Control method, apparatus and device based on enhanced sgpio

Abstract

The application provides a control method, device and equipment based on enhanced SGPIO, belongs to the field of data processing, is applied to a backboard controller of a server, the server further comprises a plurality of intelligent backboards, the backboard controller is in communication connection with the plurality of intelligent backboards, and the method comprises the following steps: acquiring first multi-modal data from the plurality of intelligent backboards through an enhanced serial general-purpose input / output (SGPIO); generating a control instruction according to the first multi-modal data, and generating second multi-modal data corresponding to the control instruction; and sending the second multi-modal data to the plurality of intelligent backboards through the enhanced SGPIO. The centralized, data-driven management and control capability for the plurality of intelligent backboards can be effectively improved, and support is provided for stable and efficient operation of a server storage system.

Description

Technical Field

[0001] This application belongs to the field of data processing, and specifically relates to a control method, apparatus and device based on enhanced SGPIO. Background Technology

[0002] Serial General Purpose Input / Output (SGPIO) technology is primarily used in Serial Attached Small Computer System Interface (SAS) or Serial Advanced Technology Attachment (SATA) storage architectures. SGPIO, through the SFF-8485 specification, coordinates with high-speed serial interfaces for multi-hard drive status monitoring and LED indicator control. However, traditional SGPIO has limited functionality, supporting only basic status signal transmission (such as hard drive activity / fault indication) and unable to handle complex control commands. Furthermore, the power supply and bandwidth allocation functions of backplane devices are fixed, preventing dynamic adjustments and timely maintenance based on actual conditions.

[0003] As data centers continue to increase their requirements for server storage density, operating efficiency, and reliability, the technical shortcomings of traditional SGPIO have become a key bottleneck restricting the performance optimization of storage systems. There is an urgent need for an enhanced SGPIO control solution that can expand functional boundaries and improve intelligence. Summary of the Invention

[0004] This application provides a control method, apparatus, and device based on enhanced SGPIO to effectively improve the centralized, data-driven management and control capabilities of server storage systems for multiple intelligent backplanes, thereby providing support for the stable and efficient operation of server storage systems.

[0005] This application provides a control method based on enhanced SGPIO, applied to a backplane controller of a server. The server further includes multiple intelligent backplanes, and the backplane controller is communicatively connected to the multiple intelligent backplanes. The method includes:

[0006] First multimodal data from the multiple smart backplanes is acquired via the enhanced serial general purpose input / output (SGPIO).

[0007] Control commands are generated based on the first multimodal data, and second multimodal data corresponding to the control commands are generated.

[0008] The second multimodal data is sent to the plurality of smart backplanes via the enhanced SGPIO.

[0009] According to the enhanced SGPIO-based control method provided in this application, the control instruction includes a bandwidth allocation instruction. The step of generating the control instruction based on the first multimodal data and generating the second multimodal data based on the control instruction includes: determining the priority and bandwidth requirement of each smart backplane based on the first multimodal data; sorting each smart backplane according to the priority to obtain multiple priority levels; allocating bandwidth to each smart backplane sequentially according to the multiple priority levels and the bandwidth requirement to obtain a bandwidth allocation value; and generating the second multimodal data of each smart backplane based on the bandwidth allocation value.

[0010] In one possible embodiment, allocating bandwidth to each smart backplane sequentially according to the plurality of priority levels and the bandwidth requirements includes: performing the following operations for each smart backplane in each priority level until bandwidth is allocated to each smart backplane or until all bandwidth has been allocated: determining the current total bandwidth requirement based on the bandwidth requirement of each smart backplane in the current priority level; obtaining the current total allocated bandwidth; determining the current remaining allocable bandwidth based on the current total bandwidth requirement and the current total allocated bandwidth; determining a target bandwidth ratio to the current total bandwidth requirement, wherein the target bandwidth is the minimum bandwidth value between the current total bandwidth requirement and the remaining allocable bandwidth; and allocating bandwidth to each smart backplane in the current priority level based on the ratio and the bandwidth requirement of each smart backplane in the current priority level.

[0011] In one possible embodiment, the control instructions include storage strategy optimization instructions. The backplane controller embeds a preset AI model. The step of generating control instructions based on the first multimodal data and generating second multimodal data based on the control instructions includes: obtaining the temperature, wear level, and input / output operations per second (IOPS) of each smart backplane based on the first multimodal data; inputting the temperature and wear level into the preset AI model to obtain a fault probability prediction value output by the preset AI model; determining the storage strategy optimization instructions for each smart backplane based on the fault probability prediction value and the IOPS; and generating second multimodal data based on the storage strategy optimization instructions. The storage strategy optimization instructions include independent disk redundant array (RAID) configuration instructions, data tiering instructions, and hibernation control instructions for each smart backplane.

[0012] In one possible embodiment, the control command includes an energy optimization command. The step of generating the control command based on the first multimodal data and generating the second multimodal data based on the control command includes: acquiring the temperature, IOPS, and load of each smart backplane based on the first multimodal data; determining the power supply mode of each smart backplane based on the temperature, the IOPS, and the load; generating the energy optimization command based on the power supply mode; and generating the second multimodal data based on the energy optimization command.

[0013] In one possible embodiment, the method further includes: obtaining the base fan speed of each smart backplane and the maximum disk power of the storage device connected to each smart backplane based on the first multimodal data; determining the target fan speed of each smart backplane based on the temperature and the base fan speed; determining the target disk power of each smart backplane based on the maximum disk power and the load; the step of generating an energy optimization instruction based on the power supply mode and generating second multimodal data based on the energy optimization instruction includes: generating an energy optimization instruction based on the power supply mode, the target fan speed and the target disk power, and generating second multimodal data based on the energy optimization instruction.

[0014] In one possible embodiment, the data frame structure of the multimodal data corresponding to the enhanced SGPIO includes at least a channel ID field, a data category field, a status data field, an AI instruction field, and a bandwidth allocation field; the channel ID field is used to identify each smart backplane, the data category field is associated with the basic data of the smart backplane, the status data field is associated with energy optimization instructions, the AI ​​instruction field is associated with storage strategy optimization instructions, and the bandwidth allocation field is associated with bandwidth allocation instructions.

