Storage resource allocation method, electronic equipment, storage medium and program product

Abstract

The invention discloses a storage resource allocation method, electronic equipment, a storage medium and a program product, and relates to the technical field of computers, and the method comprises the following steps: receiving a storage resource allocation request sent by any service subsystem, starting a counter corresponding to the storage resource allocation request, when a target impedance value corresponding to a target occupancy index value of the service subsystem is not greater than a count value of a counter, processing of the storage resource allocation request is started, and the larger the occupancy index value is, the smaller the impedance value is; and if the central resource management module is in a non-preemptive mode and the residual allocated storage resource quantity does not meet the storage resource allocation request, refusing to allocate the storage resources, and updating the central resource management module to be in a preemptive mode of which the preemptive mark is the identifier of the service subsystem, so as to ensure that the central resource management module is exclusively occupied by the service subsystem. The technical problem that the system performance and stability are reduced due to strategies in related technologies is solved, and the technical effect of improving the system performance and stability is achieved.

Description

Technical Field

[0001] The present application relates to the field of computer technologies, and in particular, to a storage resource allocation method, an electronic device, a storage medium, and a program product. Background Art

[0002] In a System on Chip (SOC) with high performance requirements, in order to improve system efficiency, multiple service subsystems need to execute concurrently and use on-chip storage to break through the storage performance bottleneck. On-chip storage has the advantages of high bandwidth and low latency, but limited capacity. Each service subsystem allocates and reclaims the on-chip storage resources configured by itself through its own dynamic storage management module. The on-chip storage resource configuration of each service subsystem is usually static and cannot adapt to the changes in the pressure of the service subsystem.

[0003] In order to improve the adaptability to pressure changes, a central resource management module is introduced to support resource sharing among multiple service subsystems and support over-allocation of resources for multiple service subsystems to improve the throughput of the service subsystem. However, this strategy may cause some tasks to be unable to obtain resources for a long time in the full-load state of the system, reducing the system performance and stability. Summary of the Invention

[0004] The present application provides a storage resource allocation method, an electronic device, a storage medium, and a program product to at least solve the problem that the strategy in the related technology reduces the system performance and stability.

[0005] The present application provides a storage resource allocation method, including: Receiving a storage resource allocation request sent by any service subsystem and delaying the processing of the storage resource allocation request; Starting a counter corresponding to the storage resource allocation request; Based on the remaining allocated storage resource amount of the central resource management module and the remaining requested storage resource amount of the service subsystem, determining the target occupancy index value of the service subsystem; Based on a preset correspondence table between the occupancy index value and the impedance value, determining the target impedance value corresponding to the target occupancy index value, where the larger the occupancy index value, the smaller the impedance value; If the target impedance value is not greater than the count value of the counter, starting to process the storage resource allocation request; If the central resource management module is in the non-preemption mode and the remaining allocated storage resource amount of the central resource management module does not meet the storage resource allocation request, updating the central resource management module to the preemption mode marked with the identifier of the service subsystem and returning a rejection allocation message to the service subsystem.

[0006] The present application also provides an electronic device, including: a memory for storing a computer program; a processor for implementing the steps of any of the above storage resource allocation methods when executing the computer program.

[0007] The present application also provides a computer-readable storage medium storing a computer program, wherein the computer program implements the steps of any of the above storage resource allocation methods when executed by a processor.

[0008] The present application also provides a computer program product including a computer program, and the computer program implements the steps of any of the above storage resource allocation methods when executed by a processor.

[0009] Through the present application, since the storage resource allocation request sent by any service subsystem is received and the processing of the storage resource allocation request is delayed; a counter corresponding to the storage resource allocation request is started; based on the remaining allocated storage resources of the central resource management module and the remaining requested storage resources of the service subsystem, the target occupancy index value of the service subsystem is determined; based on a preset correspondence table between the occupancy index value and the impedance value, the target impedance value corresponding to the target occupancy index value is determined, and the larger the occupancy index value, the smaller the impedance value; if the target impedance value is not greater than the count value of the counter, the processing of the storage resource allocation request is started, that is, the larger the target occupancy index value, the shorter the waiting time for the storage resource allocation request to be processed; if the central resource management module is in a non-preemptive mode and the remaining allocated storage resources of the central resource management module do not satisfy the storage resource allocation request, the central resource management module is updated to a preemptive mode marked with the identifier of the service subsystem, and a rejection allocation message is returned to the service subsystem to ensure that the central resource management module is exclusive to the service subsystem before the processing of the storage resource allocation request is completed. Through the above content, the situation where some tasks cannot obtain resources for a long time under the full-load state of the system is avoided. Therefore, the technical problem that the strategy in the related art reduces the system performance and stability can be solved, and the technical effect of improving the system performance and stability is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] To more clearly illustrate the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0011] Figure 1 It is a schematic diagram of the architecture of a system-on-chip in the related art; Figure 2 It is a schematic diagram of the structure of a storage resource allocation system provided by an embodiment of the present application; Figure 3 Flow diagram of a storage resource allocation method provided by an embodiment of the present application; Figure 4 Schematic diagram of the principle of the allocation impedance mechanism provided by an embodiment of the present application; Figure 5 Schematic diagram of the process where four service subsystems concurrently send storage resource allocation requests to the central resource management module, and the central resource management module processes them, provided by an embodiment of the present application; Figure 6 Flow diagram of another storage resource allocation method provided by an embodiment of the present application; Figure 7 Corresponding schematic diagram of different occupancy metric values and different latency processing durations provided by an embodiment of the present application; Figure 8 Interaction schematic diagram of the interface arbitration logic module and the core arbitration logic module corresponding to the service subsystem provided by an embodiment of the present application; Figure 9 Schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0012] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present application.

[0013] It should be noted that in the description of the present application, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such a process, method, article or device. The terms "first", "second", etc. in the present application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0014] To enable those skilled in the art of the present technology to better understand the solution of the present application, the present application will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0015] In a system-on-chip with high-performance requirements, multiple relatively independent business subsystems usually need to run in parallel to improve the overall system efficiency. To address the issue of storage performance bottlenecks, such systems often adopt on-chip storage to provide storage services for data and control information. Among them, on-chip storage can be Static Random-Access Memory (SRAM for short).

