CN-122018792-A - Intelligent management method, system and device for flash memory data

CN122018792ACN 122018792 ACN122018792 ACN 122018792ACN-122018792-A

Abstract

The invention discloses an intelligent management method, system and device of flash memory data, wherein the method comprises the steps of obtaining physical parameters and state parameters of each physical block in a flash memory; the method comprises the steps of judging whether each physical block stores a data object, calling a preset bidding strategy to perform bidding operation on physical parameters of the physical blocks to obtain virtual bids of the data objects when the physical blocks store the data objects, calling the preset bidding strategy to perform bidding operation on state parameters of the physical blocks to obtain virtual bids of the physical blocks, matching the virtual bids of the data objects with the virtual bids of the physical blocks based on a preset bilateral auction algorithm to obtain matching results, and performing intelligent management on the data objects according to the matching results, wherein the intelligent management comprises data migration. The invention carries out double-side auction on the virtual bid of each data object and the virtual bid of each physical block so as to match the data object to the corresponding physical block, thereby reducing the access delay of high-value data and improving the utilization rate of storage resources.

Inventors

Chen Kuangxiang
GAO WEI

Assignees

联和存储科技(江苏)有限公司

Dates

Publication Date: 20260512
Application Date: 20260115

Claims (10)

1. The intelligent management method of the flash memory data is characterized by comprising the following steps of: acquiring physical parameters and state parameters of each physical block in the flash memory; judging whether each physical block stores a data object or not; When the physical block stores the data object, calling a preset bidding strategy to perform bidding operation on physical parameters of the physical block to obtain a virtual bid of the data object; invoking a preset quotation strategy to perform quotation operation processing on the state parameters of each physical block to obtain virtual quotations of each physical block; based on a preset double-side auction algorithm, matching the virtual bid of each data object with the virtual bid of each physical block to obtain a matching result; And carrying out intelligent management on each data object according to the matching result, wherein the intelligent management comprises data migration.
2. The intelligent management method of flash memory data according to claim 1, wherein the obtaining physical parameters and status parameters of each physical block in the flash memory comprises: Acquiring a preset acquisition period, and acquiring physical parameters of each physical block according to the preset acquisition period, wherein the physical parameters of the physical block comprise one or more of error rate change gradients, programming time sequence increments and erasing time sequence increments; respectively carrying out weighted operation on physical parameters of each physical block to obtain the health degree of each physical block; Collecting real-time input/output request quantity, bandwidth occupancy rate and concurrent operation number of each physical block in real time; Respectively carrying out load quantization calculation on the real-time input/output request quantity, the bandwidth occupancy rate and the concurrent operation number of each physical block to obtain the load factor of each physical block; And taking the health degree of each physical block and the load factor of each physical block as the state parameters of each physical block.
3. The intelligent management method of flash memory data according to claim 2, wherein said invoking a preset bidding strategy to perform a bidding operation on physical parameters of the physical block to obtain a virtual bid of the data object comprises: Inputting the physical parameters of the physical blocks into a preset value chain model to obtain the expected residual life and the value attenuation risk of the data object; Acquiring a preset life weight of the expected remaining life and a preset risk weight of the value attenuation risk; And calculating the virtual bid of the data object according to the expected remaining life, the preset life weight, the value attenuation risk and the preset risk weight.
4. The intelligent management method of flash data according to claim 3, wherein said calculating a virtual bid for said data object based on said expected remaining life, said preset life weight, said risk of value decay, and said preset risk weight comprises: Bid=α×(1/RTL)+β×|ΔR|; wherein Bid represents a virtual Bid for the data object, α represents the preset lifetime weight, RTL represents the expected remaining lifetime, β represents the preset risk weight, and Δr represents the value decay risk.
5. The intelligent management method of flash memory data according to claim 2, wherein said invoking a preset quotation policy performs a quotation operation on the status parameters of each physical block to obtain a virtual quotation of each physical block, and comprises: For any physical block, acquiring a preset health weight of the health degree and a preset load weight of the load factor; And calculating to obtain the virtual quotation of the physical block according to the health degree, the preset health weight, the load factor and the preset load weight.
6. The intelligent management method of flash memory data according to claim 5, wherein the calculating the virtual quote of the physical block according to the health degree, the preset health weight, the load factor and the preset load weight includes: Ask=γ×(1/H)+δ×L; Wherein Ask represents the virtual quotation of the physical block, gamma represents the preset health weight, H represents the health degree, delta represents the preset load weight, and L represents the load factor.
7. The intelligent management method of flash memory data according to claim 1, wherein the matching the virtual bid of each data object with the virtual bid of each physical block based on the preset double-sided auction algorithm to obtain a matching result comprises: performing ascending order sequencing on the virtual bids of the data objects to obtain a bid sequencing table; taking the first sequence number as the current sequence number, and taking the virtual bid of the data object corresponding to the current sequence number as the current bid; Comparing the current bid with the virtual bid of each physical block to obtain all virtual bids smaller than or equal to the current bid, and taking all virtual bids as candidate bids; Determining the candidate bid with the lowest bid from each candidate bid as a target bid, and matching the physical block corresponding to the target bid with the data object corresponding to the current bid to obtain a sub-matching result; judging whether the current sequence number is the final sequence number in the bid ordering table; if the current sequence number is not the final sequence number in the bid sequencing table, adding one to the current sequence number to update the current sequence number, and re-executing the step of taking the virtual bid of the data object corresponding to the current sequence number as the current bid according to the updated current sequence number; And if the current sequence number is the final sequence number in the bid ordering table, collecting all sub-matching results, and taking the set of all sub-matching results as the matching result.
8. The intelligent management method for flash memory data according to claim 7, wherein said intelligent management of each of said data objects according to said matching result comprises: And for any sub-matching result of the matching results, migrating the data object from a current physical block to the physical block corresponding to the target bid, wherein the current physical block is a physical block storing the data object when calculating the virtual bid of the data object.
9. An intelligent management system for flash memory data, characterized in that the intelligent management system for flash memory data comprises a controller and a flash memory, wherein the controller is configured inside the flash memory, the flash memory comprises a plurality of physical blocks, and the controller can execute the intelligent management method for flash memory data according to any one of claims 1 to 8.
10. An intelligent management device for flash memory data, comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores instructions for execution by the at least one processor to enable the at least one processor to perform the intelligent management method of flash data of any one of claims 1 to 8.