[0015] This application also provides a control device based on enhanced SGPIO, applied to a backplane controller of a server, wherein the server further includes multiple intelligent backplanes, and the backplane controller is communicatively connected to the multiple intelligent backplanes. The device includes:

[0016] An acquisition unit is used to acquire first multimodal data from the plurality of smart backplanes via enhanced SGPIO;

[0017] The generation unit is configured to generate control instructions based on the first multimodal data, and generate second multimodal data corresponding to the control instructions;

[0018] The transmitting unit is used to transmit the second multimodal data to the plurality of smart backplanes via the enhanced SGPIO.

[0019] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement any of the above-described control methods based on enhanced SGPIO.

[0020] This application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements any of the above-described control methods based on enhanced SGPIO.

[0021] This application also provides a computer program product, including a computer program that, when executed by a processor, implements any of the above-described enhanced SGPIO-based control methods.

[0022] The control method, apparatus, and device based on enhanced SGPIO provided in this application are applied to a server backplane controller. The server further includes multiple intelligent backplanes, and the backplane controller is communicatively connected to the multiple intelligent backplanes. The backplane controller first acquires first multimodal data from the multiple intelligent backplanes through enhanced SGPIO, then generates control commands based on the first multimodal data, and generates second multimodal data corresponding to the control commands. Finally, it sends the second multimodal data to the multiple intelligent backplanes through enhanced SGPIO. Thus, this solution, through enhanced SGPIO communication between the server backplane controller and multiple intelligent backplanes, realizes the acquisition of multimodal data, generation and distribution of control commands, effectively improving the centralized, data-driven management and control capabilities of multiple intelligent backplanes, and providing support for the stable and efficient operation of the server storage system. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of the composition of a server provided in this application.

[0025] Figure 2 This is a flowchart illustrating a control method based on enhanced SGPIO provided in this application.

[0026] Figure 3 This is one of the functional unit block diagrams of a control device based on enhanced SGPIO provided in this application.

[0027] Figure 4This is the second functional unit block diagram of a control device based on enhanced SGPIO provided in this application.

[0028] Figure 5 This is a schematic diagram of the structure of an electronic device provided in this application. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0030] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0031] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0032] As data centers continue to increase their requirements for server storage density, operating efficiency, and reliability, the technical shortcomings of traditional SGPIO have become a key bottleneck restricting the performance optimization of storage systems.

[0033] To address the aforementioned problems, this application provides a control method, apparatus, and device based on enhanced SGPIO. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0034] Please see Figure 1The server 100 includes a backplane controller 101 and multiple smart backplanes 102. The backplane controller 101 is communicatively connected to each of the smart backplanes 102 via the SGPIO protocol. The backplane controller 101 serves as the core of the server's storage system management, acquiring multimodal data uploaded by each smart backplane 102 in real time via the SGPIO protocol, such as the operating status and load of connected storage devices. Based on this data, it generates corresponding control commands and then sends these commands to the corresponding smart backplanes 102 via the SGPIO protocol, achieving centralized scheduling and management of the multiple smart backplanes 102 and their connected storage devices. The smart backplanes 102 are functional components in the server that connect storage devices, such as hard disk drives (HDDs), solid-state drives (SSDs), and Non-Volatile Memory Express (NVMe) devices, to the backplane controller 101, acting as a bridge for data interaction and command execution between the two. On the one hand, it collects multi-dimensional operating data of the storage devices it is connected to and uploads it to the backplane controller 101 via the SGPIO protocol. On the other hand, it receives control commands issued by the backplane controller 101, parses them, and adjusts the operating parameters of the connected storage devices to ensure that the storage devices operate stably according to the management and control requirements.

[0035] Please see Figure 2 The enhanced SGPIO-based control method is applied to the backplane controller 101 mentioned above, and the enhanced SGPIO-based control method specifically includes the following steps.

[0036] S201, first multimodal data from the plurality of smart backplanes is acquired via enhanced SGPIO.

[0037] Among them, enhanced SGPIO, compared with traditional SGPIO, breaks through the limitation of only transmitting basic status signals, and has the ability to support the parallel transmission of multi-dimensional and multi-type data, making it a key carrier for realizing multimodal data interaction. Multiple intelligent backplanes correspond to multiple sets of storage device-related components that need to be managed in the server. Each intelligent backplane collects the operating data of the storage devices it is connected to, thus forming the first multimodal data. This first multimodal data covers the basic operating status of the storage devices, such as their presence and activity status; and hardware parameters, such as temperature, load, and other multi-dimensional information, not just single-type data. It provides comprehensive and real-time raw data support for the backplane controller to subsequently analyze the overall status of the storage devices and generate targeted control commands, essentially establishing the foundation for the controller's status perception of multiple intelligent backplanes and their connected storage devices.

[0038] S202, generate control commands based on the first multimodal data, and generate second multimodal data corresponding to the control commands.

[0039] The first multimodal data serves as a multi-dimensional status reference for multiple smart backplanes and connected storage devices. The backplane controller analyzes and judges this raw data to identify the current operational needs of the storage system, such as whether device operating status needs adjustment or resource configuration needs optimization, and then generates targeted control commands. These commands directly target specific adjustments to the smart backplanes and associated storage devices, rather than being generic commands without a basis. Since enhanced SGPIO requires specific data structures for signal transmission, control commands cannot be directly issued. Therefore, the generated control commands need to be encapsulated into a multimodal data format that conforms to the transmission requirements of enhanced SGPIO, namely the second multimodal data. Essentially, this provides a transmission format adapted to the communication interface for the control commands, ensuring that the commands can be accurately transmitted to the target smart backplane through enhanced SGPIO. It is important to note that the second multimodal data has a different function from the first multimodal data. For example, the former carries the data for issuing control commands, while the latter is the data collected reflecting device status. Both adapt to the multi-data type transmission capabilities of enhanced SGPIO, supporting the formation of a closed-loop control system.

[0040] S203, the second multimodal data is sent to the plurality of smart backplanes via the enhanced SGPIO.