[0016] Compared with the traditional off-chip Double Data Rate (DDR) storage method, on-chip storage has significant advantages: the access bandwidth is dozens of times higher, and the read and write latency is only in the nanosecond range. That is to say, on-chip storage has the advantages of high bandwidth and low latency. However, due to cost limitations, its storage capacity is much lower than that of DDR.

[0017] Inside each business subsystem, an independent dynamic storage management module is usually equipped to be responsible for the allocation and recycling of the on-chip storage resources of that business subsystem. At the same time, the resource configuration of each business subsystem is often statically fixed, lacking the ability of flexible adjustment and unable to respond flexibly to changes in system pressure.

[0018] To improve the adaptability of business subsystems to pressure changes, a central resource management module is introduced in related technologies to support resource sharing among multiple business subsystems. And there is a resource limit for the maximum allowable application of each business subsystem inside the central resource management module, and at the same time, it supports over-configuration of resources for multiple business subsystems, that is, the sum of the resource limits for the maximum allowable application of all business subsystems can exceed the total amount of on-chip storage resources actually managed by the central resource management module.

[0019] In this way, when some business subsystems are in an idle state, other business subsystems that need more on-chip storage resources can dynamically obtain additional on-chip storage resources, thereby enhancing their own business processing capabilities and throughput.

[0020] Figure 1 It is a schematic diagram of the architecture of the system-on-chip in related technologies. As Figure 1As shown in the figure, the system on a chip includes multiple service subsystems, namely service subsystem 1, service subsystem 2, ..., service subsystem n, and also includes a central resource management module and on-chip storage. Each service subsystem is responsible for executing specific service functions, and each service subsystem is equipped with a dynamic storage management module, which is used to manage the on-chip storage resources related to the corresponding service subsystem, such as allocating and reclaiming on-chip storage resources, etc., to ensure that the storage requirements of the service subsystem are met. The central resource management module is used to monitor and manage the usage of on-chip storage resources of the entire system on a chip, and allocate resources to each service subsystem and its dynamic storage management module to ensure the efficient operation of the overall system. The on-chip storage is divided into multiple slices, such as slice 1, slice 2, ..., slice m, which are used to store information such as data and program codes during the operation of the system. The service subsystem accesses the corresponding slice of the on-chip storage through the system bus according to the requirements, and performs data read or write operations.

[0021] It can be understood that the on-chip storage resources exist in the form of slices. The service subsystems, the central resource management module, and the on-chip storage are connected through the system bus. The system bus is a channel for transmitting data, addresses, and control signals. The service subsystem can interact with the on-chip storage through the system bus to read the required data or write the processing results. At the same time, the central resource management module also monitors and controls the usage of resources of the service subsystem and the on-chip storage through the system bus.

[0022] However, in the system full-load state, the management strategy of the on-chip storage resources by the central resource management module in the related technology has a situation where a certain task is stuck and unable to execute for a long time because it cannot obtain the required on-chip storage resources. The reasons for this situation may be uneven resource allocation, some resources being occupied by other tasks for a long time, or the resource scheduling algorithm of the system being unfair. The resource demand is in a starvation waiting state for a long time, the processing delay of the task continues to increase, and even enters the exception throwing processing flow. Although this does not cause the system to deadlock, it greatly restricts the release of the overall system performance, generates unnecessary performance fluctuations, and reduces the system performance and stability.

[0023] To solve the above problems, an embodiment of the present application provides a storage resource allocation method, which includes: receiving a storage resource allocation request sent by any business subsystem, and delaying the processing of the storage resource allocation request; starting a counter corresponding to the storage resource allocation request; determining a target occupancy index value of the business subsystem based on the remaining allocated storage resource amount of the central resource management module and the remaining applied storage resource amount of the business subsystem; determining a target impedance value corresponding to the target occupancy index value based on a preset correspondence table between the occupancy index value and the impedance value, where the larger the occupancy index value, the smaller the impedance value; if the target impedance value is not greater than the count value of the counter, start processing the storage resource allocation request; if the central resource management module is in a non-preemptive mode and the remaining allocated storage resource amount of the central resource management module does not satisfy the storage resource allocation request, update the central resource management module to a preemptive mode marked with the identifier of the business subsystem, and return a rejection allocation message to the business subsystem. The method provided by the above solution, when the target occupancy index value of the business subsystem is larger, sets the waiting time for processing its storage resource allocation request to be shorter; when the central resource management module is in a non-preemptive mode and the remaining allocated storage resource amount of the central resource management module does not satisfy the storage resource allocation request, updates the central resource management module to a preemptive mode marked with the identifier of the business subsystem, ensuring that the central resource management module is exclusive to the business subsystem before the storage resource allocation request is processed, avoiding the situation that some tasks cannot obtain resources for a long time in the system full-load state, realizing more efficient and balanced resource allocation among multiple business subystems, avoiding performance bottlenecks caused by uneven resource allocation, and improving the overall performance and stability of the system.

[0024] Combined with the specific application environment architecture or specific hardware architecture on which the execution of the storage resource allocation method depends, the specific application environment architecture or specific hardware architecture is described herein.

[0025] The storage resource allocation method, electronic device, storage medium and program product provided by the embodiments of the present application are applicable to the allocation of on-chip storage resources. As Figure 2 shown, it is a schematic structural diagram of the storage resource allocation system based on the embodiments of the present application, mainly including a central resource management module and multiple business subsystems. Among them, for the scenario where multiple business subsystems share on-chip storage resources through the central resource management module, in order to more efficiently and evenly allocate resources among the business subsystems, the central resource management module includes a core arbitration logic module and multiple interface arbitration logic modules, and one interface arbitration logic module is used to connect to one business subsystem.

