can complete the closed write, add the current write closure sequence number, its corresponding logical block address, write status, host confirmation status, the amount of data written, and the written data in the local closed content. Use the updated cumulative byte count as the cumulative byte count when reading the next write closure sequence number, and continue reading the next write closure sequence number. When a closed write operation is completed, the current write closure sequence number, its corresponding logical block address, write status, host confirmation status, the amount of data written, and the written data are assigned to the proxy closure content. All write closure sequences following the current write closure sequence number, along with their corresponding logical block addresses, write status, host confirmation status, the amount of data written, and the written data, are sequentially assigned to the proxy closure content, and reading subsequent write closure sequences is stopped. When all write closure sequences have been read and the updated cumulative byte count is consistently less than or equal to the amount of data the source SSD can complete the closed write operation, all closure content is assigned to the local closure content, and the proxy closure content is set to empty. When the number of bytes corresponding to the first written closure sequence number exceeds the amount of data the source SSD can complete the closed write operation, the local closure content is set to empty, and all closure content is assigned to the proxy closure content.

[0046] S7: Based on the SSD power-down closure topology table, write local closure content and proxy closure content separately to generate power-down closure records.

[0047] S7.1: It should be noted that, following the order of the topology records in the SSD power-down closure topology table, the SSD identifier corresponding to each topology record is read sequentially. The consistency of the SSD identifier corresponding to each topology record is checked against the SSD identifier corresponding to the source SSD. If the two SSD identifiers match, the fault domain, closed storage area, and emergency transfer relationship corresponding to the source SSD are read from the topology record with the matching SSD identifiers. The starting physical location and writable capacity are also read from the closed storage area. The local closed content is read sequentially from smallest to largest according to the write closure sequence number. The starting physical location of the closed storage area is determined as the write position corresponding to the first write closure sequence number in the table. The system first assigns the first write closure sequence number, along with the corresponding logical block address, write status, host confirmation status, amount of data to be written, and the data itself to the corresponding write location. After each write closure sequence number is completed, the corresponding write location is summed with the number of bytes occupied by the corresponding content to obtain the write location corresponding to the next write closure sequence number. The system then continues to write the next write closure sequence number, along with the corresponding logical block address, write status, host confirmation status, amount of data to be written, and the data itself, until all local closed content is written to the closed storage area of ​​the source SSD, while maintaining the original order of the write closure sequence numbers within the local closed content.

[0048] S7.2: It should be noted that, from the SSD power-down closed topology table, the emergency transmission relationship corresponding to the source SSD is read. The sending SSD identifier, receiving SSD identifier, and transmission direction recorded in each emergency transmission relationship are read sequentially. The sending SSD identifier is checked against the source SSD identifier. Emergency transmission relationships where the sending SSD identifier matches the source SSD identifier, the sending end of the transmission direction is the source SSD, and the receiving end of the transmission direction is the receiving SSD are retained. The receiving SSD is then identified as a candidate SSD. Based on the SSD identifier of the candidate SSD, the fault domain corresponding to the candidate SSD is read from the SSD power-down closed topology table. Candidate SSDs with the same fault domain as the source SSD are excluded. The remaining candidate SSDs are arranged according to the character order of their SSD identifiers. The candidate SSD ranked first is identified as the proxy SSD. Based on the SSD identifier of the proxy SSD, the closed storage area corresponding to the proxy SSD is read from the SSD power-down closed topology table, and the closed storage area corresponding to the proxy SSD is then determined. The system reads the starting physical location and writable capacity from the combined storage area, counts the total number of bytes occupied by the proxy closure content, and reads the proxy closure content sequentially from smallest to largest according to the write closure sequence number. Each write closure sequence number, along with its corresponding logical block address, write status, host confirmation status, write data volume, and write data, is transferred to the proxy SSD according to the emergency transfer relationship. The starting physical location of the closed storage area corresponding to the proxy SSD is determined as the write position corresponding to the first write closure sequence number. After completing the writing of the current write closure sequence number and its corresponding logical block address, write status, host confirmation status, write data volume, and write data, the current write position is summed with the number of bytes occupied by the content corresponding to the current write closure sequence number to obtain the write position corresponding to the next write closure sequence number. This process of transmission and writing is repeated from smallest to largest according to the write closure sequence number until all the proxy closure content is written to the closed storage area of ​​the proxy SSD.

