WO2022252546A1 - 一种信息调节方法、设备及存储介质 - Google Patents

一种信息调节方法、设备及存储介质 Download PDF

Info

Publication number
WO2022252546A1
WO2022252546A1 PCT/CN2021/136434 CN2021136434W WO2022252546A1 WO 2022252546 A1 WO2022252546 A1 WO 2022252546A1 CN 2021136434 W CN2021136434 W CN 2021136434W WO 2022252546 A1 WO2022252546 A1 WO 2022252546A1
Authority
WO
WIPO (PCT)
Prior art keywords
data transmission
current
threshold
current limiting
transmission interface
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2021/136434
Other languages
English (en)
French (fr)
Inventor
邓城初
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
WeBank Co Ltd
Original Assignee
WeBank Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by WeBank Co Ltd filed Critical WeBank Co Ltd
Publication of WO2022252546A1 publication Critical patent/WO2022252546A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • the present application relates to the field of artificial intelligence, and in particular to an information adjustment method, device and storage medium.
  • the single-interface pressure test is usually used to measure the CPU usage of the transactions per second (Transactions Per Second, TPS) of a single interface on a single machine, and then calculate the target TPS and TPS of all interfaces.
  • TPS Transactions Per Second
  • the embodiment of the present application expects to provide an information adjustment method, device and storage medium, which solves the problem of low flexibility in the current limit configuration process and realizes an intelligent automatic adjustment of the current limit threshold. , by dynamically adjusting the current limiting parameter value of the data transmission interface according to the real-time transmission conditions of all data transmission interfaces, the flexibility and adjustment efficiency of the current limiting configuration process are effectively improved.
  • an information adjustment method the method includes:
  • the value of the current limiting parameter of each of the first data transmission interfaces is adjusted to the corresponding value of the current limiting threshold to be adjusted.
  • an information conditioning device comprising: a memory, a processor, and a communication bus; wherein:
  • the memory is used to store executable instructions
  • the communication bus is used to realize the communication connection between the processor and the memory
  • the processor is configured to execute the information adjustment program stored in the memory to implement the steps of the information adjustment method described in any one of the above.
  • a storage medium stores an information adjustment program, and when the information adjustment program is executed by a processor, the steps of the information adjustment method described in any one of the foregoing are implemented.
  • the current limit threshold after determining at least one first data transmission interface that is currently in the current limiting mode, determine the first current limiting threshold corresponding to the current limiting parameter of each first data transmission interface, and obtain at least one first current limiting threshold. flow threshold, and analyze the CPU occupancy rate of at least one first current current limiting threshold to determine the corresponding current limiting threshold of each data transmission interface to be adjusted, so as to adjust the current limiting parameter value of each first data transmission interface.
  • the current limit parameter value is dynamically adjusted, which effectively improves the flexibility and adjustment efficiency of the current limit configuration process.
  • FIG. 1 is a schematic flow diagram of an information adjustment method provided in an embodiment of the present application
  • FIG. 2 is a schematic flowchart of another information adjustment method provided by the embodiment of the present application.
  • FIG. 3 is a schematic flowchart of another information adjustment method provided by the embodiment of the present application.
  • FIG. 4 is a schematic flowchart of an information adjustment method provided by another embodiment of the present application.
  • FIG. 5 is a schematic flowchart of another information adjustment method provided by another embodiment of the present application.
  • FIG. 6 is a schematic diagram of a system architecture for implementing an information adjustment method provided by an embodiment of the present application.
  • FIG. 7 is a schematic structural diagram of an information adjustment device provided by an embodiment of the present application.
  • An embodiment of the present application provides an information adjustment method. Referring to FIG. 1, the method is applied to an information adjustment device, and the method includes the following steps:
  • Step 101 Determine at least one first data transmission interface that is currently in a current limiting mode.
  • the first data transmission interface generally refers to an interface provided by the information adjustment device for communicating with an external Internet system, and is used for receiving request information sent by an external application.
  • the current limiting mode refers to a way to limit the data transmission interface when the current number of requests of the data transmission interface exceeds the upper threshold of the data transmission interface processing requests. Post-processing, or sending requests exceeding the upper threshold to other interfaces of the same type for processing. In this way, it is possible to prevent the service system provided by the information adjustment device from collapsing due to too many requests, and effectively ensure that the information adjustment device can provide services normally, continuously and stably.
  • the information adjustment device can usually be a server device, which can be used to run a service system that provides services, or it can be an electronic device that runs some application systems with a large amount of information interaction.
  • Step 102 Determine a first current current limiting threshold corresponding to the current limiting parameter of each first data transmission interface, to obtain at least one first current current limiting threshold.
  • the first current current limiting threshold corresponding to the current limiting parameter of each first data transmission interface may be a preset current limiting upper limit value.
  • Step 103 analyzing the CPU usage of at least one first current current limiting threshold, and determining the corresponding current limiting threshold of each first data transmission interface to be adjusted.
  • a preset analysis method is used to analyze the CPU occupancy rate of each first current limit value, and the corresponding current limit threshold to be adjusted for each data transmission interface is determined, wherein the preset analysis method
  • the method can be a trained target neural network model, or some linear algorithm with determined weight coefficients, such as a matrix operation algorithm with determined weight coefficients, a polynomial regression algorithm, or a ridge regression algorithm.
  • Step 104 adjusting the value of the current limiting parameter of each first data transmission interface to the corresponding current limiting threshold to be adjusted.
  • each first data transmission interface is currently in the current limiting mode, the corresponding dynamic adjustment is made to the value of the current limiting parameter of each first data transmission interface, wherein the adjusted to-be-adjusted
  • the current limiting threshold can make the CPU usage rate of the information adjustment device not exceed the preset CPU usage rate, thereby ensuring the operation efficiency of the information adjustment device.
  • the current limit threshold after determining at least one first data transmission interface that is currently in the current limiting mode, determine the first current limiting threshold corresponding to the current limiting parameter of each first data transmission interface, and obtain at least one first current limiting threshold. flow threshold, and analyze the CPU occupancy rate of at least one first current current limiting threshold to determine the corresponding current limiting threshold of each data transmission interface to be adjusted, so as to adjust the current limiting parameter value of each first data transmission interface.
  • the current limit parameter value is dynamically adjusted, which effectively improves the flexibility and adjustment efficiency of the current limit configuration process.
  • the embodiments of the present application provide an information adjustment method, as shown in FIG. 2 , the method is applied to an information adjustment device, and the method includes the following steps:
  • Step 201 Determine at least one second data transmission interface that is in a data transmission mode at the current moment.
  • the information adjustment device may be a bank server device that provides various bank business services, and the data transmission mode refers to a mode in which the interface survives and the interface can normally transmit data information.
  • the at least one second data transmission interface refers to all surviving data transmission interfaces currently used by the information adjustment device to provide services. Exemplarily, it is assumed that the information adjustment device determines that the at least one second data transmission interface in the data transmission mode at the current moment is interface 1, interface 2, interface 3, interface 4, and interface 5.
  • Step 202 Determine a first current current limiting threshold corresponding to a current limiting parameter of each second data transmission interface.
  • the information adjustment device determines the first current current limiting threshold set by each second data transmission interface for its current limiting parameter. Exemplarily, it is determined that the first current limit threshold of interface 1 is a1, the first current limit threshold of interface 2 is a2, the first current limit threshold of interface 3 is a3, and the first current limit threshold of interface 4 is The threshold is a4 and the first current current limiting threshold of interface 5 is a5.
  • Step 203 Determine the current request quantity value of each second data transmission interface.
  • the information adjustment device determines the current request quantity value of each second data transmission interface included in the at least one second data transmission interface, and obtains at least one current request quantity value. Exemplarily, it is determined that the current request quantity value of interface 1 is q1, the current request quantity value of interface 2 is q2, the current request quantity value of interface 3 is q3, the current request quantity value of interface 4 is q4, and the current request quantity value of interface 5 is The requested quantity value is q5.
  • Step 204 From the at least one second data transmission interface, determine the second data transmission interface whose current request quantity value is greater than the corresponding first current current limiting threshold, and obtain at least one first data transmission interface.
  • the current request quantity value of each second data transmission interface is compared with its corresponding first current current limiting threshold to obtain the second data whose current request quantity value is greater than the corresponding first current current limiting threshold transmission interface, and determine the second data transmission interface whose current request quantity value is greater than the corresponding first current limiting threshold as at least one first data transmission interface.
  • at least one first data transmission interface includes one first data transmission interface and at least two first data transmission interfaces.
  • each current request quantity value is smaller than the corresponding first current current limiting threshold, that is, a first data transmission interface cannot be determined, that is to say, there is no need to A current limiting operation is performed on at least one second data transmission interface.
  • Step 205 Determine the first current current limiting threshold corresponding to the current limiting parameter of each first data transmission interface, and obtain at least one first current current limiting threshold.
  • the first current current limiting threshold corresponding to each first data transmission interface may be directly obtained from the determined first current current limiting threshold corresponding to at least one second data transmission interface.
  • the determined at least one first data transmission interface is at least one second data transmission interface, that is, interface 2, interface 4, and interface 5 among interface 1, interface 2, interface 3, interface 4, and interface 5, and the corresponding
  • the first current limit thresholds corresponding to the current limit parameters of each first data transmission interface included in the at least one first data transmission interface are in turn: the first current limit threshold corresponding to interface 2 is a2, and the first current limit threshold corresponding to interface 4 is a2.
  • the first current limiting threshold is a4 and the first current limiting threshold corresponding to the interface 5 is a5.
  • Step 206 Analyze the CPU usage of at least one first current current limiting threshold, and determine the corresponding current limiting threshold of each first data transmission interface to be adjusted.
  • the first current limiting threshold corresponding to interface 2 is a2
  • the first current limiting threshold corresponding to interface 4 is a4
  • the first current limiting threshold corresponding to interface 5 is a5.
  • the current limiting threshold to be adjusted corresponding to interface 2 the current limiting threshold to be adjusted corresponding to interface 4, and the current limiting threshold to be adjusted corresponding to interface 5 are obtained.
  • it may be obtained by independently analyzing and determining the first current current limiting threshold of each corresponding first data transmission interface.
  • Step 207 adjusting the value of the current limiting parameter of each first data transmission interface to the corresponding current limiting threshold to be adjusted.
  • the information adjustment device determines the current limiting threshold to be adjusted for each first data transmission interface, it adjusts the value of the current limiting parameter of each first data transmission interface to the corresponding current limiting threshold to be adjusted, However, the values of the current limiting parameters of the at least one second data transmission interface other than the at least one first data transmission interface are not adjusted.
  • the current limit threshold after determining at least one first data transmission interface that is currently in the current limiting mode, determine the first current limiting threshold corresponding to the current limiting parameter of each first data transmission interface, and obtain at least one first current limiting threshold. flow threshold, and analyze the CPU occupancy rate of at least one first current current limiting threshold to determine the corresponding current limiting threshold of each data transmission interface to be adjusted, so as to adjust the current limiting parameter value of each first data transmission interface.
  • the current limit parameter value is dynamically adjusted, which effectively improves the flexibility and adjustment efficiency of the current limit configuration process.
  • An embodiment of the present application provides an information adjustment method, as shown in FIG. 3 , the method is applied to an information adjustment device, and the method includes the following steps:
  • Step 301 Determine at least one second data transmission interface that is in a data transmission mode at the current moment.
  • Step 302. Determine a first current current limiting threshold corresponding to a current limiting parameter of each second data transmission interface.
  • Step 303 Determine the current request quantity value of each second data transmission interface.
  • Step 304 From the at least one second data transmission interface, determine the second data transmission interface whose current request quantity value is greater than the corresponding first current limiting threshold, and obtain at least one first data transmission interface.
  • Step 305 Determine a first current current limiting threshold corresponding to the current limiting parameter of each first data transmission interface, to obtain at least one first current current limiting threshold.
  • Step 306 acquiring a trained target neural network model.
  • the trained target neural network model may be obtained by performing model training on a back propagation (Back Propagation, BP) neural network using a large amount of historical data.
  • the historical data includes the current limiting threshold of each data transmission interface, and on the premise of the current limiting threshold of each data transmission interface, the corresponding information adjusts the CPU usage of the device.
  • step 306 can be realized by the following steps: obtaining historical sample data; cleaning historical sample data, eliminating error data, and obtaining target sample data; using target sample data to perform model training on the initialized neural network model to obtain the target neural network model .
  • Step 307. Determine a preset step value of each first data transmission interface.
  • the preset step value of each first data transmission interface may be preset, or may be calculated according to actual conditions.
  • Step 308 Based on the first current limit threshold of at least one first data transmission interface and the corresponding preset step value of each first data transmission interface, use the target neural network model to perform CPU usage prediction processing, and determine the corresponding The current limiting threshold to be adjusted for each first data transmission interface.
  • the input to the target neural network model Layer so that the target neural network model predicts the knowledge after adjusting the first current limit threshold of the corresponding first data transmission interface based on the preset step value of each first data transmission interface, and then according to the prediction result, to Determine the current limit threshold to be adjusted corresponding to each first data transmission interface.
  • Step 309 adjusting the value of the current limiting parameter of each first data transmission interface to the corresponding current limiting threshold to be adjusted.
  • step 307 may be implemented by steps 307a or 307b-307c; wherein, when the number of at least one first data transmission interface is 1, choose to execute step 307a, and at least one When the number of the first data transmission interface is greater than or equal to 2, choose to perform steps 307b-307c:
  • Step 307a if at least one first data transmission interface includes one first data transmission interface, determine a preset step value corresponding to the first data transmission interface as a unit step value.
  • the unit step value may be a preset positive integer value, for example, the unit step value may be 1.
  • Step 307b If at least one first data transmission interface includes at least two first data transmission interfaces, determine a proportional relationship between first current limiting thresholds of the at least two first data transmission interfaces.
  • a proportional calculation is performed on the first current limiting threshold of the at least one first data transmission interface to obtain a corresponding proportional relationship.