[0026] An exclusive resource preemption mechanism is introduced into the core arbitration logic module to avoid the problem of resource starvation, that is, a business subsystem fails to obtain resources after a long period of repeated retries, resulting in task timeouts. Among them, the exclusive preemption mechanism is that when a business subsystem applies for certain storage resources from the core arbitration logic module, but the remaining storage resources of the core arbitration logic module cannot meet the request, the core arbitration logic module enters the preemption mode, and other business subsystems cannot obtain any storage resource allocation before the request of the business subsystem is met or the request times out. It can be understood that the initial storage resources in the core arbitration logic module are all on-chip storage resources of the on-chip system.

[0027] It is understandable that, when only the exclusive preemption mechanism is applied, the business subsystem with a larger average value of resource requests or more intensive resource requests will obtain more storage resources, especially when the High Water Mark (HWM) of resource requests of each business subsystem, i.e., the maximum allowed resource limit, is over-allocated, i.e., the sum of the High Water Marks of resource requests of each business subsystem is greater than or equal to the total allocatable storage resources of the system on chip.

[0028] In order to make the resource allocation between various business subsystems more balanced, the allocation impedance mechanism is introduced in the interface arbitration logic module to balance the resource allocation between multiple business subsystems and block the allocation requests of business subsystems that have occupied more resources. The core idea of ​​the allocation impedance mechanism is to flexibly suppress new requests from business subsystems with higher storage resource usage, so that other business subsystems with relatively less storage resource usage can obtain resources with higher priority; at the same time, when storage resources are close to exhaustion, the allocation is delayed to buy more time for storage resources to be recovered.

[0029] An embodiment of the present application provides a storage resource allocation method, which is applied to a central resource management module. The central resource management module is used to manage all on-chip storage resources of the system on chip, including allocating and reclaiming on-chip storage resources. Figure 3 A flowchart of a storage resource allocation method provided in an embodiment of the present application, such as Figure 3 As shown, the process includes the following steps: Step S301: receiving a storage resource allocation request sent by any business subsystem, and delaying processing of the storage resource allocation request.

[0030] Referring to the foregoing, the service subsystem sends a storage resource allocation request to the corresponding interface arbitration logic module in the central resource management module. After receiving the storage resource allocation request sent by its corresponding service subsystem, any interface arbitration logic module first masks the request from the core arbitration logic module, that is, delays processing the storage resource allocation request. It can be understood that the storage resource is an on-chip storage resource.

[0031] Step S302, start a counter corresponding to the storage resource allocation request.

[0032] While masking the storage resource allocation request, immediately start a counter corresponding to the storage resource allocation request, so that the counter corresponding to the storage resource allocation request starts counting incrementally from 0. This counter is an impedance counter.

[0033] It can be understood that the clock frequency at which the counter operates determines how often the counter counts. Exemplarily, if the counter operates at a clock frequency of 500 megahertz (MHz), then the counter increments by 1 every 2 nanoseconds.

[0034] Step S303, based on the remaining allocated storage resources of the central resource management module and the remaining requested storage resources of this service subsystem, determine the target occupancy metric value of this service subsystem.

[0035] Among them, the remaining allocated storage resources of the central resource management module are the remaining allocable storage resources of the current central resource management module, and the remaining requested storage resources of this service subsystem are the remaining requestable storage resources of this service subsystem. Among them, the remaining requested storage resources of this service subsystem are determined by the difference between the storage resource volume water level line of this service subsystem, that is, the highest water level line for resource requests, and the occupied storage resources of this service subsystem.

[0036] Step S304, based on a pre-set correspondence table between the occupancy metric value and the impedance value, determine the target impedance value corresponding to the target occupancy metric value. The larger the occupancy metric value, the smaller the impedance value.

[0037] It should be noted that the pre-set correspondence table between the occupancy metric value and the impedance value is configured during the initialization of the system-on-chip.

[0038] Step S305, if the target impedance value is not greater than the count value of the counter, start processing the storage resource allocation request.

[0039] It can be understood that the larger the occupancy index value is, the smaller the impedance value is. Then, the larger the target occupancy index value is, the shorter the time required for the target impedance value not to exceed the count value of the counter. That is to say, the larger the target occupancy index value of the service subsystem is, the shorter the delay in processing the storage resource allocation request of the service subsystem is, so as to balance the resource allocation among service subsystems.

[0040] Furthermore, since the storage resources are dynamically allocated and recycled, the target occupancy index value of the service subsystem is constantly changing, and the corresponding target impedance value is also constantly changing. When the target impedance value does not exceed the count value of the counter, the interface arbitration logic module cancels the shielding of the current storage resource allocation request, so that it enters the core arbitration logic module for final resource allocation arbitration.

[0041] It can be seen from steps S301 to S305 that the present application realizes the allocation impedance mechanism by simply adding an impedance counter and a corresponding relationship table of a preset occupancy index value and an impedance value, with relatively low complexity. Figure 4 It is a schematic diagram of the principle of the allocation impedance mechanism provided by the embodiment of the present application. As Figure 4 shown, the interface arbitration logic module introduces the allocation impedance mechanism. Every time the interface arbitration logic module receives a storage resource allocation request sent by the corresponding service subsystem, it triggers the allocation impedance mechanism, first shields the storage resource allocation request from the core arbitration logic module, and immediately starts the counter corresponding to the storage resource allocation request, so that the counter starts to increment from 0, that is, the count value++. Periodically calculate the occupancy index value of the service subsystem, and then look up the corresponding impedance value from the impedance value look-up table with the periodically calculated occupancy index value of the service subsystem as the input, that is, periodically update the corresponding look-up table value, and look up the corresponding impedance value from the impedance value look-up table based on the updated look-up table value. Among them, the impedance value look-up table is a corresponding relationship table of a preset occupancy index value and an impedance value. If the impedance value found from the impedance value look-up table is less than or equal to the count value, immediately release the shielding of the storage resource allocation request so that it enters the core arbitration logic module for final resource allocation arbitration.

