CN118113214A - Dynamic allocation algorithm for memory cache space - Google Patents

Abstract

The present invention discloses a memory cache space dynamic allocation algorithm, the data storage process includes step S1. monitoring the operation and load of the memory; step S2. introducing demand parameters according to the monitoring results, dynamically adjusting the size of the cache space, and allocating the adjusted cache space to the data to be stored; step S3. storing the data to be stored in the cache space and writing it to the corresponding storage unit. The present invention provides a memory cache space dynamic allocation algorithm, which uses software to coordinate the processing of data cache, so as to better balance load calculations, optimize the allocation of memory cache space, improve the read and write performance and efficiency of the memory, avoid frequent writing to the same storage unit, and reduce the response time and power consumption of the memory.

Description

A dynamic allocation algorithm for memory cache space

Technical Field

The present invention relates to the technical field of data storage, and more specifically, to a memory cache space dynamic allocation algorithm.

Background technique

With the emergence of new technologies such as communications, computing, and artificial intelligence, the amount of global data is growing exponentially. With the establishment of data centers, computing centers, and clouds, the demand for data storage is rising. In terms of hardware used to store data, low-power, ultra-high-density NAND memories, such as solid-state drives (SSDs), are common data storage hardware. A solid-state drive is a hard disk made of a solid-state electronic storage chip array. The solid-state drive includes a control unit and a storage unit (FLASH storage chip or DRAM storage chip), of which the storage unit flash memory (NAND Flash) is the main storage medium of the solid-state drive.

Starting from the host side, the user sends an access request to the SSD at the operating system level, and then the file system converts the access request into an access command to the storage medium that complies with the common agreement between the software and hardware through the driver, mainly including read, write and erase commands, etc., and then sends the command to the SSD control unit through the connection interface. The SSD control unit responds after receiving the command, accesses the storage unit according to a specific timing, and then passes the access result to the host side through the connection interface again. For the SSD control unit level, its input is commands and data, and its output is data and command status. However, the medium life of the storage unit is limited, and each storage unit has a certain number of writes. If the same storage unit is written frequently, bad blocks will be generated, affecting the user experience.

Summary of the invention

In order to overcome the problem of bad blocks caused by frequent reading and writing of the same storage unit, the present invention provides a memory cache space dynamic allocation algorithm.

The technical solution of the present invention is as follows:

A memory cache space dynamic allocation algorithm, the data storage process includes

Step S1. Monitoring the operation and load of the memory;

Step S2. According to the monitoring results, introduce demand parameters, dynamically adjust the size of the cache space, and allocate the adjusted cache space to the data to be stored;

Step S3: Store the data to be stored in the cache space and write it into the corresponding storage unit.

In the above-mentioned memory cache space dynamic allocation algorithm, in step S1, the parameters of the monitoring memory include

(1) Read and write operation frequency, that is, the number of read and write operations performed by the memory per unit time.

(2) Data size, that is, the amount of data stored in the memory, including the amount of data stored and the amount of data to be stored.

(3) Access mode, that is, the mode used by the memory when performing read and write operations, such as random access, continuous access, etc.

(4) Available space, that is, the amount of space remaining in the memory that can be used to store data.

(5) Used space, that is, the size of the space in the memory that has stored data.

(6) The number of concurrent accesses, that is, the number of users or processes accessing the memory at the same time.

(7) Access speed, that is, the speed at which the memory responds to read and write operations.

(8) Error repair capability, that is, the memory's ability to detect and repair errors.

(9) Data consistency, that is, the integrity and consistency of data in the memory.

(10) Access rights, i.e., the access rights setting of the storage device, including user authentication, access rights control, etc.

(11) Data backup and recovery, that is, the storage device’s ability to back up and restore data.

(12) Energy consumption, that is, the energy consumption of the memory, including energy consumption in the running state and energy consumption in the standby state.

The above-mentioned memory cache space dynamic allocation algorithm sets multiple demand parameters, monitors the performance parameters corresponding to the demand parameters, compares the monitoring data with the preset threshold of the demand parameters, and when the detection data triggers the threshold, adjusts the cache space adjustment mode preset according to the demand parameters.

Furthermore, the adjustment of multiple demand parameters is performed simultaneously, or the comparison and calculation are performed in a cyclic rolling manner according to a set sequence, and the comparison and cache space adjustment of all demand parameters are completed one by one.

Furthermore, in step S2, the demand parameters include IOPS demand, bandwidth demand and delay demand, and the IOPS demand, bandwidth demand and delay demand each have a demand threshold.