[0041] The second multimodal data, being an instruction carrier encapsulated to match the enhanced SGPIO transmission specification, ensures data integrity and real-time performance during transmission. This avoids instruction delays or loss caused by format incompatibility and limited bandwidth in traditional interfaces, providing reliable support for control signal transmission. Multiple smart backplanes correspond to different storage device groups within the server, and the control actions required by each smart backplane may differ. Therefore, the second multimodal data matched to each smart backplane must be selectively distributed to the corresponding target, rather than being broadcast uniformly. This ensures that each smart backplane only receives instruction carriers relevant to its own control needs, preventing erroneous instruction execution or resource waste. The control instructions generated by the backplane controller are delivered to the smart backplanes, providing a direct basis for the subsequent intelligent backplane instruction parsing and adjustment of the connected storage device's operating status.

[0042] As can be seen, in this embodiment, the enhanced SGPIO communication between the server backplane controller and multiple smart backplanes enables the acquisition of multimodal data, the generation and distribution of control commands, and effectively improves the centralized and data-driven management and control capabilities of multiple smart backplanes, providing support for the stable and efficient operation of the server storage system.

[0043] In one possible embodiment, the control command includes a bandwidth allocation command. The step of generating the control command based on the first multimodal data and generating the second multimodal data based on the control command includes: determining the priority and bandwidth requirement of each smart backplane based on the first multimodal data; sorting each smart backplane according to the priority to obtain multiple priority levels; allocating bandwidth to each smart backplane sequentially according to the multiple priority levels and the bandwidth requirement to obtain a bandwidth allocation value; and generating the second multimodal data of each smart backplane based on the bandwidth allocation value.

[0044] The first multimodal data includes key information such as load intensity and service type of the storage devices connected to each smart backplane, including core database services and general archiving services. This data can be used as a basis for priority determination. For example, smart backplanes carrying high-real-time services and with consistently high loads will be assigned higher priority. Based on the actual data transmission volume of the devices and the service's rate requirements, the specific bandwidth requirements of each smart backplane can be quantified. By sorting the smart backplanes into different levels (such as high, medium, and low priority levels), the aim is to ensure that high-priority backplanes associated with core services can obtain bandwidth resources first, avoiding performance impacts on core services due to resource contention, and providing a logical framework for subsequent orderly allocation.

[0045] The allocation process processes each level sequentially from highest to lowest priority. First, it satisfies the bandwidth requirements of all backplanes within the highest priority level, or, when total bandwidth is limited, prioritizes allocating their required core resources before processing lower priority levels. The final bandwidth allocation value is a balance between the requirements of each backplane and the total system bandwidth capacity, avoiding resource waste and preventing insufficient resources for critical backplanes. Since enhanced SGPIO requires signal transmission in a specific data format, bandwidth allocation values ​​cannot be directly issued. Therefore, the bandwidth allocation value corresponding to each smart backplane needs to be encapsulated as second multimodal data conforming to the enhanced SGPIO transmission specification. This ensures that allocation instructions accurately match the target smart backplane and maintain data integrity during transmission, providing an executable signal carrier for subsequent adjustments to bandwidth parameters by the smart backplanes according to the allocation value.

[0046] As can be seen, in this embodiment, by accurately extracting key information such as load intensity and service type of the intelligent backplane associated storage devices based on the first multimodal data, the priority and actual bandwidth requirements of each intelligent backplane are determined. Then, an ordered allocation logic is established through priority hierarchical sorting, and the total bandwidth and the requirements of each backplane are balanced hierarchically to complete the bandwidth allocation. The allocation result is encapsulated into second multimodal data conforming to the enhanced SGPIO transmission specification. This significantly improves the accuracy of bandwidth resource allocation in the server storage system, avoiding resource waste in the traditional fixed allocation mode and preventing performance bottlenecks caused by insufficient resource preemption for core services. At the same time, by prioritizing the bandwidth requirements of intelligent backplanes associated with high real-time and high-load core services, the stability and response efficiency of core service operation are effectively enhanced, avoiding problems such as stuttering and latency caused by insufficient bandwidth. Furthermore, the second multimodal data encapsulation design, which conforms to transmission specifications, ensures that bandwidth allocation commands can be accurately matched to the target intelligent backplane and transmitted completely, guaranteeing the effective execution of allocation results. Ultimately, this enhances the flexibility and efficiency of server storage system resource scheduling, better adapts to the differences in bandwidth requirements under different business scenarios, and provides strong support for the stable operation and business support capabilities of the server storage system.

[0047] In one possible embodiment, allocating bandwidth to each smart backplane sequentially according to the plurality of priority levels and the bandwidth requirements includes: performing the following operations for each smart backplane in each priority level until bandwidth is allocated to each smart backplane or until all bandwidth has been allocated: determining the current total bandwidth requirement based on the bandwidth requirement of each smart backplane in the current priority level; obtaining the current total allocated bandwidth; determining the current remaining allocable bandwidth based on the current total bandwidth requirement and the current total allocated bandwidth; determining a target bandwidth ratio to the current total bandwidth requirement, wherein the target bandwidth is the minimum bandwidth value between the current total bandwidth requirement and the remaining allocable bandwidth; and allocating bandwidth to each smart backplane in the current priority level based on the ratio and the bandwidth requirement of each smart backplane in the current priority level.

[0048] This solution follows the principle of prioritizing high-priority requests, first addressing the bandwidth needs of high-priority devices and then processing the lower-priority ones. This fundamentally prevents high-priority intelligent backplanes associated with core services from experiencing bandwidth shortages due to low-priority devices preempting resources, thus ensuring the resource priority of core services.

[0049] Within a single priority level, the bandwidth requirements of all intelligent backplanes at that level are aggregated to clarify the overall resource demands of that level, providing data for subsequent assessments of whether the demands can be met. Then, real-time monitoring of system resource usage based on the total currently allocated bandwidth prevents over-allocation of bandwidth due to overlooking already occupied resources. Finally, a boundary threshold for resource allocation is established based on the remaining allocable bandwidth, defining the maximum bandwidth limit that can be obtained by the current level and preventing unrestricted allocation exceeding the system's total bandwidth capacity.