[0042] It can be understood that all interface arbitration logic modules share the corresponding relationship table of the preset occupancy index value and the impedance value.

[0043] Step S306, if the central resource management module is in the non-preemption mode and the remaining allocated storage resources of the central resource management module do not meet the storage resource allocation request, then update the central resource management module to the preemption mode marked with the service subsystem identifier, and return a rejection allocation message to the service subsystem.

[0044] Among them, the central resource management module is in a preemption mode marked with the business subsystem identifier, indicating that the central resource management module is monopolized by this business subsystem before the resource allocation for the current storage resource allocation request of this business subsystem is completed, and will not process the storage resource allocation requests of other business subsystems. Each business subsystem corresponds to a unique business subsystem identifier.

[0045] This step S306 is implemented by the core arbitration logic module. That is to say, the remaining allocated storage resource amount of the central resource management module is the remaining allocated storage resource amount of the core arbitration logic module. Updating the central resource management module to be in a preemption mode marked with the business subsystem identifier means updating the core arbitration logic module to be in a preemption mode marked with the business subsystem identifier. It can be understood that returning a rejection allocation message to this business subsystem means rejecting the allocation of storage resources to this business subsystem.

[0046] It should be noted that there are three possible reasons for a storage resource allocation request of a business subsystem to be rejected: 1. Reaching the highest water mark HWM of the resource application preset for this business subsystem: The upper limit of the storage resources currently allowed for each business subsystem preset by the system-on-chip is called the highest water mark of the resource application. If the sum of the storage resources currently allocated to this business subsystem, that is, the amount of storage resources occupied by this business subsystem, and the amount of storage resources applied for by the new storage resource allocation request of this business subsystem is greater than the HWM, the interface arbitration logic module will directly reject this storage resource allocation request.

[0047] 2. Resource pool exhaustion (Resource Run Out, abbreviated as: RRO): When a certain business subsystem passes the judgment of whether it reaches the HWM of this business subsystem in the interface arbitration logic module, that is, it does not reach the HWM of this business subsystem, it will further apply to the central arbitration logic module for the corresponding storage resources. However, if the remaining storage resources of the core arbitration logic module at this time cannot meet this application and the exclusive preemption flag is "empty" or "marked as occupied by the current business subsystem", the core arbitration logic module will directly reject this storage resource allocation request. Among them, the exclusive preemption flag being "empty" indicates that the current core arbitration logic module is in a non-preemption mode, and the exclusive preemption flag being "marked as occupied by the current business subsystem" indicates that the current core arbitration logic module is in a preemption mode marked with the business subsystem identifier.

[0048] 3. Resource pool preemption (Occupied, abbreviated as: OCP): When a certain business subsystem passes the judgment on whether it reaches the HWM of the business subsystem in the interface arbitration logic module, it will further apply to the core arbitration logic module for the corresponding storage resources. However, if the exclusive preemption flag is "marked as occupied by other business subsystems", the core arbitration logic will directly reject this storage resource allocation request. The exclusive preemption flag being "marked as occupied by other business subsystems" indicates that the current core arbitration logic module is in the preemption mode marked by other business subsystems.

[0049] It can be understood that the rejection allocation information returned to the business subsystem in step S306 is the rejection allocation information of the resource pool exhaustion type.

[0050] Figure 5 This is a schematic flowchart of the process where four business subsystems provided in the embodiments of the present application concurrently send storage resource allocation requests to the central resource management module and the central resource management module processes them. As Figure 5As shown, the central resource management module manages the on-chip storage resources by means of a to-be-allocated page table and a pointer list. The business subsystem 1 sends a request 1.1 to the central resource management module through the dynamic storage management module. After the interface arbitration logic module determines that the business subsystem 1 has reached the preset resource application highest watermark, it returns a rejection allocation message to the business subsystem. The type of the rejection allocation message is the HWM type, that is, it has reached the preset resource application highest watermark. After waiting for a preset time, the business subsystem 1 initiates a request 1.1 again. The business subsystem 1 has not reached the preset resource application highest watermark, but at this time the exclusive preemption mark of the core arbitration logic module is 2, that is, the current core arbitration logic module is in the preemption mode with the preemption mark as the business subsystem 2. It returns a rejection allocation message to the business subsystem. The type of the rejection allocation message is the OCP type. After waiting for a preset time, business subsystem 1 initiates request 1.1 again. Business subsystem 1 has not reached the preset resource application maximum watermark, but at this time the exclusive preemption mark of the core arbitration logic module is 2, that is, the current core arbitration logic module is in the preemption mode marked as business subsystem 2, and a rejection allocation information is returned to the business subsystem, and the type of the rejection allocation information is OCP type. After waiting for a preset time, business subsystem 1 initiates request 1.1 again. Business subsystem 1 has not reached the preset resource application maximum watermark, the core arbitration logic module is in the non-preemption mode, and the remaining allocated storage resources of the core arbitration logic module do not meet request 1.1, then the core arbitration logic module is updated to be in the preemption mode marked as business subsystem 1, and a rejection allocation information is returned to the business subsystem, and the type of the rejection allocation information is RRO type. After waiting for a preset time, business subsystem 1 initiates request 1.1 again. Business subsystem 1 has not reached the preset resource application maximum watermark. The core arbitration logic module is in the preemptive mode marked as business subsystem 1, but the remaining allocated storage resources of the core arbitration logic module do not meet request 1.1, and a rejection allocation message is returned to the business subsystem. The type of the rejection allocation message is RRO type. After waiting for a preset time, business subsystem 1 initiates request 1.1 again. Business subsystem 1 has not reached the preset resource application maximum watermark. The core arbitration logic module is in the preemptive mode marked as business subsystem 1, and the remaining allocated storage resources of the core arbitration logic module meet request 1.1. The storage resources corresponding to request 1.1 are allocated to business subsystem 1, and the processing of request 1.1 ends.