Furthermore, the cache space adjustment process is

Step A1. Initialize the cache space size to a fixed value;

Step A2: Monitor the storage unit's IOPS, bandwidth, and latency

Step A3. If the IOPS exceeds the threshold A, increase the cache size by Y each time until the IOPS drops below the threshold A or the cache size reaches the upper limit;

Step A4. If the bandwidth usage exceeds threshold B, increase the cache size by Z each time until the bandwidth usage drops below threshold B or the cache reaches the upper limit;

Step A5. If the delay exceeds the threshold C, increase the cache size by W each time until the delay drops below the threshold C or the cache reaches the upper limit;

Step A6. If the above three conditions are not met, but the cache space is less than the lower limit D, then reduce the cache space size, reducing E each time until the cache space is not less than the lower limit D;

Step A7. Repeat steps A2-A6 to dynamically adjust the cache space size.

Furthermore, after the adjustment of the demand parameters is completed, according to the size of the data to be stored, the cache space adjusted based on the demand parameters is used as the initial space, and further dynamic adjustment is performed taking into account the size of the data to be stored and the cache space.

Furthermore, the dynamic adjustment process of taking into account the size of the data to be stored and the cache space includes:

Step B1. Initialize the size and read/write speed of the cache space.

Step B2: Set the maximum expansion threshold and the minimum reduction threshold of the cache space.

Step B3. When the size of the data to be stored is greater than or equal to the maximum expansion threshold, the cache space size is expanded to twice the existing cache space; when the size of the data to be stored is less than the minimum reduction threshold, the cache space size is reduced to 0.5 times the existing cache space; when the size of the data to be stored is less than the maximum expansion threshold and greater than the minimum reduction threshold, decide whether to expand or reduce the cache space size based on the utilization of the cache space.

Furthermore, when allocating the data to be stored to the corresponding cache space, the IOPS, bandwidth and latency information of the storage system are re-monitored to evaluate whether the current cache space meets the demand. If not, the adjustment process of the cache space size corresponding to the demand parameters is returned, and the cache space size is readjusted until it is evaluated that the current cache space meets the demand.

In the above-mentioned memory cache space dynamic allocation algorithm, the control unit tracks and records the number of writes to each storage unit, and prioritizes them according to the number of writes. When writing data to be stored into the storage unit, the data is preferentially written into the storage unit with a high priority.

The beneficial effect of the present invention according to the above scheme is that the present invention provides a dynamic allocation algorithm for memory cache space, which uses software to collaboratively process data cache to better balance load operations, optimize the allocation of memory cache space, improve the read and write performance and efficiency of the memory, avoid frequent writing to the same storage unit, and reduce the response time and power consumption of the memory.

Detailed ways

In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

A memory cache space dynamic allocation algorithm, the data storage process includes

Step S1: Monitor the operation and load of the memory.

Parameters for monitoring memory include

(1) Read and write operation frequency, that is, the number of read and write operations performed by the memory per unit time.

(2) Data size, that is, the amount of data stored in the memory, including the amount of data stored and the amount of data to be stored.

(3) Access mode, that is, the mode used by the memory when performing read and write operations, such as random access, continuous access, etc.

(4) Available space, that is, the amount of space remaining in the memory that can be used to store data.

(5) Used space, that is, the size of the space in the memory that has stored data.

(6) The number of concurrent accesses, that is, the number of users or processes accessing the memory at the same time.

(7) Access speed, that is, the speed at which the memory responds to read and write operations.

(8) Error repair capability, that is, the memory's ability to detect and repair errors.

(9) Data consistency, that is, the integrity and consistency of data in the memory.

(10) Access rights, i.e., the access rights setting of the storage device, including user authentication, access rights control, etc.

(11) Data backup and recovery, that is, the storage device’s ability to back up and restore data.

(12) Energy consumption, that is, the energy consumption of the memory, including energy consumption in the running state and energy consumption in the standby state.

The operation and load conditions of the memory are monitored in step S1 mainly to obtain basic conditions of the memory regarding data transmission and storage, so as to obtain input numerical parameters for allocation calculation for subsequent calculations.

Specifically, the monitoring parameters of the storage device can be set arbitrarily and can be selected according to the needs. The quantity and parameter type can be set arbitrarily.

Step S2: According to the monitoring results, the demand parameters are introduced, the size of the cache space is dynamically adjusted, and the adjusted cache space is allocated to the data to be stored.