  • Step 307c based on the proportional relationship and the unit step value, determine a preset step value corresponding to each first data transmission interface.
  • the unit step value determines the preset step value of interface 2 1.
  • the preset step value of port 4 is 2, and the default step value of port 5 is 1.
  • step 308 can be implemented by steps 308a-308h, or steps 308a and 308i-308p;
  • the number of second interfaces of the second data transmission interface is the same, choose to execute steps 308a-308h, if the number of second interfaces is greater than the number of first interfaces, steps 308a and 308i-308p:
  • Step 308a Determine the sum of the first current current limiting threshold and the corresponding preset step value of each first data transmission interface to obtain at least one first reference current limiting threshold.
  • the first reference current limiting threshold of the first data transmission interface is the first current current limiting threshold+1.
  • first reference current limiting threshold first current current limiting threshold+preset step value
  • first reference current limiting threshold a21 corresponding to interface 2 is a2+1
  • first reference current limiting threshold a41 corresponding to interface 4 is a4+2
  • first reference current limiting threshold a51 corresponding to interface 5 is a5+1.
  • Step 308b if the number of the first interface of the at least one first data transmission interface is the same as the number of the second interface of the at least one second data transmission interface, through the target neural network model, perform the central processing unit on at least one first reference current limiting threshold (Central Processing Unit, CPU) occupancy rate prediction to obtain the first predicted occupancy rate.
  • CPU Central Processing Unit
  • the number of first interfaces of at least one first data transmission interface is the same as the number of second interfaces of at least one second data transmission interface, indicating that the current request quantity values of all data transmission interfaces of the information adjustment device exceed If the corresponding first current current limiting threshold is determined, the current limiting parameters of all data transmission interfaces need to be adjusted. At this time, if the first interface number of at least one first data transmission interface is 1, the second number of at least one second data transmission interface is also 1.
  • Input at least one first reference current-limiting threshold corresponding to at least one first data transmission interface to the input layer of the target neural network model, and use the target neural network model to predict that in the case of at least one first reference current-limiting threshold, information adjustment
  • the occupancy rate of the central processing unit (Central Processing Unit, CPU) corresponding to the device is used to obtain the first predicted occupancy rate.
  • CPU Central Processing Unit
  • Step 308c acquiring the current CPU usage at the current moment.
  • the information adjustment device when the current occupancy rate of the CPU at the current moment and the current limiting parameter of each second data transmission interface included in at least one second data transmission interface are the first current current limiting threshold, the information adjustment device corresponds to The actual CPU usage.
  • Step 308d Determine a first difference between the first predicted occupancy rate and the current occupancy rate.
  • the first difference the first predicted occupancy rate ⁇ the current occupancy rate.
  • step 308d it can choose to execute step 308e, or choose to execute steps 308f-308h; where, if the first difference is greater than or equal to the target error, choose to execute step 308e, if the first difference is smaller than the target error , choose to execute steps 308f to 308h:
  • Step 308e if the first difference is greater than or equal to the target error, determine the current-limiting threshold to be adjusted for each first data transmission interface as the corresponding first reference current-limiting threshold.
  • the target error is an error experience value obtained from a large number of experiments, or an error experience value set by the user according to actual needs.
  • the target error can be continuously corrected and adjusted.
  • the first difference is greater than or equal to the target error, it is determined that the adjusted first reference current limit threshold meets the requirements, therefore, it is determined that the adjusted current limit threshold of each first data transmission interface is the corresponding first reference limit flow threshold.
  • the first current limit threshold corresponding to at least one first data transmission interface is: the first current limit threshold of interface 1 is a1, the first current limit threshold of interface 2 is a2, and the first current limit threshold of interface 3 is The first current limiting threshold is a3, the first current limiting threshold of interface 4 is a4, and the first current limiting threshold of interface 5 is a5.
  • the determined preset step value of interface 1 is b1, interface The preset step value of 2 is b2, the default step value of interface 3 is b3, the preset step value of interface 4 is b4, and the preset step value of interface 5 is b5. Therefore, the determined interface 1
  • the target neural network model determines that the current limiting threshold to be adjusted for interface 1 is a11, the current limiting threshold for interface 2 is a21, and the current limiting threshold for interface 3 is determined to be a11.
  • the current limiting threshold to be adjusted is a31, the current limiting threshold to be adjusted of interface 4 is a41, and the current limiting threshold to be adjusted of interface 5 is a51.
  • Step 308f if the first difference is smaller than the target error, determine the sum of each first reference current-limiting threshold and the corresponding preset step value to obtain at least one second reference current-limiting threshold.
  • Step 308g updating at least one first reference current limiting threshold to at least one second reference current limiting threshold.
  • Step 308h return to execute "predict the CPU occupancy rate of at least one first reference current limiting threshold through the target neural network model to obtain the first predicted occupancy rate", until after n times of looping, when the first difference is greater than or equal to the target In the case of an error, or in the case of n being greater than or equal to the first preset number of times, determine that the current-limiting threshold to be adjusted for each first data transmission interface is the corresponding updated first reference current-limiting threshold.
  • the updated first reference current limiting threshold is the sum of n times the preset step value corresponding to the corresponding first data transmission interface and the corresponding first current limiting threshold, and n is a positive value greater than or equal to 1. integer.
  • the target neural network model is used to predict the updated a11 of interface 1, a21 of interface 2, a31 of interface 3, a41 of interface 4, and a51 of interface 5 to obtain the corresponding first prediction occupancy, and then determine a first reference difference between the first predicted occupancy and the current occupancy.
  • the first preset number of times is an experience value of times obtained from a large number of experiments, or may be an experience value of times set by the user according to actual needs.
  • Step 308i If the second interface number is greater than the first interface number, determine a first current current limiting threshold of at least one third data transmission interface.
  • the at least one third data transmission interface is a data transmission interface in the at least one second data transmission interface except the at least one first data transmission interface.
  • the number of second interfaces is greater than the number of first interfaces, indicating that at least one first data transmission interface is a part of the data transmission interface included in at least one second data transmission interface, that is, only some transmission interfaces are in the current limiting mode , determine at least one third data transmission interface that is not in the current limiting mode in the information adjustment device, and determine the first current current limiting threshold of the at least one third data transmission interface.
  • at least one second data transmission interface is composed of at least one first data transmission interface and at least one third data transmission interface.
  • Step 308j Using the target neural network model, perform CPU occupancy prediction on at least one first reference current limiting threshold and the first current current limiting threshold of at least one third data transmission interface to obtain a second preset occupancy.
  • the request traffic of at least one second data transmission interface has an impact on the CPU occupancy rate of the information adjustment device. Therefore, when using the target When the neural network model predicts the CPU usage, at least one current limiting parameter of the second data transmission interface needs to be considered. Therefore, when determining the current limiting parameter of at least one first data transmission interface through the target neural network model, it is also necessary to input the first current limiting threshold of at least one third data transmission interface to obtain an accurate at least one first The current limit threshold to be adjusted for the data transmission interface.
  • Step 308k acquiring the current CPU usage at the current moment.
  • Step 308l Determine a second difference between the second preset occupancy rate and the current occupancy rate.
  • step 308l it can choose to execute step 308m, or choose to execute steps 308n-308p; where, if the second difference is greater than or equal to the target error, choose to execute step 308m, if the second difference is less than or equal to Target error, choose to execute steps 308n-308p:
  • Step 308m if the second difference is greater than or equal to the target error, determine the current-limiting threshold to be adjusted for each first data transmission interface as the corresponding first reference current-limiting threshold.
  • Step 308n if the second difference is smaller than the target error, determine the sum of each first reference current-limiting threshold and the corresponding preset step value to obtain at least one second reference current-limiting threshold.
  • Step 308o updating at least one first reference current limiting threshold to at least one second reference current limiting threshold.
  • Step 308p return to execute step 308j, until after looping m times, if the second difference is greater than or equal to the target error, or m is greater than or equal to the second preset number of times, determine each first data transmission interface
  • the current limiting threshold to be adjusted is the corresponding updated first reference current limiting threshold.
  • the updated first reference current limiting threshold is the sum of m times the preset step value corresponding to the corresponding first data transmission interface and the corresponding first current limiting threshold, and m is a positive value greater than or equal to 1. integer.
  • the information adjustment device is further configured to perform the following steps: If it detects that there is an interface in current-limiting mode among at least one second data transmission interface, return to execute " Determine at least one first data transmission interface that is currently in the current-limiting mode, until the number of interfaces in the current-limiting mode in the at least one second data transmission interface is zero, or the updated data included in the at least one second data transmission interface
  • the current limiting parameter corresponding to the at least one first data transmission interface is adjusted to the corresponding first to-be-adjusted current limiting threshold, the determined first actual CPU occupancy rate is greater than or equal to the preset occupancy rate.
  • the working mode of each data transmission interface is constantly checked, that is, whether it is in the current limiting mode. Once it is detected that there is an interface in the current limiting mode in at least one second data transmission interface, then from the step 301 starts to be executed repeatedly, and adjusts the current limiting parameters of the data transmission interface in the current limiting mode to ensure the efficiency of data transmission.
  • the information adjustment device may also be used to perform steps 310-312:
  • Step 310 acquiring the first actual CPU usage rate corresponding to the value of the current limiting parameter of each first data transmission interface when the current limiting threshold is to be adjusted.
  • the information adjustment device can collect the first value of the CPU of the corresponding information adjustment device when the value of the current limit parameter of each first data transmission interface is adjusted to the corresponding current limit threshold to be adjusted. actual occupancy.
  • Step 311 determining the corresponding target predicted occupancy rate obtained through the prediction of the target neural network model when the value of the current limiting parameter of each first data transmission interface is the current limiting threshold to be adjusted.
  • the corresponding target predicted occupancy rate can be determined according to the aforementioned implementation process, if the first difference is greater than or is equal to the target error, the target predicted occupancy rate can be determined as the first predicted occupancy rate; if the second difference is greater than or equal to the target error, the target predicted occupancy rate can be determined as the second predicted occupancy rate.
  • Step 312 Update the target neural network model based on the target predicted occupancy rate and the first actual occupancy rate.
  • the information adjustment device performs feedback processing according to the first actual occupancy rate obtained after adjusting at least one first data transmission interface, and updates and adjusts the target neural network model to make the target neural network model more realistic
  • the application scenario ensures the accuracy of the target neural network model and effectively ensures the reliability of the prediction process.
  • the information adjustment device is also used to perform steps 313-314:
  • Step 313 If it is detected that in at least one second data transmission interface, the current request quantity value of each second data transmission interface is less than the corresponding first current limiting threshold, determine the ratio of each first current limiting threshold to a preset value The product of , get the corresponding feedback current limit threshold.
  • the feedback current limiting threshold is smaller than the first current current limiting threshold.
  • the preset ratio is an empirical ratio value obtained from a large number of experiments, and may also be an empirical ratio value set by the user according to actual needs. If the at least one second data transmission interface included in the information adjustment device does not need to perform current limiting processing, determine the product of each first current current limiting threshold and the preset ratio to obtain the corresponding feedback current limiting threshold.
  • the at least one first current limit threshold corresponding to at least one second data transmission interface is sequentially: the first current limit threshold of interface 1 is a1, the first current limit threshold of interface 2 is a2, and the first current limit threshold of interface 2 is a2.
  • the first current limiting threshold of interface 3 is a3, the first current limiting threshold of interface 4 is a4, and the first current limiting threshold of interface 5 is a5.
  • the preset ratio is ⁇
  • the corresponding interface 1 The feedback current-limiting threshold of interface 2 is recorded as ⁇ *a1
  • the feedback current-limiting threshold of interface 2 is recorded as ⁇ *a2
  • the feedback current-limiting threshold of interface 3 is recorded as ⁇ *a3
  • the feedback current-limiting threshold of interface 4 is recorded as ⁇ *a4
  • the feedback current limiting threshold of interface 5 is denoted as ⁇ *a5.
  • Step 314 if it is detected that among at least one second data transmission interface, the current request quantity value of each second data transmission interface is less than the corresponding feedback current limiting threshold, adjust the value of the current limiting parameter of each second data transmission interface To initialize the current limit threshold.
  • the initial current limit threshold is the default current limit threshold set by the information adjustment device for each second data transmission interface.
  • the system of the device is set by default.
  • the current request quantity value of interface 1 is q1
  • the current request quantity value of interface 2 is q2
  • the current request quantity value of interface 3 is q3
  • the current request quantity value of interface 4 is q4
  • the current request quantity value of interface 5 is The quantity value is q5, therefore, when q1 ⁇ *a1, q2 ⁇ *a2, q3 ⁇ *a3, q4 ⁇ *a4 and q5 ⁇ *a5, adjust the current limit of each second data transmission interface
  • the value of the parameter is the initial current limit threshold.
  • the embodiment of the present application provides a system architecture diagram for implementing an information adjustment method, as shown in FIG. Dictionary service (Remote Dictionary Server, Redis) 46 and application server agent (Agent) 47; Wherein:
  • the configuration module 41 is used to provide service operation and maintenance personnel (referred to as operation and maintenance personnel) with an interface for filling in the configuration information of each data transmission interface and the adjustable flag bit.
  • the configuration information includes at least the aforementioned initialized current limiting threshold. Suppose there are three second data transmission interfaces A, B, and C.
  • the configuration information for the three second data transmission interfaces A, B, and C will be Write the configuration table rate_config preset in the database 45 for recording the configuration information of each second data transmission interface, and write the initial current limiting thresholds for the three second data transmission interfaces A, B and C into the database 45
  • the naming method of the initial current limit threshold of each second data transmission interface can be recorded as "interface name_limit”.
  • the information adjustment method provided by this application is executed only when the adjustable flag is marked as adjustable, and the information adjustment method provided by this application is not executed when the adjustable flag is marked as non-adjustable.
  • the database 45 also includes the number of requests received for each time slice of each second data transmission interface, and the naming method of the number of requests received by each time slice of each second data transmission interface is record It is "interface name_request_cnt"; an agent_upload_data_cal table used to record the current limit threshold of each second data transmission interface in each time slice, the number of requests for each second data transmission interface, and the CPU usage rate.
  • the acquisition module 42 is used to receive the data of each time slice returned by the application server agent 47, which may be JavaScript object notation (JavaScript Object Notation, JSON) data.
  • JSON JavaScript Object Notation