[0051] It can be understood that the processing flows of Request 1.2 of Business Subsystem 1, Request 2.1 of Business Subsystem 2, Request 2.2 of Business Subsystem 2, Request 2.3 of Business Subsystem 2, Request 3.1 of Business Subsystem 3, Request 4.1 of Business Subsystem 4, and Request 4.2 of Business Subsystem 4 are similar to the processing flow of Request 1.1 of Business Subsystem 1, and will not be elaborated here. Among them, as Figure 5 shown, the requests are all storage resource allocation requests. If a request is rejected, the request is a dashed line, and the reason for rejection is noted below the request signal. If the request obtains storage resource allocation, the request information is a solid line.

[0052] The storage resource allocation method provided by the embodiment of the present application realizes more efficient and balanced resource allocation among multiple business subsystems by introducing an exclusive preemption mechanism and an allocation impedance mechanism, avoids performance bottlenecks caused by uneven resource allocation, and avoids the problem that a certain business subsystem causes task timeouts due to being unable to obtain the required resources for a long time. Thereby, it reduces the occurrence of task processing delays and exception throwing processing flows, increases the stability of the system, and reduces performance fluctuations caused by uneven resource allocation or tasks waiting for resources for a long time, thereby improving the overall performance of the system.

[0053] Moreover, the storage resource allocation method provided by the embodiment of the present application allows dynamic adjustment of resource allocation when the pressure of the business subsystem changes, so that the resource configuration is no longer statically fixed, but can flexibly respond to the changes in the pressure of the business subsystem. When other business subsystems are in an idle state, a certain business subsystem can obtain more storage resources, thereby enhancing its business throughput.

[0054] Through the allocation impedance mechanism, new requests of business subsystems with a relatively high resource occupancy are flexibly suppressed, so that business subsystems with less resource occupancy can obtain resources with a higher priority, increasing the flexibility of resource allocation. When resources are approaching exhaustion, the allocation impedance mechanism can delay allocation and strive for more time for resources to be recycled and restored, optimizing the resource recycling process.

[0055] An embodiment of the present application provides a storage resource allocation method, which is applied to a central resource management module, Figure 6 is a flowchart of the storage resource allocation method provided by the embodiment of the present application. As Figure 6 shown, the process includes the following steps: Step S601, receive a storage resource allocation request sent by any business subsystem, and delay processing the storage resource allocation request. For details, please refer to Figure 3 Step S301 of the embodiment shown, which will not be elaborated here.

[0056] Step S602: Start a counter corresponding to the storage resource allocation request. For details, please refer to Figure 3 Step S302 of the embodiment shown in

[0057] Step S603: Based on the remaining allocated storage resources of the central resource management module and the remaining requested storage resources of this business subsystem, determine the target occupancy index value of this business subsystem.

[0058] Specifically, the above Step S603 includes: Step S6031: Obtain the working mode of the central resource management module.

[0059] Among them, the central resource management module has two working modes, namely the business subsystem priority mode and the central resource management module priority mode. The working mode of the central resource management module is pre-set by technicians according to the actual application scenario. Among them, in the business subsystem priority mode, even if the core arbitration logic module still has a large amount of storage resources to be allocated, as long as the number of storage resources already allocated to this business subsystem is close to reaching the HWM, there will be an obvious resource application lag effect from the interface arbitration logic module. And in the central resource management module priority mode, that is, the core arbitration logic module priority mode, only when the core arbitration logic module only has a small amount of storage resources to be allocated, the interface arbitration logic modules corresponding to each business subsystem will start to have a lag effect on the resource applications of the business subsystems, and the intensity depends on the current actual storage resource occupancy of each business subsystem. The more the storage resource occupancy, the greater the intensity, that is, the longer the waiting time for delayed processing.

[0060] Step S6032: Based on the working mode, the remaining allocated storage resources of the central resource management module, and the remaining requested storage resources of this business subsystem, determine the target occupancy index value of this business subsystem.

[0061] Step S604: Based on the pre-set correspondence table between the occupancy index value and the impedance value, determine the target impedance value corresponding to the target occupancy index value. The larger the occupancy index value, the smaller the impedance value. For details, please refer to Figure 3 Step S304 of the embodiment shown in

[0062] Step S605: If the target impedance value is not greater than the count value of the counter, start processing the storage resource allocation request. For details, please refer to Figure 3 Step S305 of the embodiment shown in

[0063] Step S606: If the central resource management module is in the non-preemptive mode and the remaining allocated storage resource amount of the central resource management module does not meet the storage resource allocation request, then update the central resource management module to the preemptive mode marked with the service subsystem identifier, and return a rejection allocation message to the service subsystem. For details, please refer to Figure 3 Step S306 of the embodiment shown, which will not be elaborated here.

[0064] The storage resource allocation method provided by the embodiment of the present application can flexibly adjust the resource allocation strategy according to the actual application scenario by clarifying the working mode of the central resource management module, which helps to achieve more efficient resource utilization in different application scenarios.

[0065] In some alternative embodiments, the above step S6032 includes: Step a1: In the case where the working mode is the service subsystem priority mode, determine the target occupancy index value of the service subsystem through the first formula. The first formula is: .

[0066] Step a2: In the case where the working mode is the central resource management module priority mode, determine the target occupancy index value of the service subsystem through the second formula. The second formula is: .

[0067] Wherein, is the target occupancy index value of the service subsystem in the service subsystem priority mode, is the target occupancy index value of the service subsystem in the central resource management module priority mode, is the weight coefficient, is the remaining allocated storage resource amount of the central resource management module, is the remaining applied storage resource amount of the service subsystem, is the normalization coefficient.

[0068] It should be noted that the weight coefficient: , wherein, The value of is preset by technicians.

[0069] The normalization coefficient: = , , wherein, The value of is preset by technicians.

[0070] In some alternative embodiments, the above storage resource allocation method further includes: Step b1, when the target occupancy index value is greater than the preset saturation value, update the target occupancy index value to the preset saturation value.

[0071] It can be understood that when the target occupancy index value calculated by the above first formula or second formula is greater than the preset saturation value, update the target occupancy index value to the preset saturation value, and determine the target impedance value based on the updated target occupancy index value. That is to say, the maximum value of the target occupancy index value is the preset saturation value.