[0049] After completing the domain-specific writing of local and proxy closed content, the logical block address corresponding to each write closure sequence number in the local and proxy closed content is read sequentially from smallest to largest according to the write closure sequence number. The SSD identifier and write position recorded during the domain-specific writing process for each write closure sequence number are also read, and the write position is determined as the corresponding closed storage area location. Using the write closure sequence number as the association basis, the write closure sequence number, logical block address, SSD identifier, and closed storage area location are bound item by item, and all binding results are summarized from smallest to largest according to the write closure sequence number to form a power-down closure record.

[0050] S8: After power is restored, read the power failure closing record, establish a recovery mapping between the logical block address and the location of the closed storage area, read the local closing content and the proxy closing content according to the recovery mapping, perform data backtracking, update the address mapping, and generate the abnormal power failure protection result.

[0051] S8.1: It should be noted that after power-on, the write closure sequence number, logical block address, SSD identifier, and closed storage area location are read sequentially from each power-down closure record; the recovery mapping is initialized to empty, and the logical block address is used as the index to query whether the same logical block address exists in the recovery mapping; if the same logical block address does not exist in the recovery mapping, the write closure sequence number, logical block address, SSD identifier, and closed storage area location are bound item by item and written to the recovery mapping; if the same logical block address exists in the recovery mapping, the write closure sequence number corresponding to the same logical block address in the recovery mapping is read, and the currently read write closure sequence number is compared with the write closure sequence number corresponding to the same logical block address in the recovery mapping; when... When the write closure sequence number read previously is greater than the write closure sequence number corresponding to the same logical block address in the recovery map, the write closure sequence number read previously, the SSD identifier, and the closed storage area location are replaced with the write closure sequence number read previously, the SSD identifier, and the closed storage area location in the recovery map. When the write closure sequence number read previously is less than or equal to the write closure sequence number corresponding to the same logical block address in the recovery map, the write closure sequence number read previously, the SSD identifier, and the closed storage area location in the recovery map remain unchanged. After completing the sequential reading and recovery map update of all power-down closure records, a recovery map is formed with the logical block address as the index and records the write closure sequence number, the SSD identifier, and the closed storage area location.

[0052] S8.2: As should be noted, the write closure sequence number, logical block address, SSD identifier, and closed storage area location in the recovery map are read sequentially; the logical block address is matched sequentially with the logical address range corresponding to each SSD in the SSD power-down closure topology table, and the SSD corresponding to the logical address range where the logical block address is located is identified as the source SSD; based on the SSD identifier in the recovery map, the SSD that stores the corresponding content is located from the SSD power-down closure topology table, and the consistency check is performed between the SSD identifier in the recovery map and the identifier of the source SSD. When the SSD identifier in the recovery map matches the identifier of the source SSD, the SSD that stores the corresponding content is identified as the source SSD, and the local closed content is read according to the closed storage area location; when the SSD identifier in the recovery map does not match the identifier of the source SSD, the SSD that stores the corresponding content is identified as the proxy SSD, and the proxy closed content is read according to the closed storage area location; the write operation begins from the closed storage area location. The process involves taking the closure sequence number, logical block address, write status, host confirmation status, and the amount of data written. Then, consecutive bytes are read according to the amount of data written to obtain the written data. Based on the SSD identifier and closed storage area location in the recovery mapping, the power-down closure record with the same SSD identifier and closed storage area location is located from the power-down closure record. The write closure sequence number and logical block address in the local closure content and the proxy closure content are compared with the write closure sequence number and logical block address in the located power-down closure record for consistency. When the write closure sequence number and logical block address are consistent, the write data corresponding to the same write closure sequence number is written to the source SSD according to the logical block address. When there is a discrepancy between the write closure sequence number and the logical block address, the write data migration for the corresponding write closure sequence number is stopped. This process of source SSD identification, content reading, consistency verification, and write data migration is repeated until all write data that has passed the consistency verification in the local closure content and the proxy closure content has been migrated back.