  • the JSON data format received by the acquisition module 42 may be as follows:
  • the acquisition module 42 After the acquisition module 42 receives the above data, it analyzes it and writes the analyzed data into the agent_upload_data_log data table in the database 45 .
  • Acquisition module 42 writes the data that analysis obtains in the agent_upload_data_log data table in the database 45 and comprises: create asynchronous task one, and asynchronous task one query records the number of requests of each second data request interface in the Redis 46 of this time slice, i.e.
  • asynchronous task 2 calculates the current limit value A_limit_sum of the time slice for A data transmission interface, and B data transmission The current-limited value B_limit_sum of the interface and the current-limited value C_limit_sum of the C data transmission interface. It should be noted that, taking the A data transmission interface as an example, the current-limited value A_limit_sum of the A data transmission interface is based on the The number of requests for this time slice is determined by the current limit threshold of A data transmission interface in this time slice. The number of requests for A data transmission interface in this time slice is less than or equal to the limit of A data transmission interface in this time slice.
  • A_limit_sum is determined to be 0.
  • A_limit_sum is determined to be the request of A data transmission interface in this time slice.
  • the number of requests is A_request_cnt, B_request_cnt, C_request_cnt; finally, time_stamp, A_limit_sum, B_limit_sum, C_limit_sum, cpu_avg, A_request_cnt, B_request_cnt, C_request_cnt are written into the preset agent_upload_data_cal table in the database 45.
  • the training module 43 is used to perform model training on the BP initialization neural network model to obtain the target neural network model, and continuously update the model of the target neural network model.
  • the training module 43 is driven by a timed task, for example, it may train once an hour.
  • the specific implementation process of the training module 43 is as follows:
  • A_limit_sum is the current limit value corresponding to the A data transmission interface in a certain time slice
  • B_limit_sum is the current limit value corresponding to the B data transmission interface in a certain time slice
  • C_limit_sum is the C data transmission interface in a certain time slice.
  • Current limiting is to reject requests at the bottom of the application. At this time, the application consumes almost no resources, so this part of the data needs to be removed, otherwise the calculation accuracy will be affected.
  • the first data set is traversed to determine whether at least one current-limited value in each group of A_limit_sum, B_limit_sum, and C_limit_sum is greater than 0. If there is at least one data transmission interface whose current-limited value is greater than 0, determine the judgment condition If it is established, each group of A_limit_sum, B_limit_sum and C_limit_sum including the judgment condition is established is eliminated from the first data set, and finally the second data set is obtained.
  • the specifications of the server devices in the same operating system are consistent, and the number of online requests processed is almost equal, so the difference in CPU utilization does not exceed 5%.
  • Traversing the second data set, using the time slice time_stamp as an index query the corresponding time slice in agent_upload_data_log whether there is a difference between the CPU usage of a certain server in the time slice and the average CPU usage cpu_avg of all servers in the time slice exceeds 0.05. If the CPU usage rate of a certain server exceeds the average CPU usage rate cpu_avg, the data of the server in the time slice is removed from the second data set to obtain the third data set.
  • the number of input variables of the BP neural network model is the number of data transmission interfaces included in the configuration table rate_config; the number of hidden layers is at least 2, and the number of output variables is 1.
  • the number of input variables of the BP neural network model is 3, and the number of hidden layers is assumed to be 3.
  • the weight coefficient and bias coefficient initialization process of the BP neural network model input each group of A_request_cnt, B_request_cnt and C_request_cnt in the fourth data set to the input layer of the BP neural network model; expected output: and each group in the fourth data set
  • the cpu_avg corresponding to A_request_cnt, B_request_cnt and C_request_cnt; the training output is input A_request_cnt, B_request_cnt, C_request_cnt to the input layer of the BP neural network model, and the output layer of the BP neural network model outputs the predicted CPU usage rate Y_cpu_avg; calculation error Err cpu_avg-Y_cpu_avg, Update the weight coefficient and bias coefficient of the BP neural network model according to the error Err, until the error between the predicted CPU usage calculated by the updated BP neural network model and the corresponding cpu_avg
  • the adjustment module 44 is configured to automatically adjust the current limiting threshold configuration of each data transmission interface in gray scale based on the principle of negative feedback.
  • the adjustment module 44 checks the data reported by the agent_upload_data_cal latest server agent 46, and checks whether the current-limited value of each data transmission interface is greater than 0, that is, whether the data request quantity of each data transmission interface is greater than the current-limited threshold of each data transmission interface . If the current limit value of the data transmission interface is greater than 0, the current limit threshold adjustment task is started, and the target neural network model is obtained as the operation model for this adjustment. For the current limit situation of different data transmission interfaces, it can be divided into single interface occurrence The following two adjustment methods are adopted for current limiting and multi-interface current limiting:
  • the current limit threshold of the current limited B data transmission interface is increased by a unit step value such as 1 to update B_request_cnt, A
  • the data transmission interface A_request_cnt is the original configuration current limiting threshold
  • the C data transmission interface C_request_cnt is the original configuration current limiting threshold.
  • Input A_request_cnt, updated B_request_cnt and C_request_cnt into the target neural network model to obtain the first predicted CPU usage, At the same time, calculate the difference between the first predicted CPU usage rate and the cpu_avg of the current time slice.
  • the difference is greater than or equal to 0.05, end; otherwise, repeat the above steps and continuously adjust the B_request_cnt of the B data transmission interface until it passes the target neural network. If the difference between the third predicted CPU usage obtained by the model prediction and cpu_avg is greater than 0.05, the loop is terminated, and the updated B_request_cnt is determined as the latest current limit threshold of the B data transmission interface, and then the latest limit of the updated B_request_cnt is set to The configuration of the flow threshold is delivered to each application; finally, the adjusted configuration and the corresponding predicted CPU usage are stored in the database 45 .
  • the updated A_request_cnt, B_request_cnt and C_request_cnt are input into the target neural network model , get the first predicted CPU occupancy rate, and calculate the cpu_avg difference between the predicted CPU occupancy rate and the current time slice at the same time, so that the preset step value of the A data transmission interface and the preset step value of the B data transmission interface are continuously used
  • the current limit threshold corresponding to the interface and the current limit threshold corresponding to the B data transmission interface send and configure the latest current limiting threshold corresponding to the A data transmission interface and the current limiting threshold corresponding to the B data transmission interface to each application, and finally, store the adjusted configuration and the corresponding third predicted CPU usage rate in the
  • the adjustment module 44 After the adjustment module 44 sends the updated configuration to each application, it collects the latest data reported by the application server agent 47 in the next time period, such as 5 seconds from the current period, to obtain the updated configuration correspondence
  • the first actual CPU occupancy rate of and calculate the error between the first actual CPU occupancy rate and the corresponding third predicted CPU occupancy rate, and negatively feed back the error to the target neural network model, and perform model training on the target neural network model Update to get the updated target neural network model.
  • the adjustment module 44 checks the data reported to the agent_upload_data_cal table by the application server agent 47 in (2), and checks whether the current-limited value of each data transmission interface is greater than 0. If there is still a current-limited value of the data transmission interface greater than 0, the current limit still exists, and (1) and (2) can be repeated until the latest actual CPU usage collected in (2) exceeds 75% or there is no current limit. If the current-limited data transmission interface is used, the adjustment module 44 terminates the adjustment task and initializes the monitoring task.
  • the adjustment module 44 is also used to detect the data reported by the agent_upload_data_cal latest application server agent 47, if it detects that the request quantity values of each data transmission interface, such as A_request_cnt, B_request_cnt and C_request_cnt, are less than the corresponding current current limit thresholds When half of the aforementioned preset ratio is 50%, the current limiting thresholds of each data transmission interface are restored to the initial current limiting thresholds in the configuration table rate_config.
  • the first current limiting threshold corresponding to the current limiting parameter of each first data transmission interface, and obtain at least one first current limiting threshold. flow threshold, and analyze the CPU occupancy rate of at least one first current current limit threshold to determine the current limit threshold to be adjusted corresponding to each first data transmission interface, so that the current limit parameters of each first data transmission interface
  • the value is adjusted to the current limit threshold to be adjusted, which solves the problem of low flexibility in the current limit configuration process, and realizes an intelligent automatic adjustment of the current limit threshold.
  • the current limit parameter value of the interface is dynamically adjusted, which effectively improves the flexibility and adjustment efficiency of the current limit configuration process.
  • the embodiments of the present application provide an information adjustment device, which can be applied to the information adjustment method provided in the embodiments corresponding to Figures 1 to 5, as shown in Figure 7, the information adjustment device 5 May include: processor 51, memory 52 and communication bus 53, wherein:
  • Memory 52 used to store executable instructions
  • Communication bus 53 for realizing the communication connection between processor 51 and memory 52;
  • the processor 51 is configured to execute the information adjustment program stored in the memory 52, so as to realize the following steps:
  • the value of the current limiting parameter of each first data transmission interface is adjusted to a corresponding to-be-adjusted current limiting threshold.
  • the processor when the processor executes the step of determining at least one first data transmission interface that is currently in the current limiting mode, it may be implemented through the following steps:
  • From the at least one second data transmission interface determine the second data transmission interface whose current request quantity value is greater than the corresponding first current current limiting threshold, to obtain at least one first data transmission interface.
  • the processor executes the step of analyzing the CPU occupancy rate of the central processing unit for at least one first current limiting threshold, and when determining the corresponding current limiting threshold of each first data transmission interface to be adjusted, it can be determined by follows these steps to achieve:
  • the target neural network model is used to perform CPU occupancy rate prediction processing, and each corresponding first data transmission interface is determined.
  • the processor when the processor executes the step of determining the preset step value corresponding to each first data transmission interface, it may be implemented through the following steps:
  • At least one first data transmission interface includes a first data transmission interface, determine that the preset step value corresponding to the first data transmission interface is a unit step value;
  • At least one first data transmission interface includes at least two first data transmission interfaces, determining a proportional relationship between the first current limiting thresholds of the at least two first data transmission interfaces;
  • a preset step value corresponding to each first data transmission interface is determined.
  • the processor executes the step based on the first current limit threshold of at least one first data transmission interface and the corresponding preset step value of each first data transmission interface, using the target neural network model to perform CPU usage prediction processing, when determining the current limiting threshold to be adjusted for each corresponding first data transmission interface, can be achieved through the following steps:
  • the target neural network model is used to predict the CPU usage of the at least one first reference current limiting threshold, and obtain a first predicted occupancy rate
  • the current-limiting threshold to be adjusted for each first data transmission interface is a corresponding first reference current-limiting threshold.
  • the processor is further configured to perform the following steps:
  • the current-limiting threshold to be adjusted for each first data transmission interface is the corresponding updated first reference current-limiting threshold; wherein, the updated first The reference current limiting threshold is the sum of n times the preset step value corresponding to the corresponding first data transmission interface and the corresponding first current limiting threshold, where n is a positive integer greater than or equal to 1.
  • the processor is further configured to perform the following steps:
  • At least one third data transmission interface is at least one second data transmission interface divided by at least one first A data transfer interface other than a data transfer interface
  • the processor is further configured to perform the following steps:
  • the processor executes the step of adjusting the value of the current limiting parameter of each first data transmission interface to the corresponding current limiting threshold to be adjusted, it is further used to perform the following steps:
  • the processor executes the step of adjusting the value of the current limiting parameter of each first data transmission interface to the corresponding current limiting threshold to be adjusted, it is further used to perform the following steps:
  • Obtaining the value of the current limiting parameter of each first data transmission interface is the first actual occupancy rate of the corresponding CPU when the current limiting threshold is to be adjusted;
  • the value of the current limiting parameter of each first data transmission interface is determined to be the current limiting threshold to be adjusted, the corresponding target predicted occupancy rate obtained through the prediction of the target neural network model;
  • the target neural network model is updated based on the target predicted occupancy rate and the first actual occupancy rate.
  • the processor after the processor performs the step of determining the current request quantity value of each second data transmission interface, it is further used to perform the following steps:
  • the current request quantity value of each second data transmission interface is less than the corresponding first current limiting threshold, determine the product of each first current limiting threshold and the preset ratio, Obtaining a corresponding feedback current limiting threshold; wherein, the feedback current limiting threshold is smaller than the first current current limiting threshold;
  • the current request quantity value of each second data transmission interface is less than the corresponding feedback current limit threshold, adjust the value of the current limit parameter of each second data transmission interface to the initialization limit flow threshold.
  • the current limit threshold after determining at least one first data transmission interface that is currently in the current limiting mode, determine the first current limiting threshold corresponding to the current limiting parameter of each first data transmission interface, and obtain at least one first current limiting threshold. flow threshold, and analyze the CPU occupancy rate of at least one first current current limiting threshold to determine the corresponding current limiting threshold of each data transmission interface to be adjusted, so as to adjust the current limiting parameter value of each first data transmission interface.
  • the current limit parameter value is dynamically adjusted, which effectively improves the flexibility and adjustment efficiency of the current limit configuration process.
  • the embodiments of the present application provide a computer-readable storage medium, referred to as a storage medium for short, where one or more programs are stored in the computer-readable storage medium, and the one or more programs can be used by one or more
  • the processor executes to implement the implementation process of the information adjustment method provided in the embodiments corresponding to FIGS. 1 to 5 , which will not be repeated here.
  • An embodiment of the present application provides an information adjustment method, device, and storage medium.
  • the method includes: determining at least one first data transmission interface that is currently in the current limiting mode; determining the current limiting parameter corresponding to each of the first data transmission interfaces The first current limiting threshold value of the first current limiting threshold value, at least one first current limiting threshold value is obtained; CPU utilization rate analysis of the central processing unit is performed on at least one of the first current limiting threshold value, and each corresponding first data transmission interface is determined the current limiting threshold to be adjusted; adjust the value of the current limiting parameter of each of the first data transmission interfaces to the corresponding current limiting threshold to be adjusted, thus solving the problem of low flexibility in the current limiting configuration process Problem, a realization method of intelligent automatic adjustment of the current limit threshold has been implemented. Through the real-time transmission of all data transmission interfaces, the current limit parameter value of the data transmission interface is dynamically adjusted, which effectively improves the flexibility and flexibility of the current limit configuration process. Regulatory efficiency.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Feedback Control In General (AREA)