Description

Intelligent management method, system and device for flash memory data Technical Field The present application relates to the field of data storage technologies, and in particular, to an intelligent management method, system, and device for flash memory data. Background With the high-speed development of technologies such as big data, artificial intelligence and edge computing, the market has put stringent requirements on the performance, reliability and service life of storage devices, and flash memory has become a mainstream storage medium by virtue of high read-write rate, low power consumption and the like. In this context, the merits of the flash management technology directly determine the overall performance of the storage device. The core aim of the flash memory management technology is to realize efficient placement of data, intelligent scheduling of cache and accurate execution of garbage collection, so that the service life of a flash memory medium is prolonged while the data access efficiency is improved. At present, the existing flash memory management technology generally performs data allocation and migration operations based on a static policy, for example, writing data into a designated physical block according to a preset priority rule, however, the static policy often adopts a cut-off allocation policy, ignoring life cycle differences (requirements of high-priority data on storage reliability and sensitivity of low-value data on storage cost) of different data objects, and causing a problem that data value and storage resource cost are difficult to match accurately. Therefore, the existing flash memory management technology is easy to cause the problems of excessively high access delay of high-value data and low utilization rate of storage resources. Disclosure of Invention The application mainly aims to provide an intelligent management method, system and device for flash memory data, and aims to solve the problems of excessively high access delay of high-value data and low utilization rate of storage resources. In order to achieve the above object, the present application provides an intelligent management method for flash memory data, which includes: acquiring physical parameters and state parameters of each physical block in the flash memory; judging whether each physical block stores a data object or not; When the physical block stores the data object, calling a preset bidding strategy to perform bidding operation on physical parameters of the physical block to obtain a virtual bid of the data object; invoking a preset quotation strategy to perform quotation operation processing on the state parameters of each physical block to obtain virtual quotations of each physical block; based on a preset double-side auction algorithm, matching the virtual bid of each data object with the virtual bid of each physical block to obtain a matching result; And carrying out intelligent management on each data object according to the matching result, wherein the intelligent management comprises data migration. In some embodiments, the acquiring the physical parameters and the state parameters of each physical block in the flash memory includes: Acquiring a preset acquisition period, and acquiring physical parameters of each physical block according to the preset acquisition period, wherein the physical parameters of the physical block comprise one or more of error rate change gradients, programming time sequence increments and erasing time sequence increments; respectively carrying out weighted operation on physical parameters of each physical block to obtain the health degree of each physical block; Collecting real-time input/output request quantity, bandwidth occupancy rate and concurrent operation number of each physical block in real time; Respectively carrying out load quantization calculation on the real-time input/output request quantity, the bandwidth occupancy rate and the concurrent operation number of each physical block to obtain the load factor of each physical block; And taking the health degree of each physical block and the load factor of each physical block as the state parameters of each physical block. In some embodiments, the invoking the preset bidding strategy to perform bidding operation on the physical parameters of the physical block to obtain the virtual bid of the data object includes: Inputting the physical parameters of the physical blocks into a preset value chain model to obtain the expected residual life and the value attenuation risk of the data object; Acquiring a preset life weight of the expected remaining life and a preset risk weight of the value attenuation risk; And calculating the virtual bid of the data object according to the expected remaining life, the preset life weight, the value attenuation risk and the preset risk weight. In some embodiments, the calculating the virtual bid for the data object based on the expected remaining life, the preset life weight