[0050] If the total demand of the current level is less than the remaining bandwidth, it means the system can fully meet the demand of that level, and the target bandwidth is the total demand. If the total demand is greater than the remaining bandwidth, the target bandwidth is the remaining bandwidth, preventing excessive resource consumption by that level from causing lower priority levels to have no bandwidth available. The target bandwidth is allocated proportionally to the demand of each smart backplane using a ratio, ensuring that devices within the same priority level do not monopolize resources due to excessively high demand, nor are they under-allocated due to low demand. For example, if the total demand of a certain level is 100Mbps and the remaining bandwidth is 80Mbps, with a ratio of 0.8, devices requiring 20Mbps are allocated 16Mbps, and devices requiring 30Mbps are allocated 24Mbps, achieving a balanced allocation of resources within the level. The next level of smart backplanes can continue to allocate bandwidth based on the remaining available bandwidth.

[0051] It is evident that this application not only safeguards the resource rights of high-priority services but also achieves efficient utilization of the system's total bandwidth, while avoiding resource allocation imbalance within a single level, providing an executable technical path for the precise and orderly scheduling of bandwidth resources.

[0052] In one possible embodiment, the control instructions include storage strategy optimization instructions. The backplane controller embeds a preset AI model. The process of generating control instructions based on the first multimodal data and generating second multimodal data based on the control instructions includes: obtaining the temperature, wear level, and input / output operations per second (IOPS) for each smart backplane based on the first multimodal data; inputting the temperature and wear level into the preset AI model to obtain a fault probability prediction value output by the preset AI model; determining the storage strategy optimization instructions for each smart backplane based on the fault probability prediction value and the IOPS; and generating second multimodal data based on the storage strategy optimization instructions. The storage strategy optimization instructions include independent disk redundant array (RAID) configuration instructions, data tiering instructions, and hibernation control instructions for each smart backplane.

[0053] Firstly, enhanced SGPIO is used to acquire multimodal data, accurately extracting three core parameters for each smart backplane-connected storage device: temperature, wear level, and IOPS. These three parameters correspond to the hardware health status, lifespan degradation trend, and workload intensity of the storage device, respectively. Temperature directly affects the operational stability of storage chips; excessively high temperatures can lead to increased data read / write latency or even errors. Wear level is a key lifespan indicator for flash memory devices, directly related to the probability of read / write failures. IOPS directly reflects the access frequency of the services carried by the device; for example, high IOPS corresponds to real-time database transactions, while low IOPS corresponds to archive data storage. These three parameters together form the data foundation for storage strategy optimization, ensuring that subsequent decisions cover both device hardware and business application dimensions, avoiding the one-sidedness caused by traditional strategies relying on only a single indicator.

[0054] Building upon this, temperature and wear levels are input into a pre-defined AI model embedded in the backplane controller. This model is a lightweight intelligent analysis model built based on the TinyML model library. Its architecture includes an input layer, a hidden layer, and an output layer. The hidden layer can include multiple units of Long Short-Term Memory (LSTM) networks. The introduction of LSTM effectively processes time-series changes in temperature and wear levels, such as a 6°C increase in temperature over three consecutive monitoring cycles and a monthly average increase in wear of 0.06, rather than relying solely on the current instantaneous value. This enables dynamic prediction of device failure risks. For example, when the wear level of an SSD connected to a smart backplane reaches 0.85 and the temperature exceeds 42°C for four consecutive cycles, the model's predicted failure probability can reach 0.75. This prediction result overcomes the limitation of traditional SGPIO, which can only issue alarms after a failure occurs. It provides a forward-looking risk basis for storage strategy optimization, shifting strategy adjustments from post-event remediation to pre-event prevention.

[0055] Subsequently, by combining the predicted failure probability with IOPS to determine storage strategy optimization instructions, the dual goals of risk control and business adaptation are achieved. Among these, Redundant Array of Independent Disks (RAID) is a technology that combines multiple independent hard drives into a disk array according to specific rules. It avoids data loss due to single-point hard drive failure through redundancy backup and accelerates transmission through data block parallel processing. Common RAID levels, such as RAID10 (balancing performance and redundancy) and RAID5 (balancing storage utilization and fault tolerance), can be flexibly selected according to the business's requirements for reliability and efficiency. Data tiering is a management method that allocates different types of data to corresponding performance-level storage resources based on data access frequency, such as frequently accessed hot data, frequently accessed warm data, and infrequently accessed cold data. Typically, hot data is stored in high-performance storage to ensure access speed, while cold data is stored in low-cost, low-speed storage to save resources, thereby optimizing storage resource utilization and reducing overall storage costs. Hibernation control is an energy management mechanism for storage devices under low load conditions, such as when IOPS is too low for a continuous period or the load is continuously low. When the preset hibernation conditions are met, the device suspends non-essential operating modules, reduces the rotation speed, or cuts off some power supply to enter a low-power hibernation state. When a wake-up signal such as a data access request is detected, it quickly resumes operation, thereby reducing energy consumption and extending the service life of the device without affecting business needs.

[0056] On the one hand, business load types are categorized based on IOPS. For example, when IOPS > 10k (high-load, high-frequency read / write business, such as real-time payment systems), to ensure business continuity and data read / write efficiency, the strategy optimization instructions will prioritize including RAID10 configuration instructions. RAID10 combines high read / write speeds with double data redundancy, which can meet the data security requirements in high-frequency read / write scenarios. When IOPS < 1% (low-access business, such as annual report archiving), the instructions will include RAID1 + hibernation control instructions. RAID1 ensures basic data redundancy, and hibernation control can reduce device power consumption under low-load conditions.

[0057] On the other hand, the predicted failure probability can serve as a risk correction factor to dynamically adjust the basic strategy. If a smart backplane's IOPS > 10k (RAID10 configuration required), but its predicted failure probability > 0.6 (high failure risk), the command will simultaneously add a data tiering instruction. This will migrate hot data (data accessed within the last hour) from that backplane to other smart backplanes in better health, while retaining cold data locally. This avoids the loss of hot data due to high-risk device failures and does not affect the storage efficiency of cold data. The thresholds in the data tiering instruction, such as 1 hour for hot data, 24 hours for warm data, and 72 hours for cold data, can be set based on the access patterns of most enterprise-level storage systems to ensure that the tiering logic conforms to actual business scenarios.