Abstract

本申请实施例公开了一种信息调节方法，该方法包括：确定当前处于限流模式的至少一个第一数据传输接口；确定每一所述第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值；对至少一个所述第一当前限流阈值进行中央处理器CPU占用率分析，确定对应的每一所述第一数据传输接口的待调整限流阈值；调整每一所述第一数据传输接口的所述限流参数的值为对应的所述待调整限流阈值。本申请实施例还公开了一种信息调节设备和存储介质。

Description

一种信息调节方法、设备及存储介质
相关申请的交叉引用
本申请基于申请号为202110624327.6、申请日为2021年6月4日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本申请作为参考。
技术领域
本申请涉及人工智能领域，尤其涉及一种信息调节方法、设备及存储介质。
背景技术
随着计算机技术的飞速发展，越来越多的技术应用在金融领域，传统金融业正在逐步向金融科技(Fintech)转变，但由于金融行业的安全性和实时性要求，也对技术提出了更高的要求。在通信过程中，为了保证为用户提供较好的业务服务，在业务流程实现过程中，通常采用限流的方式来防止突如其来的数据流量洪峰造成通信系统或通信下游系统资源耗尽而崩溃。目前常用的限流阈值配置过程中，通常采用单接口压测的方式测得单机单接口每秒处理的交易笔数(Transactions Per Second，TPS)的CPU占用率后，计算所有接口的目标TPS与单机单接口单TPS的CPU占用率的乘积后求累加和，然后以单机CPU不超过45％为基准线，得出单机器的单机限流配置。
但是，目前上述配置过程中，实现单口压测时，不能完全还原实际生产环境，导致单接口压测得出的单机单接口TPS的CPU占用率与实际生产环境中的实际CPU占用率有一定的误差，导致得到的单机限流配置误差较大，且设置的单机限流配置误差实时性较差。
发明内容
为解决上述技术问题，本申请实施例期望提供一种信息调节方法、设备及存储介质，解决了目前限流配置过程灵活性较低的问题，实现了一种智能自动调节限流阈值的实现方法，通过对全部数据传输接口的实时传输情况对数据传输接口的限流参数值进行动态调整，有效提高了限流配置过程的灵活性和调节效率。
本申请的技术方案是这样实现的：
第一方面，一种信息调节方法，所述方法包括：
确定当前处于限流模式的至少一个第一数据传输接口；
确定每一所述第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值；
对至少一个所述第一当前限流阈值进行中央处理器CPU占用率分析，确定对应的每一所述第一数据传输接口的待调整限流阈值；
调整每一所述第一数据传输接口的所述限流参数的值为对应的所述待调整限流阈值。
第二方面，一种信息调节设备，所述设备包括：存储器、处理器和通信总线；其中：
所述存储器，用于存储可执行指令；
所述通信总线，用于实现所述处理器和所述存储器之间的通信连接；
所述处理器，用于执行所述存储器中存储的信息调节程序，实现如上述任一项所述的信息调节方法的步骤。
第三方面，一种存储介质，所述存储介质上存储有信息调节程序，所述信息调节程序被处理器执行时实现如上述任一项所述的信息调节方法的步骤。
本申请实施例中，确定当前处于限流模式的至少一个第一数据传输接口后，确定每一第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值，并对至少一个第一当前限流阈值进行CPU占用率分析，来确定对应的每一数据传输接口的待调整限流阈值，以将每一第一数据传输接口的限流参数值调整为待调整限流阈值，解决了目前限流配置过程灵活性较低的问题，实现了一种智能自动调节限流阈值的实现方法，通过对全部数据传输接口的实时传输情况对数据传输接口的限流参数值进行动态调整，有效提高了限流配置过程的灵活性和调节效率。
附图说明
图1为本申请实施例提供的一种信息调节方法的流程示意图；
图2为本申请实施例提供的另一种信息调节方法的流程示意图；
图3为本申请实施例提供的又一种信息调节方法的流程示意图；
图4为本申请另一实施例提供的一种信息调节方法的流程示意图；
图5为本申请另一实施例提供的另一种信息调节方法的流程示意图；
图6为本申请实施例提供的一种实现信息调节方法的系统架构示意图；
图7为本申请实施例提供的一种信息调节设备的结构示意图。
具体实施方式
为了使本申请的目的、技术方案和优点更加清楚，下面将结合附图对本申请作进一步地详细描述，所描述的实施例不应视为对本申请的限制，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其它实施例，都属于本申请保护的范围。
除非另有定义，本文所使用的所有的技术和科学术语与属于本申请的技术领域的技术人员通常理解的含义相同。本文中所使用的术语只是为了描述本申请实施例的目的，不是旨在限制本申请。
本申请的实施例提供一种信息调节方法，参照图1所示，方法应用于信息调节设备，该方法包括以下步骤：
步骤101、确定当前处于限流模式的至少一个第一数据传输接口。
在本申请实施例中，第一数据传输接口通常指的是信息调节设备提供的与外部互联网系统进行通信的接口，用于接收外部应用发送的请求信息。限流模式指的是数据传输接口的当前请求数量超过数据传输接口处理请求的上限阈值时，对数据传输接口进行限流的一种方式，限流的方式例如可以是使超出上限阈值的请求稍后处理，或将超出上限阈值的请求发送至同类的其他接口处理。这样，可以防止由于请求过多导致信息调节设备提供的服务系统崩溃的情况，有效保证了信息调节设备能够正常持续稳定的提供服务。信息调节设备通常可以是服务器设备，可以用于运行提供服务的服务系统，也可以是运行一些具有大量信息交互的应用系统的电子设备。
步骤102、确定每一第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值。
在本申请实施例中，每一第一数据传输接口的限流参数对应的第一当前限流阈值可以是预先设置的一个限流上限值。
步骤103、对至少一个第一当前限流阈值进行CPU占用率分析，确定对应的每一第一数据传输接口的待调整限流阈值。
在本申请实施例中，采用预设的分析方法对每一第一当前限流值进行CPU占用率分析，确定得到对应的每一数据传输接口的待调整限流阈值，其中，预设的分析方法可以是已训练好的目标神经网络模型，也可以是一些已经确定好的权重系数的线性算法，例如可以是已确定权重系数的矩阵运算算法、多项式回归算法或岭回归算法等。
步骤104、调整每一第一数据传输接口的限流参数的值为对应的待调整限流阈值。
在本申请实施例中，由于每一第一数据传输接口当前处于限流模式，因此，对每一第一数据传输接口的限流参数的值进行相应的动态调整，其中，调整后的待调整限流阈值能够使信息调节设备的CPU的占用率不超过CPU的预设占用率，从而保证信息调节设备的运行效率。
本申请实施例中，确定当前处于限流模式的至少一个第一数据传输接口后，确定每一第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值，并对至少一个第一当前限流阈值进行CPU占用率分析，来确定对应的每一数据传输接口的待调整限流阈值，以将每一第一数据传输接口的限流参数值调整为待调整限流阈值，解决了目前限流配置过程灵活性较低的问题，实现了一种智能自动调节限流阈值的实现方法，通过对全部数据传输接口的实时传输情况对数据传输接口的限流参数值进行动态调整，有效提高了限流配置过程的灵活性和调节效率。
基于前述实施例，本申请的实施例提供一种信息调节方法，参照图2所示，该方法应用于信息调节设备，该方法包括以下步骤：
步骤201、确定当前时刻处于数据传输模式的至少一个第二数据传输接口。
在本申请实施例中，信息调节设备可以是银行提供银行各种业务服务的服务器设备，数据传输模式指的是接口存活且接口能够正常传输数据信息的模式。至少一个第二数据传输接口指的是信息调节设备当前用于提供服务的存活的所有数据传输接口。示例性的，假设信息调节设备确定当前时刻处于数据传输模式的至少一个第二数据传输接口为接口1、接口2、接口3、接口4和接口5。
步骤202、确定每一第二数据传输接口的限流参数对应的第一当前限流阈值。
在本申请实施例中，信息调节设备确定每一第二数据传输接口针对其限流参数设置的第一当前限流阈值。示例性的，确定得到接口1的第一当前限流阈值为a1、接口2的第一当前限流阈值为a2、接口3的第一当前限流阈值为a3、接口4的第一当前限流阈值为a4和接口5的第一当前限流阈值为a5。
步骤203、确定每一第二数据传输接口的当前请求数量值。
在本申请实施例中，信息调节设备确定至少一个第二数据传输接口包括的每一第二数据传输接口的当前请求数量值，得到至少一个当前请求数量值。示例性的，确定得到接口1的当前请求数量值为q1、接口2的当前请求数量值为q2、接口3的当前请求数量值为q3、接口4的当前请求数量值为q4和接口5的当前请求数量值为q5。
步骤204、从至少一个第二数据传输接口中，确定当前请求数量值大于对应的第一当前限流阈值的第二数据传输接口，得到至少一个第一数据传输接口。
在本申请实施例中，将每一第二数据传输接口的当前请求数量值与其对应的第一当前限流阈值进行比较，得到当前请求数量值大于对应的第一当前限流阈值的第二数据传输接口，并将当前请求数量值大于对应的第一当前限流阈值的第二数据传输接口确定为至少一个第一数据传输接口。其中，至少一个第一数据传输接口包括一个第一数据传输接口的情况和至少两个第一数据传输接口的情况。
需说明的是，针对至少一个第二数据传输接口，存在每一当前请求数量值均小于对应的第一当前限流阈值的情况，即一个第一数据传输接口也确定不到，也就是说无需对至少一个第二数据传输接口进行限流操作。
步骤205、确定每一第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值。
在本申请实施例中，每一第一数据传输接口对应的第一当前限流阈值可以是直接从确定的至少一个第二数据传输接口对应的第一当前限流阈值中获取到的。示例性的，假设确定的至少一个第一数据传输接口为至少一个第二数据传输接口即接口1、接口2、接口3、接口4和接口5中的接口2、接口4和接口5，对应的可以确定至少一个第一数据传输接口包括的每一第一数据传输接口的限流参数对应的第一当前限流阈值依次为：接口2对应的第一当前限流阈值为a2、接口4对应的第一当前限流阈值为a4和接口5对应的第一当前限流阈值为a5。
步骤206、对至少一个第一当前限流阈值进行CPU占用率分析，确定对应的每 一第一数据传输接口的待调整限流阈值。
在本申请实施例中，对接口2对应的第一当前限流阈值为a2、接口4对应的第一当前限流阈值为a4和接口5对应的第一当前限流阈值为a5进行分析，确定得到接口2对应的待调整限流阈值，接口4对应的待调整限流阈值和接口5对应的待调整限流阈值。在确定每一第一数据传输接口的待调整限流阈值时，可以是对对应的每一第一数据传输接口的第一当前限流阈值进行分别独立分析确定得到的。但在一些应用场景下，由于每一数据传输接口的限流阈值的调整均会对信息调节设备的CPU占用率造成较大的影响，因此，在调整每一第一数据传输接口的限流参数的值时，可能需要对信息调节设备的全部数据传输接口的限流参数的值进行全局分析，来确定每一第一数据传输接口的待调整限流阈值。
步骤207、调整每一第一数据传输接口的限流参数的值为对应的待调整限流阈值。
在本申请实施例中，信息调节设备确定每一第一数据传输接口的待调整限流阈值后，将每一第一数据传输接口的限流参数的值调整为对应的待调整限流阈值，而至少一个第二数据传输接口中除至少一个第一数据传输接口外的接口的限流参数的值则不进行调整。
需要说明的是，本实施例中与其它实施例中相同步骤和相同内容的说明，可以参照其它实施例中的描述，此处不再赘述。
本申请实施例中，确定当前处于限流模式的至少一个第一数据传输接口后，确定每一第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值，并对至少一个第一当前限流阈值进行CPU占用率分析，来确定对应的每一数据传输接口的待调整限流阈值，以将每一第一数据传输接口的限流参数值调整为待调整限流阈值，解决了目前限流配置过程灵活性较低的问题，实现了一种智能自动调节限流阈值的实现方法，通过对全部数据传输接口的实时传输情况对数据传输接口的限流参数值进行动态调整，有效提高了限流配置过程的灵活性和调节效率。
本申请的实施例提供一种信息调节方法，参照图3所示，该方法应用于信息调节设备，该方法包括以下步骤：
步骤301、确定当前时刻处于数据传输模式的至少一个第二数据传输接口。
步骤302、确定每一第二数据传输接口的限流参数对应的第一当前限流阈值。
步骤303、确定每一第二数据传输接口的当前请求数量值。
步骤304、从至少一个第二数据传输接口中，确定当前请求数量值大于对应的第一当前限流阈值的第二数据传输接口，得到至少一个第一数据传输接口。
步骤305、确定每一第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值。
步骤306、获取已训练好的目标神经网络模型。