[0053] After completing the write data rollback corresponding to each write closure sequence number, the write completion confirmation returned by the source SSD is read. The write completion confirmation includes the write result and write position. Based on the write closure sequence number and logical block address saved when the write data rollback instruction was issued, the write completion confirmation is associated with the corresponding write closure sequence number and logical block address. When the write result indicates that the write data rollback is complete, the write position in the write completion confirmation is read. Using the logical block address as the lookup index for address mapping, the identifier and storage location of the SSD currently corresponding to the logical block address in the address mapping are read. The identifier of the SSD currently corresponding to the logical block address is updated to the identifier of the source SSD, and the logical block address is updated... The current corresponding storage location is updated to the write location in the write completion confirmation. After the address mapping update is completed, the logical block address is used again as the query index for address mapping to read the updated SSD identifier and storage location. The updated SSD identifier is checked for consistency with the source SSD identifier, and the updated storage location is checked for consistency with the write location in the write completion confirmation. When the SSD identifier and storage location are consistent, the address mapping update check is confirmed to be successful. The write closure sequence number of the completed write data migration and address mapping update is bound to the logical block address item by item. All binding results are summarized in ascending order of write closure sequence number to generate the abnormal power loss protection result.

[0054] This embodiment also provides an SSD abnormal power loss protection system for enterprise-level storage, including: The acquisition module collects the identifier, fault domain, logical address range, closed memory area, closed write parameters, emergency transfer relationship and built-in capacitor parameters of each SSD, and establishes an SSD power failure closed topology table; The recording module receives a host write request and identifies the source SSD based on the logical address range. It then assigns a write closure sequence number, records the logical block address, data cache location, write status, and host confirmation status, and forms a write closure record. The verification module performs sequential verification on write closure records whose status is confirmed by the host to determine the persistence boundary. The interception module intercepts the write closed records that the host has confirmed as confirmed after receiving an abnormal power failure signal from the source SSD, taking the persistent boundary as the interception starting point, and forms the write records to be protected. The encapsulation module performs closed encapsulation based on the write record to be protected, forming closed content; The segmentation module performs closure segmentation based on the data volume of the closure content, closure write parameters, and built-in capacitor parameters, forming local closure content and proxy closure content. The write module performs domain-based writing of local closed content and proxy closed content according to the SSD power-down closed topology table, generating power-down closed records; The module is established, and after power-on, the power-down closing record is read. A recovery mapping between the logical block address and the location of the closed storage area is established. Based on the recovery mapping, the local closed content and the proxy closed content are read, the data is back-transferred, and the address mapping is updated to generate the abnormal power-down protection result.

[0055] In summary, this invention effectively improves protection efficiency by: defining persistent boundaries to lock in the range of critical data that needs protection during abnormal power outages; intelligently dividing the data to be protected into local closed content and proxy closed content based on the real-time energy status of the SSD; and using a cross-device collaboration mechanism to urgently migrate data exceeding the local capacitor's capacity to other healthy SSDs, thereby overcoming the physical limitations of a single device's capacitor energy and mitigating the risk of data loss due to energy depletion to some extent. Finally, by generating power outage closure records, a "data map" is provided for the recovery process, enabling direct location and rapid data migration after power restoration, simplifying the recovery process, shortening business interruption time, and improving the data reliability and service availability of enterprise-level storage solutions in abnormal power outage scenarios.

[0056] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for protecting SSDs from abnormal power loss in enterprise-level storage, characterized in that: include, Collect the identifier, fault domain, logical address range, closed memory area, closed write parameters, emergency transfer relationship and built-in capacitor parameters of each SSD, and establish an SSD power failure closed topology table; After receiving a host write request and identifying the source SSD based on the logical address range, a write closure sequence number is assigned, and the logical block address, data cache location, write status, and host confirmation status are recorded to form a write closure record. Sequential verification is performed on write closure records whose host confirmation status is confirmed to determine the persistence boundary; After receiving an abnormal power failure signal, the source SSD uses the persistent boundary as the starting point for interception and performs boundary interception on the write closure record that the host has confirmed as confirmed, forming the write record to be protected. The record to be protected is closed and encapsulated to form closed content; The closure is segmented based on the data volume of the closure content, the closure write parameters, and the built-in capacitor parameters to form local closure content and proxy closure content. Based on the SSD power-down closure topology table, the local closure content and the proxy closure content are written in separate domains to generate power-down closure records; After power is restored, the power failure closure record is read, a recovery mapping between the logical block address and the location of the closed storage area is established, the local closed content and the proxy closed content are read according to the recovery mapping, the address mapping is updated after data migration, and the abnormal power failure protection result is generated.