Set multiple demand parameters, i.e., the first demand parameter, the second demand parameter, ..., etc., monitor the performance parameters corresponding to the demand parameters, compare the monitoring data with the preset threshold of the demand parameters, and when the detection data triggers the threshold, adjust the cache space adjustment mode preset by the demand parameters. The adjustment of multiple demand parameters can be carried out simultaneously, or the comparison and calculation can be carried out in a circular rolling manner according to the set sequence, and all the comparisons and cache space adjustments of the demand parameters are completed one by one.

In this embodiment, when dynamically adjusting the cache space, QoS requirements (Quality of Service) are introduced. The QoS requirements are mainly to ensure that the storage system can meet the access performance requirements of different applications or workloads, including

(1) IOPS (input/output operations per second) requirements: For applications that require high-frequency random reads and writes, such as databases, higher IOPS is required to ensure the normal operation of the system.

(2) Bandwidth requirements: For applications that require continuous reading and writing of large files, such as video editing and big data analysis, it is necessary to ensure that the storage system has a sufficiently high bandwidth to prevent data transmission from becoming a performance bottleneck.

(3) Latency requirements: For applications with high real-time requirements, such as financial transaction systems, it is necessary to ensure low latency of the storage system to ensure the real-time nature of transactions.

Based on the above QoS requirements, the adjustment process of cache space is as follows:

Step A1. Initialize the cache space size to a fixed value. Part of the space in the memory is divided as the cache space. According to the computing power of the control unit and the required storage space of the storage unit, a comprehensive judgment is made. If the computing power is strong and the memory write space is sufficient, more cache space can be allocated. In this embodiment, it is set to X.

Step A2: Monitor the IOPS, bandwidth and latency of the storage unit. In other embodiments, the information corresponding to the demand parameters is monitored.

Step A3. If the IOPS exceeds the threshold A, increase the cache space size by Y each time until the IOPS drops below the threshold A or the cache space reaches the upper limit.

Step A4. If the bandwidth usage exceeds threshold B, the cache space size is increased by Z each time until the bandwidth usage drops below threshold B or the cache space reaches an upper limit.

Step A5. If the delay exceeds the threshold C, increase the size of the cache space by W each time until the delay drops below the threshold C or the cache space reaches an upper limit.

Step A6. If the above three conditions are not met, but the cache space is less than the lower limit D, reduce the cache space size, reducing E each time until the cache space is not less than the lower limit D.

Step A7. Repeat steps A2-A6 to dynamically adjust the cache space size.

Among them, A, B, and C are respectively the preset threshold values of IOPS, bandwidth, and latency, and D, E, W, X, Y, and Z are respectively the capacity sizes of the adjusted cache space or storage unit. In a specific embodiment, the units of these values are all GB.

After the cache space size is adjusted, it is dynamically adjusted based on the size of the data to be stored, taking into account the size of the data to be stored and the cache space. The process of adjusting the cache space is as follows.

Step B1. Initialize the size of the cache space and the read/write speed. Based on the results of the dynamic adjustment of the cache space in steps A1 to A7, initialize and set the corresponding read/write speed. Usually, the size of the cache space is positively correlated with the read/write speed. The larger the cache space, the faster the read/write speed.

Step B2: Set the maximum expansion threshold and the minimum reduction threshold of the cache space.

Step B3. When the size of the data to be stored is greater than or equal to the maximum expansion threshold, the cache space size is expanded to twice the existing cache space;

When the size of the data to be stored is less than the minimum reduction threshold, the cache space size is reduced to 0.5 times the existing cache space;

When the size of the data to be stored is less than the maximum expansion threshold and greater than the minimum reduction threshold, the cache space size is decided whether to expand or reduce based on the cache space utilization rate. If the cache space utilization rate is greater than 80%, the cache space size is expanded to twice the existing cache space. If the cache space utilization rate is less than 50%, the cache space size is reduced to 0.5 times the existing cache space.

After the adjustment is completed, the data to be stored needs to be allocated to the corresponding cache space. At this time, the IOPS, bandwidth and latency information of the storage system are re-monitored to evaluate whether the current cache space meets the demand. If not, return to step A3-step A6 to readjust the cache space size until the current cache space is evaluated to meet the demand. The adjustment ends with the readjustment of the relationship between the cache space and the demand parameters as the adjustment end benchmark, in order to ensure that data storage is prioritized to meet the demand. In other embodiments, the dynamic adjustment between the size of the data to be stored and the cache space can be set as the end point to prioritize the storage write speed of the data to be stored.

Step S3: Store the data to be stored in the cache space and write it into the corresponding storage unit.