[0072] When the target occupancy index value calculated by the above first formula or second formula is not greater than the preset saturation value, determine the target impedance value based on the target occupancy index value calculated by the above first formula or second formula.

[0073] The storage resource allocation method provided by the embodiments of the present application ensures the accuracy of the obtained target impedance value by setting the maximum value of the target occupancy index value to the preset saturation value when the target occupancy index value is greater than the preset saturation value.

[0074] In some optional embodiments, the above storage resource allocation method further includes: Step c1, determine the preset saturation value through the third formula, and the third formula is:

[0075] Wherein, is the preset saturation value, is the preset saturation parameter. The value of b can be 3, 4, or 5, which is specifically preset by those skilled in the art.

[0076] In some optional embodiments, the above storage resource allocation method further includes: Step d1, obtain different occupancy index values.

[0077] Step d2, for any occupancy index value, determine the impedance value corresponding to the occupancy index value through the fourth formula to obtain the impedance values corresponding to different occupancy index values.

[0078] Step d3, based on different occupancy index values and the impedance values corresponding to different occupancy index values, establish a corresponding relationship table between the preset occupancy index values and impedance values.

[0079] Wherein, the fourth formula is:

[0080] Wherein, is the maximum delay processing duration, is the occupancy index value, is the impedance adjustment coefficient, For the corresponding impedance value.

[0081] It can be understood that during the initialization of the system-on-chip, and are set by the technical staff. That is to say, and are given. At this time, different values correspond to different impedance values. This impedance value is a count value, but essentially represents the delay processing duration of the storage resource allocation request. Because each time the counter counts, there is a fixed time interval, based on the count value, the delay processing duration can be obtained. Figure 7 This is a corresponding schematic diagram of different occupancy index values and different delay processing durations provided by the embodiments of the present application. As Figure 7 shown, in , the counter works at a clock frequency of 500 MHz. When changes, different occupancy index values correspond to different delay processing durations. Among them,

[0082] The smaller the value, the longer the delay processing duration.

[0082] In some alternative embodiments, before starting the counter corresponding to the storage resource allocation request, the above storage resource allocation method further includes: Step e1, based on the amount of storage resources to be applied for in the storage resource allocation request and the amount of occupied storage resources of the service subsystem, determine the current total amount of storage resources applied for by the service subsystem.

[0083] Step e2, if the current total amount of storage resources applied for is not greater than the storage resource amount water level line of the service subsystem, then execute the step of starting the counter corresponding to the storage resource allocation request.

[0084] The storage resource allocation method provided by the embodiments of the present application ensures more fair and reasonable resource allocation among different service subsystems by restricting the maximum amount of storage resources that each service subsystem can apply for.

[0085] In some alternative embodiments, the above storage resource allocation method further includes: Step f1, if the current total amount of storage resources applied for is greater than the storage resource amount water level line of the service subsystem, then return a rejection allocation message to the service subsystem to reject the storage resource allocation request of the service subsystem.

[0086] The storage resource allocation method provided by the embodiments of the present application can avoid subsequent processing of invalid storage resource allocation requests by directly returning a rejection allocation message when the current total amount of storage resources applied for is greater than the storage resource amount water level line of the service subsystem.

[0087] In some optional implementations, the above storage resource allocation method further includes: Step g1: If the central resource management module is in the preemption mode where the preemption identifier is not the identifier of the business subsystem, a rejection allocation message is returned to the business subsystem to reject the storage resource allocation request of the business subsystem.

[0088] The storage resource allocation method provided in the embodiment of the present application can avoid conflicts and confusion caused by multiple business subsystems preempting resources at the same time, and ensure that resource allocation is executed according to the established priority rules, by rejecting the storage resource allocation request of the current business subsystem if the central resource management module is in the preemption mode of other business subsystems.

[0089] In some optional implementations, the above storage resource allocation method further includes: Step h1: if the central resource management module is in the preemption mode whose preemption mark is the identifier of the business subsystem, it is determined whether the remaining allocated storage resources of the central resource management module meet the storage resource allocation request.

[0090] Step h2: if the remaining allocated storage resources of the central resource management module meet the storage resource allocation request, the storage resources corresponding to the storage resource allocation request are allocated to the business subsystem, and the central resource management module is updated to be in non-preemptive mode.

[0091] After receiving the allocation rejection information, the business subsystem waits for a first preset time period and initiates a storage resource allocation request again until the storage resource amount corresponding to the storage resource allocation request is obtained or the request duration of the storage resource allocation request exceeds a preset duration threshold.

[0092] It is understandable that after receiving the allocation rejection information, any business subsystem waits for the first preset time period and re-initiates the storage resource allocation request that was rejected until the storage resource amount corresponding to the storage resource allocation request is obtained or the storage resource allocation request times out. The storage resource allocation request times out when the request duration of the storage resource allocation request exceeds the preset duration threshold. The first preset time period and the preset duration threshold are set by the technical staff and are not specifically limited here.

[0093] The business subsystem executes the next storage resource allocation request when the storage resource amount corresponding to a storage resource allocation request is obtained or when a storage resource allocation request times out.

[0094] In some optional implementations, the above storage resource allocation method further includes: Step i1, if the central resource management module is in non-preemptive mode and the remaining allocated storage resource amount of the central resource management module meets the storage resource allocation request, allocate the storage resource amount corresponding to the storage resource allocation request to the service subsystem, and update the remaining allocated storage resource amount of the central resource management module.

[0095] The storage resource allocation method provided by the embodiments of the present application can directly complete resource allocation if the central resource management module has sufficient remaining allocated storage resource amount and is currently in non-preemptive mode, which can significantly shorten the time of resource allocation and improve the response speed of the system.

[0096] In some optional embodiments, the above storage resource allocation method further includes: Step j1, at every second preset time period, update the target occupancy index value of the service subsystem based on the remaining allocated storage resource amount of the central resource management module and the remaining applied storage resource amount of the service subsystem.