[0058] Finally, since enhanced SGPIO requires signal transmission with a specific data frame structure, storage strategy optimization instructions cannot be directly issued. Therefore, they need to be encapsulated into second multimodal data that conforms to the frame structure.

[0059] As can be seen, in this embodiment, by relying on the first multimodal data to obtain the temperature, wear and IOPS of the intelligent backplane, using the preset AI model to predict the probability of failure, and combining the probability of failure and IOPS to generate storage strategy optimization instructions including RAID configuration, data tiering and hibernation control, the intelligence and accuracy of server storage strategy formulation can be improved, and the operational reliability, resource utilization efficiency and energy consumption optimization level of the storage system can be enhanced.

[0060] In one possible embodiment, the control command includes an energy optimization command. The step of generating the control command based on the first multimodal data and generating the second multimodal data based on the control command includes: acquiring the temperature, IOPS, and load of each smart backplane based on the first multimodal data; determining the power supply mode of each smart backplane based on the temperature, the IOPS, and the load; generating the energy optimization command based on the power supply mode; and generating the second multimodal data based on the energy optimization command.

[0061] Firstly, three core parameters—temperature, IOPS, and load—are extracted from the first multimodal data for each smart backplane. The selection of these three parameters directly relates to the device's energy requirements and operational safety. Temperature reflects the device's heat dissipation pressure; excessively high temperatures require increased heat dissipation energy consumption to ensure hardware safety, while suitable temperatures reduce heat dissipation and operational energy consumption. IOPS reflects the level of business activity; high IOPS means the business has high requirements for device response speed, requiring a higher power supply to avoid performance bottlenecks. Low IOPS indicates that the device is under low load, with room for energy saving. Load directly reflects the device's resource utilization; high load requires sufficient power to support data processing, while low load allows for the reduction of unnecessary power supply. These three parameters together constitute the data benchmark for judging the actual energy requirements of the device, ensuring that subsequent adjustments to the power supply mode do not affect business operations while maximizing energy savings.

[0062] Based on this, the power supply mode of the intelligent backplane can be determined according to temperature, IOPS, and load, enabling dynamic switching of the power supply mode. For example, if an intelligent backplane is detected to have IOPS < 100, load < 20%, and temperature < 38℃ for three consecutive monitoring cycles, it indicates that the device is in a state of low service demand and low heat dissipation pressure. At this time, its power supply mode is switched to sleep mode to minimize energy consumption. If the detected temperature > 45℃ and load > 70% (regardless of IOPS), it indicates that the device is operating under high load and high heat dissipation pressure, and it needs to be switched to normal mode (full speed operation) to ensure device performance and hardware safety. When IOPS is at a medium level, such as 100 ≤ IOPS ≤ 5000, load 30%-60%, and temperature 38℃-45℃, it switches to light load mode (50% speed) or energy-saving mode (30% speed) to find the optimal solution between performance and energy consumption. This dynamic mode switching based on multiple parameters completely changes the drawback of the fixed power supply mode in traditional SGPIO systems, enabling precise matching of energy allocation with the actual needs of the device.

[0063] Subsequently, energy optimization instructions are generated based on the determined power supply mode. These instructions not only include power supply mode switching commands, such as switching from sleep mode to light load mode, but also integrate mode-adaptive operating parameters, such as the device speed threshold in light load mode and the core power supply voltage in energy-saving mode, ensuring that the intelligent backplane clearly understands the specific energy management actions to be performed. Finally, the energy optimization instructions are used to generate second-mode multimodal data.

[0064] As can be seen, this embodiment reduces unnecessary energy consumption through hibernation and energy-saving modes under low load, meeting the development needs of green data centers. It also ensures equipment performance and operational safety through normal mode under high load, and ensures effective command delivery through the reliable transmission of enhanced SGPIO. Ultimately, without affecting business operations, it significantly improves the energy utilization efficiency of the storage system, extends the service life of the equipment, and provides strong support for the overall economic efficiency and stability of the server operation.

[0065] In one possible embodiment, the method further includes: obtaining the base fan speed of each smart backplane and the maximum disk power of the storage device connected to each smart backplane based on the first multimodal data; determining the target fan speed of each smart backplane based on the temperature and the base fan speed; determining the target disk power of each smart backplane based on the maximum disk power and the load; the step of generating an energy optimization instruction based on the power supply mode and generating second multimodal data based on the energy optimization instruction includes: generating an energy optimization instruction based on the power supply mode, the target fan speed and the target disk power, and generating second multimodal data based on the energy optimization instruction.

[0066] First, the baseline fan speed and maximum disk power of the connected storage devices are acquired through multimodal data acquisition. Supplementing this data allows for precise capture of the device's basic heat dissipation and energy consumption status. The baseline fan speed reflects the current operating intensity of the cooling system, providing a benchmark for subsequent speed adjustments. The maximum disk power clarifies the upper limit of the storage device's energy consumption, defining a safe boundary for power regulation and preventing performance limitations due to excessive energy saving. Based on this, the target fan speed is determined according to temperature and the baseline fan speed, achieving a dynamic balance between heat dissipation and energy consumption. When the smart backplane temperature exceeds a preset threshold, the target speed can be appropriately increased from the baseline fan speed to ensure sufficient heat dissipation and maintain hardware stability. When the temperature is below the threshold, the target speed can be reduced to avoid ineffective energy consumption caused by high-speed fan operation. Furthermore, differentiated speed control can be implemented based on the temperature distribution differences of different smart backplanes, further improving heat dissipation efficiency.

[0067] Determining the target disk power based on maximum disk power and load enables precise matching of energy consumption and performance. When the load is low, the target disk power can be appropriately reduced without exceeding the maximum disk power, cutting unnecessary energy consumption. When the load rises to a critical threshold, the target disk power can be increased to a level suitable for the load, ensuring that the read / write performance of the storage device meets business needs and avoiding business delays due to insufficient power. In the energy optimization command generation stage, the power supply mode, target fan speed, and target disk power are integrated, making the command more comprehensive and executable. The command not only specifies the overall power supply mode of the device, such as hibernation, light load, and normal operation, but also precisely specifies the fan operating speed and disk operating power. These parameters are reliably transmitted through enhanced SGPIO communication between the backplane controller and the intelligent backplane, ensuring precise implementation of control actions.