在本申请实施例中，已训练好的目标神经网络模型可以是对反向传播(Back Propagation，BP)神经网络采用大量历史数据进行模型训练得到的。其中，历史数据包括每一数据传输接口的限流阈值，和在每一数据传输接口的限流阈值的前提下，对应的信息调节设备的CPU占用率。
其中，步骤306可以由以下步骤来实现：获取历史样本数据；对历史样本数据进行清洗，剔除误差数据，得到目标样本数据；采用目标样本数据对初始化神经网络模型进行模型训练，得到目标神经网络模型。
步骤307、确定每一第一数据传输接口的预设步进值。
在本申请实施例中，每一第一数据传输接口的预设步进值可以是预先设置的，也可以是根据实际情况进行计算得到的。
步骤308、基于至少一个第一数据传输接口的第一当前限流阈值、对应的每一第一数据传输接口的预设步进值，采用目标神经网络模型进行CPU占用率预测处理，确定对应的每一第一数据传输接口的待调整限流阈值。
在本申请实施例中，采用每一第一数据传输接口的预设步进值对对应的每一第一数据传输接口的第一当前限流阈值进行调整后，输入至目标神经网络模型的输入层，以便目标神经网络模型对基于每一第一数据传输接口的预设步进值对对应的第一数据传输接口的第一当前限流阈值调整后的知进行预测，然后根据预测结果，来确定对应的每一第一数据传输接口的待调整限流阈值。
步骤309、调整每一第一数据传输接口的限流参数的值为对应的待调整限流阈值。
基于前述实施例，在本申请其他实施例中，步骤307可以由步骤307a或307b～307c来实现；其中，在至少一个第一数据传输接口的数量为1时，选择执行步骤307a，在至少一个第一数据传输接口的数量大于或等于2时，选择执行步骤307b～307c：
步骤307a、若至少一个第一数据传输接口包括一个第一数据传输接口，确定第一数据传输接口对应的预设步进值为单位步进值。
在本申请实施例中，单位步进值可以是预先设置好的一个正整数值，例如单位步进值可以是1。
步骤307b、若至少一个第一数据传输接口包括至少两个第一数据传输接口，确定至少两个第一数据传输接口的第一当前限流阈值之间的比例关系。
在本申请实施例中，在至少一个第一数据传输接口的数量大于或等于2时，对至少一个第一数据传输接口的第一当前限流阈值进行比例计算，得到对应的比例关系。示例性的，在至少一个第一数据传输接口为接口2、接口4和接口5时，由于接口2对应的第一当前限流阈值为a2、接口4对应的第一当前限流阈值为a4和接口5对应的第一当前限流阈值为a5，因此，可以确定得到至少一个第一数据传输接口的第一当前限流阈值之间的比例关系可以记为接口2:接口4:接口5＝a2:a4:a5，假 设a2：a4：a5＝1:2:1。
步骤307c、基于比例关系和单位步进值，确定每一第一数据传输接口对应的预设步进值。
在本申请实施例中，假设单位步进值为1时，对应的根据比例关系a2：a4：a5＝1:2:1和单位步进值1，确定得到接口2的预设步进值为1，接口4的预设步进值为2，接口5的预设步进值为1。
基于前述实施例，在本申请其他实施例中，步骤308可以由步骤308a～308h，或步骤308a和308i～308p来实现；其中，若至少一个第一数据传输接口的第一接口数量与至少一个第二数据传输接口的第二接口数量相同，选择执行步骤308a～308h，若第二接口数量大于第一接口数量，步骤308a和308i～308p：
步骤308a、确定每一第一数据传输接口的第一当前限流阈值与对应的预设步进值的和值，得到至少一个第一参考限流阈值。
在本申请实施例中，在至少一个第一数据传输接口的数量等于1时，该第一数据传输接口的第一参考限流阈值＝第一数据传输接口的第一当前限流阈值+预设步进值，其中，预设步进值为单位步进值。示例性的，第一数据传输接口的第一参考限流阈值为第一当前限流阈值+1。
在至少一个第一数据传输接口的数量大于或等于2时，第一参考限流阈值的计算公式可以记为：第一参考限流阈值＝第一当前限流阈值+预设步进值，这样，接口2对应的第一参考限流阈值a21为a2+1，接口4对应的第一参考限流阈值a41为a4+2，接口5对应的第一参考限流阈值a51为a5+1。
步骤308b、若至少一个第一数据传输接口的第一接口数量与至少一个第二数据传输接口的第二接口数量相同，通过目标神经网络模型，对至少一个第一参考限流阈值进行中央处理器(Central Processing Unit，CPU)占用率预测，得到第一预测占用率。
在本申请实施例中，至少一个第一数据传输接口的第一接口数量与至少一个第二数据传输接口的第二接口数量相同，表明信息调节设备的全部数据传输接口的当前请求数量值均超过了其对应的第一当前限流阈值，需对全部数据传输接口的限流参数进行调节。此时，若至少一个第一数据传输接口的第一接口数量为1时，至少一个第二数据传输接口的第二数量也为1。将与至少一个第一数据传输接口对应的至少一个第一参考限流阈值输入至目标神经网络模型的输入层，通过目标神经网络模型预测在至少一个第一参考限流阈值的情况下，信息调节设备对应的中央处理器(Central Processing Unit，CPU)的占用率，得到第一预测占用率。
步骤308c、获取当前时刻CPU的当前占用率。
在本申请实施例中，当前时刻CPU的当前占用率为至少一个第二数据传输接口包括的每一第二数据传输数据传输接口的限流参数为第一当前限流阈值时，信息调节设备对应的实际CPU占用率。
步骤308d、确定第一预测占用率与当前占用率之间的第一差值。
在本申请实施例中，第一差值＝第一预测占用率-当前占用率。
其中，信息调整设备执行步骤308d之后，可以选择执行步骤308e，或者选择执行步骤308f～308h；其中，若第一差值大于或等于目标误差，选择执行步骤308e，若第一差值小于目标误差，选择执行步骤308f～308h：
步骤308e、若第一差值大于或等于目标误差，确定每一第一数据传输接口的待调整限流阈值为对应的第一参考限流阈值。
在本申请实施例中，目标误差为根据大量实验得到的一个误差经验值，或者是用户根据实际需求设定的一个误差经验值，在实际使用过程中，目标误差可以进行不断的校正调整。在第一差值大于或等于目标误差时，确定调整得到的每一第一参考限流阈值符合要求，因此，确定每一第一数据传输接口的待调整限流阈值为对应的第一参考限流阈值。
示例性的，至少一个第一数据传输接口的第一接口数量与至少一个第二数据传输接口的第二接口数量相同时，即接口1、接口2、接口3、接口4和接口5均处于限流模式时，假设至少一个第一数据传输接口对应的第一当前限流阈值分别为：接口1的第一当前限流阈值为a1、接口2的第一当前限流阈值为a2、接口3的第一当前限流阈值为a3、接口4的第一当前限流阈值为a4和接口5的第一当前限流阈值为a5，对应的，确定的接口1的预设步进值为b1、接口2的预设步进值为b2、接口3的预设步进值为b3、接口4的预设步进值为b4和接口5的预设步进值为b5，因此，确定得到的接口1的第一参考限流阈值为a11＝a1+b1、接口2的第一参考限流阈值为a21＝a2+b2、接口3的第一参考限流阈值为a31＝a3+b3、接口4的第一参考限流阈值为a41＝a4+b4和接口5的第一参考限流阈值为a51＝a5+b5，此时，若通过目标神经网络模型对a11、a21、a31、a41和a51进行预测后，得到的第一预测占用率与当前占用率的第一差值大于或等于目标误差时，确定接口1的待调整限流阈值为a11、接口2的待调整限流阈值为a21、接口3的待调整限流阈值为a31、接口4的待调整限流阈值为a41和接口5的待调整限流阈值为a51。
步骤308f、若第一差值小于目标误差，确定每一第一参考限流阈值与对应的预设步进值的和值，得到至少一个第二参考限流阈值。
在本申请实施例中，若第一差值小于目标误差，继续对每一第一参考限流阈值与对应的预设步进值进行求和处理，得到至少一个第二参考限流阈值。示例性的，对应的，可以确定接口1的第二参考限流阈值为a12＝a11+b1＝a1+2b1、接口2的第二参考限流阈值为a22＝a21+b2＝a2+2b2、接口3的第二参考限流阈值为a32＝a31+b3＝a3+2b3、接口4的第二参考限流阈值为a42＝a41+b4＝a4+2b4、和接口5的第二参考限流阈值为a52＝a51+b5＝a5+2b5。
步骤308g、更新至少一个第一参考限流阈值为至少一个第二参考限流阈值。
在本申请实施例中，将计算得到的接口1的第二参考限流阈值a12＝a1+2b1赋值 给接口1的第一参考限流阈值a11、接口2的第二参考限流阈值a22＝a2+2b2赋值给接口2的第一参考限流阈值a21、接口3的第二参考限流阈值a32＝a3+2b3赋值给接口3的第一参考限流阈值a31、接口4的第二参考限流阈值a42＝a4+2b4赋值给接口4的第一参考限流阈值a41、和接口5的第二参考限流阈值a52＝a5+2b5赋值给接口5的第一参考限流阈值a51。
步骤308h、返回执行“通过目标神经网络模型，对至少一个第一参考限流阈值进行CPU占用率预测，得到第一预测占用率”，直至循环n次后，在第一差值大于或等于目标误差的情况下，或者n大于或等于第一预设次数的情况下，确定每一第一数据传输接口的待调整限流阈值为对应的更新后的第一参考限流阈值。
其中，更新后的第一参考限流阈值为对应的第一数据传输接口对应的预设步进值的n倍与对应的第一当前限流阈值的和值，n为大于或等于1的正整数。
在本申请实施例中，通过目标神经网络模型，对更新后的接口1的a11、接口2的a21、接口3的a31、接口4的a41和接口5的a51进行预测，得到对应的第一预测占用率，然后确定该第一预测占用率与当前占用率之间的第一参考差值。
若第一参考差值大于或等于目标误差，确定每一第一数据传输接口的待调整限流阈值为对应的更新后的第一参考限流阈值，即若第一参考差值大于或等于目标误差，确定对接口1的待调整限流阈值为a11＝a1+2b1、接口2的待调整限流阈值为a21＝a2+2b2、接口3的待调整限流阈值为a31＝a3+2b3、接口4的待调整限流阈值为a41＝a4+2b4和接口5的待调整限流阈值为a51＝a5+2b5。
若第一参考差值小于目标误差，再次计算更新后的a11、a21、a31、a41和a51与各自对应的预设步进值的和值，得到接口1的第二参考限流阈值为a12＝a1+3b1、接口2的第二参考限流阈值为a22＝a2+3b2、接口3的第二参考限流阈值为a32＝a3+3b3、接口4的第二参考限流阈值为a42＝a4+3b4、和接口5的第二参考限流阈值为a52＝a5+3b5，然后将a12＝a1+3b1赋值给a11、a22＝a2+3b2赋值给a21、a32＝a3+3b3赋值给a31、a42＝a4+3b4赋值给a41、a52＝a5+3b5赋值给a51后，继续采用目标神经网络模型对更新后的a11、a21、a31、a41和a51进行CPU占用率预测处理，得到对应的预测占用率与当前占用率的差值，若对应的预测占用率与当前占用率的差值大于或等于目标差值，确定各接口对应的待调整限流阈值为更新后的a11、a21、a31、a41和a51，否则，继续重复执行前述步骤，直至更新后的a11、a21、a31、a41和a51通过目标神经网络模型预测得到的预测占用率与同一时间对应的当前占用率的差值大于或等于目标误差，或者得到更新后的a11、a21、a31、a41和a51的迭代次数超过第一预设次数时，结束操作。
其中，第一预设次数为根据大量实验得到的一个次数经验值，也可以是用户根据实际需求设定的一个次数经验值。
步骤308i、若第二接口数量大于第一接口数量，确定至少一个第三数据传输接口的第一当前限流阈值。
其中，至少一个第三数据传输接口为至少一个第二数据传输接口中除至少一个第一数据传输接口外的数据传输接口。
在本申请实施例中，第二接口数量大于第一接口数量，表明至少一个第一数据传输接口是至少一个第二数据传输接口中包括的部分数据传输接口，即只有部分传输接口处于限流模式时，确定信息调节设备中未处于限流模式的至少一个第三数据传输接口，并确定至少一个第三数据传输接口的第一当前限流阈值。其中，至少一个第二数据传输接口由至少一个第一数据传输接口与至少一个第三数据传输接口组成。
步骤308j、通过目标神经网络模型，对至少一个第一参考限流阈值和至少一个第三数据传输接口的第一当前限流阈值进行CPU占用率预测，得到第二预设占用率。
在本申请实施例中，由于至少一个第二数据传输接口当前正在传输数据，因此，至少一个第二数据传输接口的请求流量均对信息调节设备的CPU的占用率有影响，因此，在采用目标神经网络模型进行CPU的占用率预测时，需考虑至少一个第二数据传输接口的限流参数。因此，在通过目标神经网络模型确定至少一个第一数据传输接口的限流参数时，也需要将至少一个第三数据传输接口的第一当前限流阈值进行输入，来得到准确的至少一个第一数据传输接口的待调整限流阈值。
步骤308k、获取当前时刻CPU的当前占用率。
步骤308l、确定第二预设占用率与当前占用率之间的第二差值。
其中，信息调节设备执行步骤308l之后，可以选择执行步骤308m，或者选择执行步骤308n～308p；其中，若第二差值大于或等于目标误差，选择执行步骤308m，若第二差值小于或等于目标误差，选择执行步骤308n～308p：
步骤308m、若第二差值大于或等于目标误差，确定每一第一数据传输接口的待调整限流阈值为对应的第一参考限流阈值。