[0097] It can be understood that since the storage resources are dynamically allocated and recycled, the remaining allocated storage resource amount of the central resource management module and the remaining applied storage resource amount of the service subsystem are constantly changing, so the target occupancy index value of the service subsystem is constantly changing. Therefore, it is necessary to periodically update the target occupancy index value of the service subsystem to determine the target impedance value based on the latest target occupancy index value and compare the target impedance value with the count value of the counter. The second preset time period is set by technicians and is not specifically limited here.

[0098] The storage resource allocation method provided by the embodiments of the present application accurately determines the target impedance value by periodically updating the target occupancy index value, ensuring the accuracy of the timing for processing the storage resource allocation request.

[0099] In some optional embodiments, the above storage resource allocation method further includes: Step k1, if the target impedance value is greater than the count value of the counter, keep the storage resource allocation request in delayed processing.

[0100] The storage resource allocation method provided by the embodiments of the present application ensures the balance of resource allocation by keeping the storage resource allocation request in delayed processing when the target impedance value is greater than the count value of the counter.

[0101] In some optional embodiments, the above storage resource allocation method further includes: Step L1, if the central resource management module is in the preemption mode marked with the business subsystem identifier, and the remaining allocated storage resource amount of the central resource management module does not meet the storage resource allocation request, then the central resource management module continues to remain in the preemption mode marked with the business subsystem identifier and returns a rejection allocation message to the business subsystem.

[0102] In some alternative embodiments, the above storage resource allocation method further includes: Step M1, monitor the processing process of any received storage resource allocation request in real time. When the request duration of any storage resource allocation request exceeds the preset duration and the storage resource amount corresponding to the storage resource allocation request has not been allocated to the business subsystem corresponding to the storage resource allocation request, an alarm is issued.

[0103] The storage resource allocation method provided by the embodiments of the present application can quickly discover and locate possible problems in the resource allocation process through real-time monitoring and alarm mechanisms, thereby avoiding service interruption caused by long-term unallocated storage resources, enhancing the reliability of the system, and ensuring the normal operation of the business subsystem.

[0104] Figure 8 It is a schematic interaction diagram between the interface arbitration logic module and the core arbitration logic module corresponding to the business subsystem provided by the embodiments of the present application. As Figure 8 shown, the interaction between the interface arbitration logic module corresponding to business subsystem 1 and the core arbitration logic module is described. The interaction between the interface arbitration logic module corresponding to other business subsystems and the core arbitration logic module is similar to the interaction between the interface arbitration logic module corresponding to business subsystem 1 and will not be elaborated here. The interaction between the interface arbitration logic module corresponding to business subsystem 1 and the core arbitration logic module includes the following processes: Inside the interface arbitration logic module corresponding to business subsystem 1, the interface arbitration logic module is initially in the standby state. After receiving the storage resource allocation request, it performs an HWM check on business subsystem 1. Performing an HWM check on business subsystem 1 means determining whether the current total requested storage resource amount of business subsystem 1 is greater than the storage resource amount water level line of this business subsystem.

[0105] If the check passes, that is, the current total requested storage resource amount of business subsystem 1 is not greater than the storage resource amount water level line of this business subsystem, then the allocation impedance mechanism is run. If the check fails, the interface arbitration logic module jumps to the rejection state to reject the storage resource allocation request of business subsystem 1.

[0106] When the target impedance value is not greater than the count value of the counter, the interface arbitration logic module forwards the storage resource allocation request to the core arbitration logic module.

[0107] The core arbitration logic module is initially in a standby state. After receiving a storage resource allocation request, if the core arbitration logic module is in a non-preemptive mode and the remaining allocable storage resources can satisfy the storage resource allocation request, the core arbitration logic module enters a pre-allocation state, makes an allocation response to the interface arbitration logic module, and completes the allocation of the storage resource amount corresponding to the storage resource allocation request. If the core arbitration logic module is in a non-preemptive mode but the remaining allocable storage resources do not satisfy the storage resource allocation request, it enters a preemptive mode, marks that the core arbitration logic module is exclusively occupied by Business Subsystem 1, and makes a rejection response of the RRO type to the interface arbitration logic module.

[0108] If the core arbitration logic module is in a preemptive mode and the core arbitration logic module is marked as exclusively occupied by other business subsystems, it makes a rejection response of the OCP type to the interface arbitration logic module.

[0109] If the core arbitration logic module is in a preemptive mode, the core arbitration logic module is marked as exclusively occupied by Business Subsystem 1, and the remaining allocable storage resources can satisfy the storage resource allocation request, the core arbitration logic module enters a pre-allocation state and makes an allocation response to the interface arbitration logic module.

[0110] If the core arbitration logic module is in a preemptive mode, the core arbitration logic module is marked as exclusively occupied by Business Subsystem 1, and the remaining allocable storage resources do not satisfy the storage resource allocation request, the core arbitration logic module makes a rejection response of the RRO type to the interface arbitration logic module.

[0111] If the exclusive occupation of the core arbitration logic module by the business subsystem times out, that is, the request duration of the storage resource allocation request exceeds the preset duration threshold, the core arbitration logic module is in a non-preemptive mode.

[0112] After receiving the allocation response or rejection response from the core arbitration logic module, the interface arbitration logic module makes a state transition accordingly, that is, jumps to the allocation state or the rejection state, and responds to Business Subsystem 1.

[0113] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0114] The embodiments of the present application also provide an electronic device, as Figure 9 shown, including a processor 901 and a memory 902. A computer program is stored in the memory 902, and the processor 901 is set to run the computer program to execute the steps in any one of the above embodiments of the storage resource allocation method.

[0115] Embodiments of the present application also provide a computer-readable storage medium storing a computer program, where the computer program is configured to execute the steps in any of the above-described embodiments of the storage resource allocation method when running.

[0116] In an exemplary embodiment, the above computer-readable storage medium may include, but is not limited to, various media capable of storing a computer program, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disc.