[0068] As can be seen, this embodiment, through collaborative analysis of multi-dimensional data and linkage control of multiple parameters, can upgrade energy optimization from an extensive mode to a refined parameter management mode. It can avoid the problems of insufficient or excessive heat dissipation by dynamically adjusting the fan speed, and balance business performance and energy consumption by precisely controlling disk power. At the same time, combined with the overall management of power supply mode, it can minimize ineffective energy consumption and significantly improve energy utilization efficiency while ensuring the stable operation of the server storage system. It can also adapt to the differences in energy demand under different load scenarios, providing strong technical support for the construction of green data centers.

[0069] In one possible embodiment, the data frame structure of the multimodal data corresponding to the enhanced SGPIO includes at least a channel ID field, a data category field, a status data field, an AI instruction field, and a bandwidth allocation field; the channel ID field is used to identify each smart backplane, the data category field is associated with the basic data of the smart backplane, the status data field is associated with energy optimization instructions, the AI ​​instruction field is associated with storage strategy optimization instructions, and the bandwidth allocation field is associated with bandwidth allocation instructions.

[0070] The enhanced SGPIO multimodal data frame structure can include a channel ID field, a data category field, a status data field, an AI command field, and a bandwidth allocation field. Specifically, the channel ID field serves as a location identifier for the frame structure. It can accurately identify each smart backplane using hexadecimal encoding (e.g., 0x01, 0x02…0xFF). This means that in scenarios where a server contains multiple smart backplanes, various commands issued by the backplane controller, such as energy optimization commands and storage strategy optimization commands, can directly target the desired smart backplane through this field, preventing mis-sending or missed commands during transmission. Especially in bandwidth allocation scenarios, smart backplanes with different priority levels need to receive differentiated bandwidth allocation values. The precise identification of the channel ID field ensures that each smart backplane only receives bandwidth parameters that match its own, providing fundamental support for hierarchical bandwidth allocation.

[0071] The data category field can act as a data classifier within the frame structure, directly associated with the basic data of the smart backplane, such as temperature, wear level, IOPS, and load. It can distinguish different types of basic data through preset identifier values. For example, 0x01 can be set as the temperature data identifier, 0x02 as the IOPS data identifier, and 0x03 as the load data identifier. When the smart backplane uploads the first multimodal data to the backplane controller, the controller can quickly identify the data type by reading this field, eliminating the need for complex data parsing and significantly improving the efficiency of basic data processing. Simultaneously, the classification function of this field also facilitates subsequent on-demand data extraction. For instance, when generating energy optimization instructions, the controller can only filter data with data category fields of 0x01 (temperature), 0x02 (IOPS), and 0x03 (load) for analysis, avoiding irrelevant data from interfering with decision-making.

[0072] The status data field serves as the energy control carrier within the frame structure, deeply bound to energy optimization commands. It can carry the core parameters of these commands. For example, when an energy optimization command determines that a smart backplane needs to enter a light-load mode, this field can store specific parameters such as the target fan speed (e.g., 1500 rpm) and target disk power (e.g., 10W) ​​in light-load mode. If the command is for hibernation mode, it can store information such as hibernation trigger delay (e.g., 30 seconds) and wake-up response threshold (e.g., wake-up when IOPS > 50). This design ensures that all control details of the energy optimization command are completely transmitted to the smart backplane, avoiding incomplete energy-saving actions due to missing parameters. Furthermore, the length of this field can be flexibly set according to the number of parameters (e.g., 4 bytes), satisfying basic control requirements without causing frame structure redundancy.

[0073] The AI ​​instruction field can serve as an intelligent policy container within the frame structure, specifically designed to carry storage policy optimization instructions. Since storage policy optimization instructions involve multiple dimensions such as RAID configuration, data tiering, and hibernation control, this field can integrate various parameters through bit partitioning. For example, the first three bits can be used to identify the RAID level (e.g., 001 for RAID1, 101 for RAID10), the middle four bits can be used to store data tiering thresholds (e.g., 0001 for hot data over 1 hour, 0010 for warm data over 24 hours), and the last bit can be used to control hibernation functionality (e.g., 0 for disabling hibernation, 1 for enabling hibernation). This compact parameter storage method is compatible with policy adjustment needs corresponding to the fault probability prediction values ​​output by the preset AI model. For example, when the fault probability is high, a combined instruction of "RAID10 configuration + data migration" can be quickly issued through this field. It also ensures that storage policy optimization instructions are efficiently transmitted in enhanced SGPIO serial transmission without the need to split into multiple frames.

[0074] The bandwidth allocation field serves as a resource quota credential within the frame structure, directly associated with bandwidth allocation instructions. It stores the specific bandwidth allocation value for the smart backplane. For example, in Mbps, represented by a 2-byte hexadecimal number, 0x0400 corresponds to 1024 Mbps. Furthermore, to accommodate priority-level allocation logic, the first four bits of this field can be reserved to identify priority. For instance, 0001 represents high priority, 0010 represents medium priority, and 0100 represents low priority. This allows the smart backplane to obtain bandwidth quotas through the field body and confirm its own resource allocation order through the priority identifier, preventing lower-priority resources from preempting higher-priority resources when bandwidth is scarce.

[0075] The data frame may also include a frame header and a Cyclic Redundancy Check (CRC) field to ensure the integrity of data transmission and prevent command errors caused by signal interference.

[0076] As can be seen, in this embodiment, by designing a multimodal data frame structure, the accuracy and efficiency of multi-type instruction and data transmission can be improved, and the orderliness and reliability of server storage system management can be enhanced.

[0077] The following describes a control device based on enhanced SGPIO provided in this application. The control device based on enhanced SGPIO described below corresponds to the control method based on enhanced SGPIO described above.

[0078] Please see Figure 3The enhanced SGPIO-based control device 300 is applied to a server backplane controller. The server also includes multiple smart backplanes. The backplane controller is communicatively connected to the multiple smart backplanes and includes: an acquisition unit 301, used to acquire first multimodal data from the multiple smart backplanes via enhanced SGPIO; a generation unit 302, used to generate control commands based on the first multimodal data and generate second multimodal data corresponding to the control commands; and a sending unit 303, used to send the second multimodal data to the multiple smart backplanes via the enhanced SGPIO.