2. The SSD abnormal power loss protection method for enterprise-level storage as described in claim 1, characterized in that: The establishment of the SSD power-down closed topology table refers to using the identifier of each SSD as the association basis, binding each item with the corresponding fault domain, logical address range, closed storage area, closed write parameters, emergency transmission relationship and built-in capacitor parameters to form a topology record for each SSD, and summarizing all topology records to form the SSD power-down closed topology table.

3. The SSD abnormal power loss protection method for enterprise-level storage as described in claim 2, characterized in that: The process of forming a closed record is specifically as follows: After receiving a host write request, read the logical block address carried in the host write request, and match the logical block address with the logical address range corresponding to each SSD in the SSD power-down closed topology table. Then, determine the SSD corresponding to the logical address range where the logical block address is located as the source SSD. Assign sequentially increasing and non-repeating write closure sequence numbers according to the order in which host write requests are received, write the write data carried by the host write request to the buffer and record the data buffer position; During the writing process of the source SSD, the write status is recorded according to whether the written data has been persisted, and the host confirmation status is recorded according to whether the host write request has returned a write success confirmation. Using the write closure sequence number as the association basis, the logical block address, data cache location, write status, and host confirmation status are bound one by one to form a write closure record.

4. The SSD abnormal power loss protection method for enterprise-level storage as described in claim 3, characterized in that: The determination of the persistence boundary specifically involves: Read the write closure sequence number, write status, and host confirmation status from the write closure record, filter the write closure records with the host confirmation status "confirmed", and sort the write closure records with the host confirmation status "confirmed" in ascending order of write closure sequence number; Starting from the smallest write closure sequence number, check whether adjacent write closure sequence numbers are consecutive, and check whether the write state corresponding to each write closure sequence number is persisted; Continue sequential verification when adjacent write closure sequence numbers are consecutive and the write status is persisted; stop sequential verification when adjacent write closure sequence numbers are not consecutive and the write status is not persisted; and determine the last write closure sequence number that passes consecutive verification as the persistence boundary.

5. The SSD abnormal power loss protection method for enterprise-level storage as described in claim 4, characterized in that: The process of forming the record to be protected is specifically as follows: Upon receiving an abnormal power failure signal, read the write closure sequence number corresponding to the persistence boundary, and read the write closure sequence number, logical block address, data cache location, write status, and host confirmation status from the write closure record; Filter the write closure records whose host confirmation status is confirmed and whose write closure sequence number is greater than the write closure sequence number corresponding to the persistence boundary, and sort the filtered write closure records in ascending order of write closure sequence number; Using the first write closure sequence number after the persistent boundary as the starting position for truncation, the logical block address, data cache location, write status, and host confirmation status corresponding to each write closure sequence number are truncated sequentially, while maintaining the order between the write closure sequences. The truncated write closure records are then aggregated to form the write record to be protected.

6. The SSD abnormal power loss protection method for enterprise-level storage as described in claim 5, characterized in that: To form a closed content, specifically: Read the write closure sequence number, logical block address, data buffer location, write status, and host confirmation status from the write record to be protected, and sort the write records to be protected in ascending order of write closure sequence number; Read the written data according to the data cache location corresponding to each write closure sequence number, and verify the correspondence between the written data and the logical block address, write status and host confirmation status; Each write closure sequence number is sequentially bound to the corresponding logical block address, write status, host confirmation status, and write data. These are then concatenated continuously according to the order of the write closure sequence numbers. The total number of bytes after concatenation is counted as the data volume of the closure content, thus forming the closure content.

7. The SSD abnormal power loss protection method for enterprise-level storage as described in claim 6, characterized in that: The formation of local closed content and proxy closed content specifically refers to: Read the amount of data in the closed content, and read the closed write parameters and built-in capacitor parameters of the source SSD from the SSD power-down closed topology table; The available capacitor energy that the source SSD can provide during an abnormal power outage is calculated based on the built-in capacitor capacity, current terminal voltage, minimum closing operating voltage and energy conversion efficiency in the built-in capacitor parameters. Calculate the amount of closed write consumption required for the source SSD to complete the unit data write and write verification based on the write time, write energy consumption and write verification time in the closed write parameters; The available capacitor energy is converted into the closed write consumption to obtain the amount of data that the source SSD can complete the closed write. Based on the cumulative number of bytes corresponding to each write closure sequence number in the cumulative closed content from smallest to largest, closed content whose cumulative byte count does not exceed the amount of data that the source SSD can complete the closure write is classified as local closed content, and closed content whose cumulative byte count exceeds the amount of data that the source SSD can complete the closure write is classified as proxy closed content.