步骤308n、若第二差值小于目标误差，确定每一第一参考限流阈值与对应的预设步进值的和值，得到至少一个第二参考限流阈值。
步骤308o、更新至少一个第一参考限流阈值为至少一个第二参考限流阈值。
步骤308p、返回执行步骤308j，直至循环m次后，在第二差值大于或等于目标误差的情况下，或者m大于或等于第二预设次数的情况下，确定每一第一数据传输接口的待调整限流阈值为对应的更新后的第一参考限流阈值。
其中，更新后的第一参考限流阈值为对应的第一数据传输接口对应的预设步进值的m倍与对应的第一当前限流阈值的和值，m为大于或等于1的正整数。
基于前述实施例，在本申请其他实施例中，信息调节设备执行步骤308之后，还用于执行以下步骤：若检测到至少一个第二数据传输接口中存在处于限流模式的接口，返回执行“确定当前处于限流模式的至少一个第一数据传输接口”，直至至少一个第二数据传输接口中处于限流模式的接口的数量为零，或者将至少一个第二数 据传输接口中包括的更新后的至少一个第一数据传输接口对应的限流参数调整为对应的第一待调整限流阈值时，确定的CPU的第一实际占用率大于或等于预设占用率。
在本申请实施例中，不停地对每一数据传输接口的工作模式即是否处于限流模式进行检查，一旦检测到至少一个第二数据传输接口中存在处于限流模式的接口，则从步骤301开始重复执行，对处于限流模式的数据传输接口的限流参数进行调整，来保证数据传输的效率。
基于前述实施例，在本申请其他实施例中，参照图4所示，信息调节设备执行步骤309之后，还可以用于执行步骤310～312：
步骤310、获取每一第一数据传输接口的限流参数的值为待调整限流阈值时对应的CPU的第一实际占用率。
在本申请实施例中，信息调节设备将至少一个第一数据传输接口的限流参数的值调整为对应的待调整限流阈值后，由于信息调节设备的数据传输接口仍然在接收服务请求，以提供对应的服务，因此，信息调节设备可以采集到在调整每一第一数据传输接口的限流参数的值为对应的待调整限流阈值的情况下，对应的信息调节设备的CPU的第一实际占用率。
步骤311、确定每一第一数据传输接口的限流参数的值为待调整限流阈值时，通过目标神经网络模型预测得到的对应的目标预测占用率。
在本申请实施例中，在每一第一数据传输接口的限流参数的值为待调整限流阈值时对应的目标预测占用率可以根据前述实现过程中确定得到，若第一差值大于或等于目标误差，可以确定目标预测占用率为第一预测占用率；若第二差值大于或等于目标误差，可以确定目标预测占用率为第二预测占用率。
步骤312、基于目标预测占用率和第一实际占用率，更新目标神经网络模型。
在本申请实施例中，信息调节设备根据调整至少一个第一数据传输接口后得到的第一实际占用率，进行反馈处理，对目标神经网络模型进行更新调节，以使目标神经网络模型更加符合实际应用场景，保证目标神经网络模型的准确性，有效保证了预测过程的可靠性。
基于前述实施例，在本申请其他实施例中，参照图5所示，信息调节设备执行步骤303之后，还用于执行步骤313～314：
步骤313、若检测到至少一个第二数据传输接口中，每一第二数据传输接口的当前请求数量值小于对应的第一当前限流阈值，确定每一第一当前限流阈值与预设比值的乘积，得到对应的反馈限流阈值。
其中，反馈限流阈值小于第一当前限流阈值。
在本申请实施例中，预设比值是一个根据大量实验得到的比值经验值，也可以是用户根据实际需求进行设定的一个比值经验值。若信息调节设备包括的至少一个第二数据传输接口均无需进行限流处理，确定每一第一当前限流阈值与预设比值的 乘积，得到对应的反馈限流阈值。示例性的，在至少一个第二数据传输接口对应的至少一个第一当前限流阈值依次为：接口1的第一当前限流阈值为a1、接口2的第一当前限流阈值为a2、接口3的第一当前限流阈值为a3、接口4的第一当前限流阈值为a4和接口5的第一当前限流阈值为a5，对应的，预设比值为γ时，对应的针对接口1的反馈限流阈值记为γ*a1、接口2的反馈限流阈值记为γ*a2、接口3的反馈限流阈值记为γ*a3、接口4的反馈限流阈值记为γ*a4和接口5的反馈限流阈值记为γ*a5。
步骤314、若检测到至少一个第二数据传输接口中，每一第二数据传输接口的当前请求数量值均小于对应的反馈限流阈值，调整每一第二数据传输接口的限流参数的值为初始化限流阈值。
在本申请实施例中，初始化限流阈值为信息调节设备针对每一第二数据传输接口设置的默认限流阈值，可以是在信息调节设备初始设置时，用户自行设备的，也可以是信息调节设备的系统自己默认设置的。示例性的，由于接口1的当前请求数量值为q1、接口2的当前请求数量值为q2、接口3的当前请求数量值为q3、接口4的当前请求数量值为q4和接口5的当前请求数量值为q5，因此，在q1＜γ*a1、q2＜γ*a2、q3＜γ*a3、q4＜γ*a4且q5＜γ*a5时，调整每一第二数据传输接口的限流参数的值为初始化限流阈值。
基于前述实施例，本申请实施例提供一种实现信息调节方法的系统架构图，参照图6所示，至少包括：配置模块41、采集模块42、训练模块43、调整模块44、数据库45、远程字典服务(Remote Dictionary Server，Redis)46和应用服务器代理(Agent)47；其中：
配置模块41，用于向服务运行维护人员(简称运维人员)提供用于填写每一数据传输接口的配置信息及可调整标记位的填写接口。配置信息至少包括前述初始化限流阈值。假设有三个第二数据传输接口A、B和C，针对三个第二数据传输接口A、B和C的配置信息提交后，将针对三个第二数据传输接口A、B和C的配置信息写入数据库45中预先设置的用于记录每一第二数据传输接口的配置信息的配置表rate_config，将针对三个第二数据传输接口A、B和C的初始化限流阈值写入数据库45中预先设置的用于记录每一第二数据传输接口的初始化限流阈值的agent_upload_data_log表中，其中，每一第二数据传输接口的初始化限流阈值的命名方式可以记为“接口名_limit”。
需说明的是，在可调整标记位标记为可调整时，才执行本申请提供的信息调节方法，若可调整标记位标记为不可调整时，不执行本申请提供的信息调节方法。
数据库45中，还包括用于记录针对每一第二数据传输接口每一时间片接收到的请求数量值，每一第二数据传输接口每一时间片接收到的请求数量值的命名方式为记为“接口名_request_cnt”；用于记录每一时间片每一第二数据传输接口的限流阈值、每一第二数据传输接口的请求数量和CPU占用率的agent_upload_data_cal表。
采集模块42，用于接收应用服务器代理47传回的每一时间片的数据，可以是JavaScript对象简谱(JavaScript Object Notation，JSON)数据。示例性的，采集模块42接收到的JSON数据格式可以如下所示：
Figure PCTCN2021136434-appb-000001
采集模块42接收到上述数据后，对其进行解析并将解析得到的数据写入数据库45中的agent_upload_data_log数据表中。采集模块42将解析得到的数据写入数据库45中的agent_upload_data_log数据表中包括：创建异步任务一，异步任务一查询在该时间片的Redis 46中记录的各个第二数据请求接口的请求数量即前述当前请求数量值，创建异步任务时确保任务表中无对应的时间切片任务；创建异步任务二，首先，异步任务二分别计算该时间片针对A数据传输接口的被限流值A_limit_sum、B数据传输接口的被限流值B_limit_sum、C数据传输接口的限流值C_limit_sum，需说明的是，以A数据传输接口为例进行说明，A数据传输接口的被限流值A_limit_sum是根据A数据传输接口在该时间片的请求数量与A数据传输接口在该时间片时的限流阈值确定得到的，在A数据传输接口在该时间片的请求数量小于或等于A数据传输接口在该时间片时的限流阈值时，确定A_limit_sum为0，在A数据传输接口在该时间片的请求数量大于A数据传输接口在该时间片时的限流阈值时，确定A_limit_sum为A数据传输接口在该时间片的请求数量减去A数据传输接口在该时间片时的限流阈值的差值；其次，计算该时间片CPU占用率的平均值cpu_avg；然后，查询该时间片的Redis 46中的各个数据传输接口的请求数量得到A_request_cnt、B_request_cnt、C_request_cnt；最后，将time_stamp、A_limit_sum、B_limit_sum、C_limit_sum、cpu_avg、A_request_cnt、B_request_cnt、C_request_cnt写入到数据库45中预先设置的agent_upload_data_cal表中。
训练模块43，用于对BP初始化神经网络模型进行模型训练，得到目标神经网络模型，并对目标神经网络模型不断的进行模型更新。
训练模块43由定时任务驱动，例如可以是每小时训练一次。训练模块43具体 实现流程如下所示：
(一)数据选取及清洗
获取数据库45包括的agent_upload_data_cal表中时间片在一段时间例如是距离当前时刻两天内的所有数据，得到第一数据集。若agent_upload_data_cal表中包括的数据的时长不足两天，停止本次训练。
根据前述描述可知，A_limit_sum是A数据传输接口在某一时间片对应的被限流值、B_limit_sum是B数据传输接口在某一时间片对应的被限流值和C_limit_sum是C数据传输接口在某一时间片对应的被限流值，如果某个数据传输接口的被限流值大于0，代表在该时间片时该数据传输接口出现限流。限流是在应用底层做请求拒绝的，此时应用几乎不消耗资源，因此这部分数据需要剔除，否则影响计算准确性。这样，遍历第一数据集，判断每一组A_limit_sum、B_limit_sum和C_limit_sum中的至少一个被限流值是否存在大于0的情况，若存在至少一个数据传输接口的被限流值大于0，确定判断条件成立，将包括判断条件成立的每一组A_limit_sum、B_limit_sum和C_limit_sum从第一数据集中剔除，最终得到第二数据集。
在实际运行环境中，同一运行系统中的服务器设备的规格是一致的，处理的联机请求数量几乎相等，因此CPU利用率差额不超过5％。遍历第二数据集，以时间片time_stamp为索引，查询agent_upload_data_log中对应时间片是否存在该时间片某一服务器CPU占用率与该时间片所有服务器的平均CPU占用率cpu_avg的差值超过0.05的。如果存在某一服务器的CPU占用率超过平均CPU占用率cpu_avg，将该时间片该服务器的数据从第二数据集中剔除，得到第三数据集。选取第三数据集中A_request_cnt、B_request_cnt、C_request_cnt和cpu_avg即前述第一实际CPU占用率字段，形成新的第四数据集。
(二)BP神经网络模型训练
BP神经网络模型的输入变量个数为配置表rate_config中包括的数据传输接口的数量；隐含层数量至少为2，输出变量个数为1。示例性的，如前述JSON数据格式所示，共有三个接口，则BP神经网络模型输入变量个数为3，假设隐含层数量为3。
对BP神经网络模型的权重系数和偏置系数初始化过程：将第四数据集中的每一组A_request_cnt、B_request_cnt和C_request_cnt输入至BP神经网络模型的输入层；期望输出：与第四数据集中每一组A_request_cnt、B_request_cnt和C_request_cnt对应的cpu_avg；训练输出为输入A_request_cnt、B_request_cnt、C_request_cnt至BP神经网络模型的输入层后，BP神经网络模型的输出层输出预测CPU占用率Y_cpu_avg；计算误差Err＝cpu_avg-Y_cpu_avg，根据误差Err来更新BP神经网络模型的权重系数与偏置系数，直至更新后的BP神经网络模型计算得到的预测CPU占用率与对应的cpu_avg之间的误差小于或等于5％时，确定此时对应的更新后的BP神经网络模型为前述目标神经网络模型。
调整模块44，用于基于负反馈原理，自动灰度调整每一数据传输接口的限流阈值配置。
(1)调整模块44检查agent_upload_data_cal最近一条服务器代理46上报的数据，检查各个数据传输接口的被限流值是否大于0，即各个数据传输接口的数据请求数量是否大于各个数据传输接口的限流阈值。若存在数据传输接口的被限流值大于0，限流阈值调整任务启动，获取目标神经网络模型作为本次调整的运营模型，针对不同数据传输接口出现限流的情况，可以分为单接口出现被限流和多接口出现限流，采用以下两种调整方式：
针对单接口出现被限流的情况：假设B数据传输接口出现限流，首先，循环将被限流的B数据传输接口的原配置的限流阈值增加单位步进值例如1来更新B_request_cnt，A数据传输接口A_request_cnt为原配置的限流阈值，C数据传输接口C_request_cnt为原配置的限流阈值，将A_request_cnt、更新后的B_request_cnt和C_request_cnt输入至目标神经网络模型中，得到第一预测CPU占用率，同时计算第一预测CPU占用率与当前时间片的cpu_avg之间的差值，若差值大于或等于0.05，结束；否则，重复上述步骤，不断调整B数据传输接口的B_request_cnt，直至通过目标神经网络模型预测得到的第三预测CPU占用率与cpu_avg之间的差值大于0.05，循环终止，并确定更新后的B_request_cnt作为B数据传输接口的最新限流阈值，然后，将更新后的B_request_cnt的最新限流阈值下发配置至各个应用；最后，将调整后的配置、对应的预测CPU占用率存储于数据库45。