[0079] In one possible embodiment, the control command includes a bandwidth allocation command. In the process of generating the control command based on the first multimodal data and generating the second multimodal data based on the control command, the generation unit 302 is specifically configured to: determine the priority and bandwidth requirement of each smart backplane based on the first multimodal data; sort each smart backplane according to the priority to obtain multiple priority levels; allocate bandwidth to each smart backplane sequentially according to the multiple priority levels and the bandwidth requirement to obtain a bandwidth allocation value; and generate the second multimodal data of each smart backplane based on the bandwidth allocation value.

[0080] In one possible embodiment, in allocating bandwidth to each smart backplane according to the plurality of priority levels and the bandwidth requirements, the generation unit 302 is specifically configured to: perform the following operations for each smart backplane in each priority level until bandwidth is allocated to each smart backplane or until all bandwidth has been allocated: determine the current total bandwidth requirement based on the bandwidth requirement of each smart backplane in the current priority level; obtain the current total allocated bandwidth; determine the current remaining allocable bandwidth based on the current total bandwidth requirement and the current total allocated bandwidth; determine the ratio of a target bandwidth to the current total bandwidth requirement, the target bandwidth being the minimum bandwidth value between the current total bandwidth requirement and the remaining allocable bandwidth; and allocate bandwidth to each smart backplane in the current priority level based on the ratio and the bandwidth requirement of each smart backplane in the current priority level.

[0081] In one possible embodiment, the control instructions include storage strategy optimization instructions. A preset AI model is embedded in the backplane controller. Regarding the generation of control instructions based on the first multimodal data and the generation of second multimodal data based on the control instructions, the generation unit 302 is specifically configured to: acquire the temperature, wear level, and input / output operations per second (IOPS) of each smart backplane based on the first multimodal data; input the temperature and wear level into the preset AI model to obtain a fault probability prediction value output by the preset AI model; determine the storage strategy optimization instructions for each smart backplane based on the fault probability prediction value and the IOPS; and generate second multimodal data based on the storage strategy optimization instructions. The storage strategy optimization instructions include independent disk redundant array (RAID) configuration instructions, data tiering instructions, and hibernation control instructions for each smart backplane.

[0082] In one possible embodiment, the control command includes an energy optimization command. Specifically, in generating the control command based on the first multimodal data and generating the second multimodal data based on the control command, the generation unit 302 is configured to: acquire the temperature, IOPS, and load of each smart backplane based on the first multimodal data; determine the power supply mode of each smart backplane based on the temperature, the IOPS, and the load; generate the energy optimization command based on the power supply mode; and generate the second multimodal data based on the energy optimization command.

[0083] In one possible embodiment, the generation unit 302 is further configured to: obtain the base fan speed of each smart backplane and the maximum disk power of the storage device connected to each smart backplane based on the first multimodal data; determine the target fan speed of each smart backplane based on the temperature and the base fan speed; determine the target disk power of each smart backplane based on the maximum disk power and the load; and in the aspect of generating energy optimization instructions based on the power supply mode and generating second multimodal data based on the energy optimization instructions, the generation unit 302 is specifically configured to: generate energy optimization instructions based on the power supply mode, the target fan speed and the target disk power, and generate second multimodal data based on the energy optimization instructions.

[0084] In one possible embodiment, the data frame structure of the multimodal data corresponding to the enhanced SGPIO includes at least a channel ID field, a data category field, a status data field, an AI instruction field, and a bandwidth allocation field; the channel ID field is used to identify each smart backplane, the data category field is associated with the basic data of the smart backplane, the status data field is associated with energy optimization instructions, the AI ​​instruction field is associated with storage strategy optimization instructions, and the bandwidth allocation field is associated with bandwidth allocation instructions.

[0085] It is understood that since the method embodiments and the device embodiments are different presentations of the same technical concept, the content of the method embodiment section in this application should be adapted to the device embodiment section in a synchronous manner, and will not be repeated here.

[0086] In the case of using integrated units, please refer to Figure 4 .exist Figure 4 The enhanced SGPIO-based control device 400 includes a processing module 412 and a communication module 411. The processing module 412 controls and manages the actions of the enhanced SGPIO-based control device 400, for example, executing the steps of the acquisition unit 301, the generation unit 302, and the transmission unit 303, and / or performing other processes described herein. The communication module 411 is used for interaction between the enhanced SGPIO-based control device 400 and other devices. Figure 4 As shown, the control device 400 based on enhanced SGPIO may further include a storage module 413, which is used for the program code and data of the control device 400 based on enhanced SGPIO.

[0087] The processing module 412 can be a processor or controller, such as a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an ASIC, an FPGA, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc. The communication module 411 can be a transceiver, RF circuitry, or a communication interface, etc. The storage module 413 can be a memory.

[0088] All relevant content in each scenario involved in the above method embodiments can be referenced from the functional descriptions of the corresponding functional modules, and will not be repeated here. The above-mentioned control device 400 based on enhanced SGPIO can perform the above... Figure 2 The control method based on enhanced SGPIO is shown.

[0089] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of the electronic device provided in this application. For example... Figure 5As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communications interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call logic instructions in the memory 530 to execute a control method based on enhanced SGPIO. This method is applied to a backplane controller of a server, the server also including multiple smart backplanes, and the backplane controller is communicatively connected to the multiple smart backplanes. The method includes: acquiring first multimodal data from the multiple smart backplanes through enhanced serial general-purpose input / output (SGPIO); generating control instructions based on the first multimodal data, and generating second multimodal data corresponding to the control instructions; and sending the second multimodal data to the multiple smart backplanes through the enhanced SGPIO.

[0090] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0091] On the other hand, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform the enhanced SGPIO-based control method provided in the above embodiments. This method is applied to a backplane controller of a server, the server further comprising multiple smart backplanes, the backplane controller being communicatively connected to the multiple smart backplanes. The method includes: acquiring first multimodal data from the multiple smart backplanes via an enhanced serial general-purpose input / output (SGPIO); generating a control instruction based on the first multimodal data, and generating second multimodal data corresponding to the control instruction; and sending the second multimodal data to the multiple smart backplanes via the enhanced SGPIO.