[0117] Embodiments of the present application also provide a computer program product. The above computer program product includes a computer program, and when the computer program is executed by a processor, it implements the steps in any of the above-described embodiments of the storage resource allocation method.

[0118] Embodiments of the present application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, it implements the steps in any of the above-described embodiments of the storage resource allocation method.

[0119] Those skilled in the art can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Skilled professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0120] The above has introduced in detail a storage resource allocation method, an electronic device, a storage medium, and a program product provided by the present application. Specific examples are used herein to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the protection scope of the claims of the present application.

Claims

1. A storage resource allocation method, characterized in that, Including: Receiving a storage resource allocation request sent by any service subsystem and delaying the processing of the storage resource allocation request; Starting a counter corresponding to the storage resource allocation request; Based on the remaining allocated storage resources of the central resource management module and the remaining applied storage resources of this service subsystem, determining the target occupancy metric value of this service subsystem; Based on a preset correspondence table between the occupancy metric value and the impedance value, determining the target impedance value corresponding to the target occupancy metric value, where the larger the occupancy metric value, the smaller the impedance value; If the target impedance value is not greater than the count value of the counter, then start processing the storage resource allocation request; If the central resource management module is in a non-preemptive mode and the remaining allocated storage resources of the central resource management module do not satisfy the storage resource allocation request, then update the central resource management module to a preemptive mode marked with the identifier of this service subsystem and return a rejection allocation message to this service subsystem.

2. The method according to claim 1, characterized in that, The determining the target occupancy metric value of this service subsystem based on the remaining allocated storage resources of the central resource management module and the remaining applied storage resources of this service subsystem includes: Obtaining the working mode of the central resource management module; Based on the working mode, the remaining allocated storage resources of the central resource management module, and the remaining applied storage resources of this service subsystem, determining the target occupancy metric value of this service subsystem.

3. The method according to claim 2, wherein The determining the target occupancy metric value of this service subsystem based on the working mode, the remaining allocated storage resources of the central resource management module, and the remaining applied storage resources of this service subsystem includes: In the case where the working mode is the service subsystem priority mode, determining the target occupancy metric value of this service subsystem through a first formula, and the first formula is: In the case where the working mode is the central resource management module priority mode, determining the target occupancy metric value of this service subsystem through a second formula, and the second formula is: Among them, is the target occupancy index value of the service subsystem in the service subsystem priority mode of the working mode, is the target occupancy index value of the service subsystem in the central resource management module priority mode of the working mode, is the weight coefficient, is the remaining allocated storage resource amount of the central resource management module, is the remaining applied storage resource amount of the service subsystem, is the normalization coefficient.

4. The method according to claim 3, wherein The method further includes: In the case where the target occupancy metric value is greater than a preset saturation value, updating the target occupancy metric value to the preset saturation value.

5. The method according to claim 4, wherein The method further includes: Determining the preset saturation value through a third formula, and the third formula is: Among them, is a preset saturation value, is a preset saturation parameter.

6. The method according to claim 1, wherein The method further includes: Obtaining different occupancy metric values; For any occupancy metric value, determining the impedance value corresponding to this occupancy metric value through a fourth formula to obtain the impedance values corresponding to different occupancy metric values; Based on different occupancy metric values and the impedance values corresponding to different occupancy metric values, establishing the preset correspondence table between the occupancy metric value and the impedance value; Wherein, the fourth formula is: Among them, is the maximum delay processing duration, is the occupancy index value, is the impedance adjustment coefficient, is the corresponding impedance value.

7. The method according to claim 1, characterized in that, Before starting the counter corresponding to the storage resource allocation request, the method further includes: Based on the storage resource amount to be applied in the storage resource allocation request and the occupied storage resources of this service subsystem, determining the current total applied storage resource amount of this service subsystem; If the current total applied storage resource amount is not greater than the storage resource amount water level line of this service subsystem, then execute the step of starting the counter corresponding to the storage resource allocation request.

8. The method according to claim 7, characterized in that The method further includes: If the current total applied storage resource amount is greater than the storage resource amount watermark of the business subsystem, a rejection allocation message is returned to the business subsystem to reject the storage resource allocation request of the business subsystem.

9. The method according to claim 1, characterized in that, The method further comprises: If the central resource management module is in the preemption mode where the preemption identifier is not the identifier of the business subsystem, it returns allocation rejection information to the business subsystem to reject the storage resource allocation request of the business subsystem.

10. The method according to claim 1, characterized in that, The method further comprises: If the central resource management module is in a preemption mode whose preemption mark is the business subsystem identifier, determining whether the remaining allocated storage resource amount of the central resource management module satisfies the storage resource allocation request; If the remaining allocated storage resources of the central resource management module meet the storage resource allocation request, the storage resources corresponding to the storage resource allocation request are allocated to the business subsystem, and the central resource management module is updated to be in non-preemptive mode; After receiving the allocation rejection information, the business subsystem waits for a first preset time period and initiates the storage resource allocation request again until the storage resource amount corresponding to the storage resource allocation request is obtained or the request duration of the storage resource allocation request exceeds a preset duration threshold.

11. The method according to claim 1, characterized in that, The method further comprises: If the central resource management module is in non-preemptive mode and the remaining allocated storage resources of the central resource management module meet the storage resource allocation request, the storage resources corresponding to the storage resource allocation request are allocated to the business subsystem, and the remaining allocated storage resources of the central resource management module are updated.

12. The method according to claim 1, characterized in that, The method further comprises: At intervals of a second preset time period, the target occupancy index value of the business subsystem is updated based on the remaining allocated storage resources of the central resource management module and the remaining applied storage resources of the business subsystem.

13. An electronic device, characterized in that, include: Memory for storing computer programs; A processor, configured to implement the steps of the storage resource allocation method according to any one of claims 1 to 12 when executing the computer program.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of the storage resource allocation method according to any one of claims 1 to 12.

15. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, the steps of the storage resource allocation method according to any one of claims 1 to 12 are implemented.

Images