针对多接口出现被限流的情况：假设A数据传输接口和B数据传输接口同时出现限流时，首先，计算A数据传输接口和B数据传输接口的初始限流阈值的比例，假设A：B＝1:3，循环开始，每次循环将被限流的数据传输接口按比例增加预设步进值，假设单位步进值为1，则A数据传输接口原配置的限流阈值增加1作为A_request_cnt，则B数据传输接口原配置的限流阈值增加3作为B_request_cnt，C数据传输接口的C_request_cnt为原配置的限流阈值，这样，将更新后的A_request_cnt、B_request_cnt和C_request_cnt输入值目标神经网络模型中，得到第一预测CPU占用率，同时计算预测CPU占用率与当前时间片的cpu_avg差值，这样，不断的采用A数据传输接口的预设步进值和B数据传输接口的预设步进值进行调整，进行循环处理，直至预测到的第三预测CPU占用率与cpu_avg之间的差值大于或等于0.05，或迭代次数达到预设次数例如10000，循环终止，从而可以得到最新的A数据传输接口对应的限流阈值和B数据传输接口对应的限流阈值。然后将最新的A数据传输接口对应的限流阈值和B数据传输接口对应的限流阈值下发配置至各个应用，最后，将调整后的配置、对应的第三预测CPU占用率存储于数据库45。
(2)调整模块44将更新后的配置下发至各应用后，在下一时间周期例如从当前周期开始的5秒后，采集应用服务器代理47上报的最新数据，得到此次更新后的配置对应的第一实际CPU占用率，并计算第一实际CPU占用率与对应的第三预测 CPU占用率之间的误差，并将该误差负反馈于目标神经网络模型，对目标神经网络模型进行模型训练更新，以得到更新后的目标神经网络模型。
(3)调整模块44检查(2)中应用服务器代理47上报至agent_upload_data_cal表的数据，检查各个数据传输接口的被限流值是否大于0。如果依旧存在有数据传输接口的被限流值大于0，则仍存在限流，可以重复(1)和(2)，直至(2)中采集的最新实际CPU占用率超过75％或不存在被限流的数据传输接口，则调整模块44调整任务终止，并初始化监测任务。
(4)调整模块44还用于对agent_upload_data_cal最近一条应用服务器代理47上报的数据进行检测时，若检测到各个数据传输接口的请求数量值例如A_request_cnt、B_request_cnt和C_request_cnt均小于各自对应的当前限流阈值的一半即前述预设比值为50％时，将各个数据传输接口的限流阈值恢复至配置表rate_config中的初始化限流阈值。
这样，实现负反馈限流阈值调整，能够在出现限流时临时使用生产空闲的25％～30％的CPU资源，有效降低限流的影响，为扩容争取时间。在一些应用场景中，为了保证运行的稳定，满足业务系统的容灾要求，通常控制CPU实际占用率在45％左右。若由于存在限流的情况进行对各个数据传输接口的限流阈值调高后，导致CPU实际占用率升高，若长时间以各个数据传输接口的较高限流阈值运行时，在后期存在请求量快速上涨时，存在部分请求处理过慢的现象。为了避免此种风险，本申请还提出在检测到各个数据传输接口的请求数量值均小于各个数据传输接口的当前限流阈值的一半时，将各个数据传输接口的限流参数的值恢复至配置表rate_config中的初始化限流阈值。
需要说明的是，本实施例中与其它实施例中相同步骤和相同内容的说明，可以参照其它实施例中的描述，此处不再赘述。
本申请实施例中，确定当前处于限流模式的至少一个第一数据传输接口后，确定每一第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值，并对至少一个第一当前限流阈值进行CPU占用率分析，来确定对应的每一第一数据传输接口的待调整限流阈值，以将每一第一数据传输接口的限流参数值调整为待调整限流阈值，解决了目前限流配置过程灵活性较低的问题，实现了一种智能自动调节限流阈值的实现方法，通过对全部数据传输接口的实时传输情况对数据传输接口的限流参数值进行动态调整，有效提高了限流配置过程的灵活性和调节效率。
基于前述实施例，本申请的实施例提供一种信息调节设备，该信息调节设备可以应用于图1～5对应的实施例提供的信息调节方法中，参照图7所示，该信息调节设备5可以包括：处理器51、存储器52和通信总线53，其中：
存储器52，用于存储可执行指令；
通信总线53，用于实现处理器51和存储器52之间的通信连接；
处理器51，用于执行存储器52中存储的信息调节程序，以实现以下步骤：
确定当前处于限流模式的至少一个第一数据传输接口；
确定每一第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值；
对至少一个第一当前限流阈值进行中央处理器CPU占用率分析，确定对应的每一第一数据传输接口的待调整限流阈值；
调整每一第一数据传输接口的限流参数的值为对应的待调整限流阈值。
在本申请其他实施例中，处理器执行步骤确定当前处于限流模式的至少一个第一数据传输接口时，可以通过以下步骤来实现：
确定当前时刻处于数据传输模式的至少一个第二数据传输接口；
确定每一第二数据传输接口的限流参数对应的第一当前限流阈值；
确定每一第二数据传输接口的当前请求数量值；
从至少一个第二数据传输接口中，确定当前请求数量值大于对应的第一当前限流阈值的第二数据传输接口，得到至少一个第一数据传输接口。
在本申请其他实施例中，处理器执行步骤对至少一个第一当前限流阈值进行中央处理器CPU占用率分析，确定对应的每一第一数据传输接口的待调整限流阈值时，可以通过以下步骤来实现：
获取已训练好的目标神经网络模型；
确定每一第一数据传输接口的预设步进值；
基于至少一个第一数据传输接口的第一当前限流阈值、对应的每一第一数据传输接口的预设步进值，采用目标神经网络模型进行CPU占用率预测处理，确定对应的每一第一数据传输接口的待调整限流阈值。
在本申请其他实施例中，处理器执行步骤确定每一第一数据传输接口对应的预设步进值时，可以通过以下步骤来实现：
若至少一个第一数据传输接口包括一个第一数据传输接口，确定第一数据传输接口对应的预设步进值为单位步进值；
若至少一个第一数据传输接口包括至少两个第一数据传输接口，确定至少两个第一数据传输接口的第一当前限流阈值之间的比例关系；
基于比例关系和单位步进值，确定每一第一数据传输接口对应的预设步进值。
在本申请其他实施例中，处理器执行步骤基于至少一个第一数据传输接口的第一当前限流阈值、对应的每一第一数据传输接口的预设步进值，采用目标神经网络模型进行CPU占用率预测处理，确定对应的每一第一数据传输接口的待调整限流阈值时，可以通过以下步骤来实现：
确定每一第一数据传输接口的第一当前限流阈值与对应的预设步进值的和值，得到至少一个第一参考限流阈值；
若至少一个第一数据传输接口的第一接口数量与至少一个第二数据传输接口的 第二接口数量相同，通过目标神经网络模型，对至少一个第一参考限流阈值进行CPU占用率预测，得到第一预测占用率；
获取当前时刻CPU的当前占用率；
确定第一预测占用率与当前占用率之间的第一差值；
若第一差值大于或等于目标误差，确定每一第一数据传输接口的待调整限流阈值为对应的第一参考限流阈值。
在本申请其他实施例中，处理器还用于执行以下步骤：
若第一差值小于目标误差，确定每一第一参考限流阈值与对应的预设步进值的和值，得到至少一个第二参考限流阈值；
更新至少一个第一参考限流阈值为至少一个第二参考限流阈值；
返回执行“通过目标神经网络模型，对至少一个第一参考限流阈值进行CPU占用率预测，得到第一预测占用率”，直至循环n次后，在第一差值大于或等于目标误差的情况下，或者n大于或等于第一预设次数的情况下，确定每一第一数据传输接口的待调整限流阈值为对应的更新后的第一参考限流阈值；其中，更新后的第一参考限流阈值为对应的第一数据传输接口对应的预设步进值的n倍与对应的第一当前限流阈值的和值，n为大于或等于1的正整数。
在本申请其他实施例中，处理器还用于执行以下步骤：
若第二接口数量大于第一接口数量，确定至少一个第三数据传输接口的第一当前限流阈值；其中，至少一个第三数据传输接口为至少一个第二数据传输接口中除至少一个第一数据传输接口外的数据传输接口；
通过目标神经网络模型，对至少一个第一参考限流阈值和至少一个第三数据传输接口的第一当前限流阈值进行CPU占用率预测，得到第二预设占用率；
获取当前时刻CPU的当前占用率；
确定第二预设占用率与当前占用率之间的第二差值；
若第二差值大于或等于目标误差，确定每一第一数据传输接口的待调整限流阈值为对应的第一参考限流阈值。
在本申请其他实施例中，处理器还用于执行以下步骤：
若第二差值小于目标误差，确定每一第一参考限流阈值与对应的预设步进值的和值，得到至少一个第二参考限流阈值；
更新至少一个第一参考限流阈值为至少一个第二参考限流阈值；
返回执行“通过目标神经网络模型，对至少一个第一参考限流阈值和至少一个第三数据传输接口的第一当前限流阈值进行CPU占用率预测，得到第二预测占用率”，直至循环m次后，在第二差值大于或等于目标误差的情况下，或者m大于或等于第二预设次数的情况下，确定每一第一数据传输接口的待调整限流阈值为对应的更新后的第一参考限流阈值；其中，更新后的第一参考限流阈值为对应的第一数据传输接口对应的预设步进值的m倍与对应的第一当前限流阈值的和值，m为大于 或等于1的正整数。
在本申请其他实施例中，处理器执行步骤调整每一第一数据传输接口的限流参数的值为对应的待调整限流阈值之后，还用于执行以下步骤：
若检测到至少一个第二数据传输接口中存在处于限流模式的接口，返回执行“确定当前处于限流模式的至少一个第一数据传输接口”，直至至少一个第二数据传输接口中处于限流模式的接口的数量为零，或者将至少一个第二数据传输接口中包括的更新后的至少一个第一数据传输接口对应的限流参数调整为对应的第一待调整限流阈值后，确定的CPU的第一实际占用率大于或等于预设占用率。
在本申请其他实施例中，处理器执行步骤调整每一第一数据传输接口的限流参数的值为对应的待调整限流阈值之后，还用于执行以下步骤：
获取每一第一数据传输接口的限流参数的值为待调整限流阈值时对应的CPU的第一实际占用率；
确定每一第一数据传输接口的限流参数的值为待调整限流阈值时，通过目标神经网络模型预测得到的对应的目标预测占用率；
基于目标预测占用率和第一实际占用率，更新目标神经网络模型。
在本申请其他实施例中，处理器执行步骤确定每一第二数据传输接口的当前请求数量值之后，还用于执行以下步骤：
若检测到至少一个第二数据传输接口中，每一第二数据传输接口的当前请求数量值小于对应的第一当前限流阈值，确定每一第一当前限流阈值与预设比值的乘积，得到对应的反馈限流阈值；其中，反馈限流阈值小于第一当前限流阈值；
若检测到至少一个第二数据传输接口中，每一第二数据传输接口的当前请求数量值均小于对应的反馈限流阈值，调整每一第二数据传输接口的限流参数的值为初始化限流阈值。
需要说明的是，本申请实施例中一个或者多个程序可被一个或者多个处理器的步骤的解释说明，可以参照1～5对应的实施例提供的方法实现过程，此处不再赘述。
本申请实施例中，确定当前处于限流模式的至少一个第一数据传输接口后，确定每一第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值，并对至少一个第一当前限流阈值进行CPU占用率分析，来确定对应的每一数据传输接口的待调整限流阈值，以将每一第一数据传输接口的限流参数值调整为待调整限流阈值，解决了目前限流配置过程灵活性较低的问题，实现了一种智能自动调节限流阈值的实现方法，通过对全部数据传输接口的实时传输情况对数据传输接口的限流参数值进行动态调整，有效提高了限流配置过程的灵活性和调节效率。
基于前述实施例，本申请的实施例提供一种计算机可读存储介质，简称为存储介质，该计算机可读存储介质存储有一个或者多个程序，该一个或者多个程序可被一个或者多个处理器执行，以实现如图1～5对应的实施例提供的信息调节方法实现 过程，此处不再赘述。
以上，仅为本申请的实施例而已，并非用于限定本申请的保护范围。凡在本申请的精神和范围之内所作的任何修改、等同替换和改进等，均包含在本申请的保护范围之内。
工业实用性
本申请实施例提供一种信息调节方法、设备及存储介质，该方法包括：确定当前处于限流模式的至少一个第一数据传输接口；确定每一所述第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值；对至少一个所述第一当前限流阈值进行中央处理器CPU占用率分析，确定对应的每一所述第一数据传输接口的待调整限流阈值；调整每一所述第一数据传输接口的所述限流参数的值为对应的所述待调整限流阈值，这样，解决了目前限流配置过程灵活性较低的问题，实现了一种智能自动调节限流阈值的实现方法，通过对全部数据传输接口的实时传输情况对数据传输接口的限流参数值进行动态调整，有效提高了限流配置过程的灵活性和调节效率。