[0092] In another aspect, this application also provides a computer program product, including a computer program that, when executed by a processor, implements any of the above-described enhanced SGPIO-based control methods. This method is applied to a backplane controller of a server, the server further including multiple smart backplanes, and the backplane controller communicatively connected to the multiple smart backplanes. The method includes: acquiring first multimodal data from the multiple smart backplanes via enhanced serial general-purpose input / output (SGPIO); generating control instructions based on the first multimodal data, and generating second multimodal data corresponding to the control instructions; and sending the second multimodal data to the multiple smart backplanes via the enhanced SGPIO.

[0093] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0094] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the methods described in the above method embodiments, wherein the computer includes an electronic device.

[0095] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0096] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0097] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0098] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0099] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0100] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0101] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage device, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0102] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A control method based on enhanced SGPIO, characterized in that, A backplane controller for a server, the server further comprising multiple intelligent backplanes, the backplane controller being communicatively connected to the multiple intelligent backplanes, the method comprising: First multimodal data from the multiple smart backplanes is acquired via the enhanced serial general purpose input / output (SGPIO). Control instructions are generated based on the first multimodal data, and the control instructions include bandwidth allocation instructions; The priority and bandwidth requirements of each smart backplane are determined based on the first multimodal data. Each smart backplane is sorted according to its priority to obtain multiple priority levels; Perform the following operations for each priority level of smart backplanes until bandwidth is allocated to each smart backplane or until all bandwidth has been allocated, to obtain the bandwidth allocation value: The total current bandwidth requirement is determined based on the bandwidth requirement of each smart backplane in the current priority level; Get the total currently allocated bandwidth; The remaining available bandwidth is determined based on the sum of the current bandwidth requirements and the sum of the currently allocated bandwidth. Determine the ratio of the target bandwidth to the sum of the current bandwidth requirements, where the target bandwidth is the minimum bandwidth value between the sum of the current bandwidth requirements and the remaining allocable bandwidth. Bandwidth is allocated to each smart backplane in the current priority level based on the ratio and the bandwidth requirement of each smart backplane in the current priority level. The second multimodal data for each smart backplane is generated based on the bandwidth allocation value; The second multimodal data is sent to the plurality of smart backplanes via the enhanced SGPIO.

2. The method according to claim 1, characterized in that, The control commands include storage strategy optimization commands. A preset AI model is embedded in the backplane controller. The process of generating control commands based on the first multimodal data and generating second multimodal data based on the control commands includes: The temperature, wear level, and input / output operations per second (IOPS) of each smart backplane are obtained based on the first multimodal data. The temperature and wear level are input into the preset AI model to obtain the fault probability prediction value output by the preset AI model; Based on the predicted failure probability and the IOPS, the storage strategy optimization instructions for each smart backplane are determined, and second multimodal data is generated based on the storage strategy optimization instructions. The storage strategy optimization instructions include independent disk redundant array RAID configuration instructions, data tiering instructions, and hibernation control instructions for each smart backplane.

3. The method according to claim 1, characterized in that, The control commands include energy optimization commands. The process of generating control commands based on the first multimodal data and generating second multimodal data based on the control commands includes: The temperature, IOPS, and load of each smart backplane are obtained based on the first multimodal data; The power supply mode of each smart backplane is determined based on the temperature, the IOPS, and the load. An energy optimization instruction is generated based on the power supply mode, and second multimodal data is generated based on the energy optimization instruction.

4. The method according to claim 3, characterized in that, The method further includes: The basic fan speed of each smart backplane and the maximum disk power of the storage device connected to each smart backplane are obtained based on the first multimodal data. The target fan speed for each smart backplate is determined based on the temperature and the base fan speed. The target disk power for each smart backplane is determined based on the maximum disk power and the load. The step of generating energy optimization instructions based on the power supply mode and generating second multimodal data based on the energy optimization instructions includes: Energy optimization instructions are generated based on the power supply mode, the target fan speed, and the target disk power, and second multimodal data is generated based on the energy optimization instructions.

5. The method according to any one of claims 1-4, characterized in that, The data frame structure of the multimodal data corresponding to the enhanced SGPIO includes at least a channel ID field, a data category field, a status data field, an AI instruction field, and a bandwidth allocation field; The channel ID field is used to identify each smart backplane, the data category field is associated with the basic data of the smart backplane, the status data field is associated with energy optimization instructions, the AI ​​instruction field is associated with storage strategy optimization instructions, and the bandwidth allocation field is associated with bandwidth allocation instructions.

6. A control device based on enhanced SGPIO, characterized in that, A backplane controller for a server, the server further comprising multiple intelligent backplanes, the backplane controller being communicatively connected to the multiple intelligent backplanes, the device comprising: An acquisition unit is used to acquire first multimodal data from the plurality of smart backplanes via enhanced SGPIO; A generation unit is configured to generate control instructions based on the first multimodal data, the control instructions including bandwidth allocation instructions; determine the priority and bandwidth requirement of each smart backplane based on the first multimodal data; sort each smart backplane according to the priority to obtain multiple priority levels; perform the following operations for each smart backplane in each priority level until bandwidth is allocated to each smart backplane or until all bandwidth has been allocated, to obtain a bandwidth allocation value: determine the current total bandwidth requirement based on the bandwidth requirement of each smart backplane in the current priority level; obtain the current total allocated bandwidth; determine the current remaining allocable bandwidth based on the current total bandwidth requirement and the current total allocated bandwidth; determine the ratio of the target bandwidth to the current total bandwidth requirement, the target bandwidth being the minimum bandwidth value between the current total bandwidth requirement and the remaining allocable bandwidth; allocate bandwidth to each smart backplane in the current priority level based on the ratio and the bandwidth requirement of each smart backplane in the current priority level; and generate second multimodal data for each smart backplane based on the bandwidth allocation value. The transmitting unit is used to transmit the second multimodal data to the plurality of smart backplanes via the enhanced SGPIO.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the control method based on enhanced SGPIO as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the control method based on enhanced SGPIO as described in any one of claims 1 to 5.

Images