The control unit tracks and records the number of writes to each storage unit, and prioritizes the numbers according to the number of writes. When writing data to be stored into a storage unit, the storage unit with a high priority (fewer writes) is given priority.

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

Claims (10)

1. A memory cache space dynamic allocation algorithm, characterized in that the data storage process includes Step S1. Monitoring the operation and load of the memory; Step S2. According to the monitoring results, introduce demand parameters, dynamically adjust the size of the cache space, and allocate the adjusted cache space to the data to be stored; Step S3: Store the data to be stored in the cache space and write it into the corresponding storage unit. 2. According to the memory cache space dynamic allocation algorithm described in claim 1, it is characterized in that in step S1, the parameters of the monitoring memory include (1) Read and write operation frequency; (2) Data size; (3) Access mode; (4) Available space; (5) Used space; (6) Number of concurrent accesses; (7) Access speed; (8) Error repair capability; (9) Data consistency; (10) Access rights; (11) Data backup and recovery; (12)Energy consumption. 3. According to a memory cache space dynamic allocation algorithm described in claim 1, it is characterized in that multiple demand parameters are set, the performance parameters corresponding to the demand parameters are monitored, and the monitoring data are compared with the threshold of the preset demand parameters. When the detection data triggers the threshold, the cache space adjustment mode preset according to the demand parameters is adjusted. 4. According to the memory cache space dynamic allocation algorithm described in claim 3, it is characterized in that the adjustment of multiple demand parameters is carried out simultaneously, or, the comparison and calculation are carried out in a cyclic rolling manner according to the set sequence and the comparison and cache space adjustment of all demand parameters are completed one by one. 5. According to a memory cache space dynamic allocation algorithm described in claim 3, it is characterized in that in step S2, the demand parameters include IOPS demand, bandwidth demand and delay demand, and the IOPS demand, bandwidth demand and delay demand each have a demand threshold. 6. According to the memory cache space dynamic allocation algorithm described in claim 5, it is characterized in that the cache space adjustment process is as follows: Step A1. Initialize the cache space size to a fixed value; Step A2: Monitor the storage unit's IOPS, bandwidth, and latency Step A3. If the IOPS exceeds the threshold A, increase the cache size by Y each time until the IOPS drops below the threshold A or the cache size reaches the upper limit; Step A4. If the bandwidth usage exceeds threshold B, increase the cache size by Z each time until the bandwidth usage drops below threshold B or the cache reaches the upper limit; Step A5. If the delay exceeds the threshold C, increase the cache size by W each time until the delay drops below the threshold C or the cache reaches the upper limit; Step A6. If the above three conditions are not met, but the cache space is less than the lower limit D, then reduce the cache space size, reducing E each time until the cache space is not less than the lower limit D; Step A7. Repeat steps A2-A6 to dynamically adjust the cache space size. 7. According to a memory cache space dynamic allocation algorithm described in claim 3, it is characterized in that after completing the adjustment of the demand parameters, based on the size of the data to be stored, the cache space adjusted based on the demand parameters is used as the initial space, and further dynamic adjustment is performed taking into account the size of the data to be stored and the cache space. 8. A memory cache space dynamic allocation algorithm according to claim 7, characterized in that the dynamic adjustment process taking into account the size of the data to be stored and the cache space includes: Step B1. Initialize the size and read/write speed of the cache space; Step B2. Set the maximum expansion threshold and the minimum reduction threshold of the cache space; Step B3. When the size of the data to be stored is greater than or equal to the maximum expansion threshold, the cache space size is expanded to twice the existing cache space; when the size of the data to be stored is less than the minimum reduction threshold, the cache space size is reduced to 0.5 times the existing cache space; when the size of the data to be stored is less than the maximum expansion threshold and greater than the minimum reduction threshold, decide whether to expand or reduce the cache space size based on the utilization of the cache space. 9. According to a memory cache space dynamic allocation algorithm described in claim 7, it is characterized in that when allocating the data to be stored to the corresponding cache space, the IOPS, bandwidth and latency information of the storage system are re-monitored to evaluate whether the current cache space meets the demand. If not, the adjustment process of the cache space size corresponding to the demand parameters is returned to readjust the cache space size until it is evaluated that the current cache space meets the demand. 10. According to a memory cache space dynamic allocation algorithm described in claim 1, it is characterized in that the control unit tracks and records the number of write times of each storage unit, and prioritizes them according to the number of write times. When writing data to be stored into the storage unit, it is written first into the storage unit with a high priority.