8. The SSD abnormal power loss protection method for enterprise-level storage as described in claim 7, characterized in that: Generate a power-down closure record, specifically: Read the SSD identifier, fault domain, closed storage area and emergency transmission relationship corresponding to the source SSD in the SSD power failure closed topology table, and write the local closed content to the closed storage area of ​​the source SSD in ascending order of write closure sequence number; Based on the emergency transfer relationship, select the SSDs that have established an emergency transfer relationship with the source SSD, exclude the SSDs with the same fault domain as the source SSD, arrange the remaining SSDs according to the character order of the SSD identifier, determine the SSD at the top of the list as the proxy SSD, and transfer the proxy closed content to the proxy SSD and write it to the closed storage area of ​​the proxy SSD in ascending order of the write closure sequence number. After completing the domain-specific writing of local and proxy closed content, read the logical block address corresponding to each write closure sequence number in the local and proxy closed content, and read the SSD identifier and closed storage area location that saves the content corresponding to each write closure sequence number. Bind the write closure sequence number, logical block address, SSD identifier and closed storage area location item by item, and summarize them according to the write closure sequence number to form a power failure closure record.

9. The SSD abnormal power loss protection method for enterprise-level storage as described in claim 8, characterized in that: The abnormal power failure protection result is generated as follows: After power is restored, the write closure sequence number, logical block address, SSD identifier, and closed storage area location in the power failure closure record are read in ascending order according to the write closure sequence number. Using the logical block address as an index, the logical block address, SSD identifier, and closed storage area location are bound item by item to establish a recovery mapping between the logical block address and the closed storage area location. Based on the SSD identifier in the recovery map, locate the source SSD that stores the local closed content and the proxy SSD that stores the proxy closed content. Based on the closed storage area location in the recovery map, read the local closed content and the proxy closed content respectively, and verify whether the write closure sequence number and logical block address in the local closed content and the proxy closed content are consistent with the write closure sequence number and logical block address in the power failure closure record. After verification, the source SSD corresponding to each logical block address is identified based on the logical address range in the SSD power-down closure topology table. The write data in the local closure content and the proxy closure content are migrated back to the source SSD in ascending order of write closure sequence number. After completing the data migration, update the address mapping corresponding to the logical block address, summarize the write closure sequence number and logical block address of the completed data migration and address mapping update, and generate the abnormal power failure protection result.

10. An SSD abnormal power loss protection system for enterprise-level storage, based on the SSD abnormal power loss protection method for enterprise-level storage according to any one of claims 1 to 9, characterized in that: include, The acquisition module collects the identifier, fault domain, logical address range, closed memory area, closed write parameters, emergency transfer relationship and built-in capacitor parameters of each SSD, and establishes an SSD power failure closed topology table; The recording module receives a host write request and identifies the source SSD based on the logical address range. It then assigns a write closure sequence number, records the logical block address, data cache location, write status, and host confirmation status, and forms a write closure record. The verification module performs sequential verification on write closure records whose status is confirmed by the host to determine the persistence boundary. The interception module intercepts the write closed records that the host has confirmed as confirmed after receiving an abnormal power failure signal from the source SSD, taking the persistent boundary as the interception starting point, and forms the write records to be protected. The encapsulation module performs closed encapsulation based on the write record to be protected, forming closed content; The segmentation module performs closure segmentation based on the data volume of the closure content, closure write parameters, and built-in capacitor parameters, forming local closure content and proxy closure content. The write module performs domain-based writing of local closed content and proxy closed content according to the SSD power-down closed topology table, generating power-down closed records; The module is established, and after power-on, the power-down closing record is read. A recovery mapping between the logical block address and the location of the closed storage area is established. Based on the recovery mapping, the local closed content and the proxy closed content are read, the data is back-transferred, and the address mapping is updated to generate the abnormal power-down protection result.