Claims (13)

  1. 一种信息调节方法，所述方法包括：
    确定当前处于限流模式的至少一个第一数据传输接口；
    确定每一所述第一数据传输接口的限流参数对应的第一当前限流阈值，得到至少一个第一当前限流阈值；
    对至少一个所述第一当前限流阈值进行中央处理器CPU占用率分析，确定对应的每一所述第一数据传输接口的待调整限流阈值；
    调整每一所述第一数据传输接口的所述限流参数的值为对应的所述待调整限流阈值。
  2. 根据权利要求1所述的方法，其中，所述确定当前处于限流模式的至少一个第一数据传输接口，包括：
    确定当前时刻处于数据传输模式的至少一个第二数据传输接口；
    确定每一所述第二数据传输接口的限流参数对应的第一当前限流阈值；
    确定每一所述第二数据传输接口的当前请求数量值；
    从至少一个所述第二数据传输接口中，确定所述当前请求数量值大于对应的所述第一当前限流阈值的第二数据传输接口，得到至少一个所述第一数据传输接口。
  3. 根据权利要求2所述的方法，其中，所述对至少一个所述第一当前限流阈值对中央处理器CPU占用率分析，确定对应的每一所述第一数据传输接口的待调整限流阈值，包括：
    获取已训练好的目标神经网络模型；
    确定每一所述第一数据传输接口的预设步进值；
    基于至少一个所述第一数据传输接口的所述第一当前限流阈值、对应的每一所述第一数据传输接口的所述预设步进值，采用所述目标神经网络模型进行CPU占用率预测处理，确定对应的每一所述第一数据传输接口的所述待调整限流阈值。
  4. 根据权利要求3所述的方法，其中，所述确定每一所述第一数据传输接口对应的预设步进值，包括：
    若至少一个所述第一数据传输接口包括一个所述第一数据传输接口，确定所述第一数据传输接口对应的预设步进值为单位步进值；
    若至少一个所述第一数据传输接口包括至少两个所述第一数据传输接口，确定至少两个所述第一数据传输接口的所述第一当前限流阈值之间的比例关系；
    基于所述比例关系和所述单位步进值，确定每一所述第一数据传输接口对应的所述预设步进值。
  5. 根据权利要求3或4所述的方法，其中，所述基于至少一个所述第一数据传输接口的所述第一当前限流阈值、对应的每一所述第一数据传输接口的所述预设步进值，采用所述目标神经网络模型进行CPU占用率预测处理，确定对应的每一所述第一数据传输接口的所述待调整限流阈值，包括：
    确定每一所述第一数据传输接口的所述第一当前限流阈值与对应的所述预设步进值的和值，得到至少一个第一参考限流阈值；
    若至少一个所述第一数据传输接口的第一接口数量与至少一个所述第二数据传输接口的第二接口数量相同，通过所述目标神经网络模型，对至少一个所述第一参考限流阈值进行CPU占用率预测，得到第一预测占用率；
    获取所述当前时刻CPU的当前占用率；
    确定所述第一预测占用率与所述当前占用率之间的第一差值；
    若所述第一差值大于或等于目标误差，确定每一所述第一数据传输接口的所述待调整限流阈值为对应的所述第一参考限流阈值。
  6. 根据权利要求5所述的方法，其中，所述方法还包括：
    若所述第一差值小于所述目标误差，确定每一所述第一参考限流阈值与对应的所述预设步进值的和值，得到至少一个第二参考限流阈值；
    更新所述至少一个第一参考限流阈值为所述至少一个第二参考限流阈值；
    返回执行所述“通过所述目标神经网络模型，对至少一个所述第一参考限流阈值进行CPU占用率预测，得到第一预测占用率”，直至循环n次后，在所述第一差值大于或等于所述目标误差的情况下，或者n大于或等于第一预设次数的情况下，确定每一所述第一数据传输接口的所述待调整限流阈值为对应的更新后的第一参考限流阈值；其中，更新后的所述第一参考限流阈值为对应的所述第一数据传输接口对应的预设步进值的n倍与对应的所述第一当前限流阈值的和值，n为大于或等于1的正整数。
  7. 根据权利要求5所述的方法，其中，所述方法还包括：
    若所述第二接口数量大于所述第一接口数量，确定至少一个第三数据传输接口的第一当前限流阈值；其中，至少一个所述第三数据传输接口为至少一个所述第二数据传输接口中除至少一个所述第一数据传输接口外的数据传输接口；
    通过所述目标神经网络模型，对至少一个所述第一参考限流阈值和至少一个所述第三数据传输接口的第一当前限流阈值进行CPU占用率预测，得到第二预测占用率；
    获取所述当前时刻CPU的当前占用率；
    确定所述第二预设占用率与所述当前占用率之间的第二差值；
    若所述第二差值大于或等于所述目标误差，确定每一所述第一数据传输接口的所述待调整限流阈值为对应的所述第一参考限流阈值。
  8. 根据权利要求7所述的方法，其中，所述方法还包括：
    若所述第二差值小于所述目标误差，确定每一所述第一参考限流阈值与对应的所述预设步进值的和值，得到至少一个第二参考限流阈值；
    更新所述至少一个第一参考限流阈值为所述至少一个第二参考限流阈值；
    返回执行所述“通过所述目标神经网络模型，对至少一个所述第一参考限流阈 值和至少一个所述第三数据传输接口的第一当前限流阈值进行CPU占用率预测，得到第二预测占用率”，直至循环m次后，在所述第二差值大于或等于所述目标误差的情况下，或者m大于或等于第二预设次数的情况下，确定每一所述第一数据传输接口的所述待调整限流阈值为对应的更新后的第一参考限流阈值；其中，更新后的所述第一参考限流阈值为对应的所述第一数据传输接口对应的预设步进值的m倍与对应的所述第一当前限流阈值的和值，m为大于或等于1的正整数。
  9. 根据权利要求3所述的方法，其中，所述调整每一所述第一数据传输接口的所述限流参数的值为对应的所述待调整限流阈值之后，所述方法还包括：
    若检测到至少一个所述第二数据传输接口中存在处于限流模式的接口，返回执行“确定当前处于限流模式的至少一个第一数据传输接口”，直至至少一个所述第二数据传输接口中处于限流模式的接口的数量为零，或者将至少一个所述第二数据传输接口中包括的更新后的至少一个第一数据传输接口对应的所述限流参数调整为对应的待调整限流阈值后，确定的CPU的第一实际占用率大于或等于预设占用率。
  10. 根据权利要求3所述的方法，其中，所述调整每一所述第一数据传输接口的所述限流参数的值为对应的所述待调整限流阈值之后，所述方法还包括：
    获取每一所述第一数据传输接口的所述限流参数的值为所述待调整限流阈值时对应的CPU的第一实际占用率；
    确定每一所述第一数据传输接口的所述限流参数的值为所述待调整限流阈值时，通过所述目标神经网络模型预测得到的对应的目标预测占用率；
    基于所述目标预测占用率和所述第一实际占用率，更新所述目标神经网络模型。
  11. 根据权利要求2所述的方法，其中，所述确定每一所述第二数据传输接口的当前请求数量值之后，所述方法还包括：
    若检测到至少一个所述第二数据传输接口中，每一所述第二数据传输接口的当前请求数量值小于对应的所述第一当前限流阈值，确定每一所述第一当前限流阈值与预设比值的乘积，得到对应的反馈限流阈值；其中，所述反馈限流阈值小于所述第一当前限流阈值；
    若检测到至少一个所述第二数据传输接口中，每一所述第二数据传输接口的当前请求数量值均小于对应的所述反馈限流阈值，调整每一所述第二数据传输接口的所述限流参数的值为初始化限流阈值。
  12. 一种信息调节设备，所述设备包括：存储器、处理器和通信总线；其中：
    所述存储器，用于存储可执行指令；
    所述通信总线，用于实现所述处理器和所述存储器之间的通信连接；
    所述处理器，用于执行所述存储器中存储的信息调节程序，实现如权利要求1至11中任一项所述的信息调节方法的步骤。
  13. 一种存储介质，所述存储介质上存储有信息调节程序，所述信息调节程序 被处理器执行时实现如权利要求1至11中任一项所述的信息调节方法的步骤。
PCT/CN2021/136434 2021-06-04 2021-12-08 一种信息调节方法、设备及存储介质 Ceased WO2022252546A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110624327.6A CN113326132B (zh) 2021-06-04 2021-06-04 一种信息调节方法、设备及存储介质
CN202110624327.6 2021-06-04

Publications (1)

Publication Number Publication Date
WO2022252546A1 true WO2022252546A1 (zh) 2022-12-08

Family

ID=77421052

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/136434 Ceased WO2022252546A1 (zh) 2021-06-04 2021-12-08 一种信息调节方法、设备及存储介质

Country Status (2)

Country Link
CN (1) CN113326132B (zh)
WO (1) WO2022252546A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120281718A (zh) * 2025-03-14 2025-07-08 四川省算云科技有限责任公司 一种基于强化学习的全局接口动态限流方法

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113326132B (zh) * 2021-06-04 2023-06-09 深圳前海微众银行股份有限公司 一种信息调节方法、设备及存储介质
CN114422439A (zh) * 2022-01-21 2022-04-29 平安国际智慧城市科技股份有限公司 接口限流方法、装置、计算机设备和存储介质
CN114553775B (zh) * 2022-02-15 2024-04-26 维沃移动通信有限公司 流量控制方法、装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107222426A (zh) * 2016-03-21 2017-09-29 阿里巴巴集团控股有限公司 控流的方法、装置及系统
CN109039937A (zh) * 2018-09-26 2018-12-18 深圳壹账通智能科技有限公司 动态限流方法、装置、计算机设备和存储介质
CN109842565A (zh) * 2018-12-15 2019-06-04 平安科技（深圳）有限公司 接口限流方法、装置、电子设备及存储介质
US20190182168A1 (en) * 2017-12-11 2019-06-13 International Business Machines Corporation Dynamic throttling thresholds
CN113326132A (zh) * 2021-06-04 2021-08-31 深圳前海微众银行股份有限公司 一种信息调节方法、设备及存储介质

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110198272B (zh) * 2018-10-12 2022-05-17 腾讯科技（深圳）有限公司 一种流量控制方法及装置、存储介质
CN112433919B (zh) * 2020-11-25 2023-01-24 深圳前海微众银行股份有限公司 一种信息告警方法、设备及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107222426A (zh) * 2016-03-21 2017-09-29 阿里巴巴集团控股有限公司 控流的方法、装置及系统
US20190182168A1 (en) * 2017-12-11 2019-06-13 International Business Machines Corporation Dynamic throttling thresholds
CN109039937A (zh) * 2018-09-26 2018-12-18 深圳壹账通智能科技有限公司 动态限流方法、装置、计算机设备和存储介质
CN109842565A (zh) * 2018-12-15 2019-06-04 平安科技（深圳）有限公司 接口限流方法、装置、电子设备及存储介质
CN113326132A (zh) * 2021-06-04 2021-08-31 深圳前海微众银行股份有限公司 一种信息调节方法、设备及存储介质

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120281718A (zh) * 2025-03-14 2025-07-08 四川省算云科技有限责任公司 一种基于强化学习的全局接口动态限流方法

Also Published As

Publication number Publication date
CN113326132B (zh) 2023-06-09
CN113326132A (zh) 2021-08-31

Similar Documents

Publication Publication Date Title
CN113326132B (zh) 一种信息调节方法、设备及存储介质
CN111459617B (zh) 基于云平台的容器化应用自动分配优化系统及其方法
TWI712900B (zh) 一種分布式集群訓練方法和裝置
US11784931B2 (en) Network burst load evacuation method for edge servers
US20210383197A1 (en) Adaptive stochastic learning state compression for federated learning in infrastructure domains
WO2017088717A1 (zh) 一种存储资源分配方法及存储资源分配系统
CN105893541B (zh) 一种基于混合存储的流式数据自适应持久化方法及系统
CN110046048B (zh) 一种基于工作量自适应快速重分配的负载均衡方法
CN112506657B (zh) 基于强化学习的面向微服务的资源管理系统
CN105677489B (zh) 离散流处理模型下批次间隔大小的动态设置系统及方法
CN115185683B (zh) 一种基于动态优化模型的云平台流处理资源分配方法
CN117473548A (zh) 基于通道的分层结构差分隐私联邦学习方法、装置
CN120186057A (zh) 分布式训练方法、装置、系统、设备和介质
CN109032630B (zh) 一种参数服务器中全局参数的更新方法
CN117971509A (zh) 异构算力集群的运行性能优化方法、装置、设备及介质
CN109560978B (zh) 网络流量检测方法、装置及系统和计算机可读存储介质
CN114841456A (zh) 交易系统的性能调节方法及装置、存储介质、电子设备
CN109688065A (zh) 参数的处理方法、装置及存储介质
CN115016928A (zh) 基于算子作用域动态感知的弹性资源分配方法及装置
JP7420228B2 (ja) 分散処理システムおよび分散処理方法
CN119440726A (zh) 异构计算集群的资源优化方法、装置、电子设备及介质
CN114297026B (zh) 自适应混沌环境的云存储服务超时值优化方法及相关设备
CN115913965A (zh) 一种基于任务时效性的关联切片资源分配方法
CN110134575B (zh) 一种服务器集群的服务能力计算方法及装置
CN112100652A (zh) 一种基于马尔科夫链的跨地理分布轨迹聚类方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21943898

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 25/01/2024)

122 Ep: pct application non-entry in european phase

Ref document number: 21943898

Country of ref document: EP

Kind code of ref document: A1