Interface adaptive scheduling method and device, equipment and medium

Abstract

The invention provides an interface adaptive scheduling method and device, equipment and a medium, and aims at solving the problem that a method provided by the related technology is not high in adaptability. The method comprises the following steps: receiving a service request; a target neural network is used for predicting the future trend of interface calling, a prediction result is obtained, and the target neural network is dynamically adjusted based on the real-time calling condition of the interface; determining a target scheduling strategy by evaluating the prediction result; the target scheduling strategy is used for adjusting the calling condition of each interface; and calling a corresponding interface for the service request through the target scheduling strategy.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to an interface adaptive scheduling method, device, equipment and medium. Background Art

[0002] With the rapid development of information technology, interface calls between various information systems and applications are becoming more and more frequent. Interface scheduling management has become a key link to ensure the efficient operation of the system. Especially in the big data and cloud computing environment, the amount of interface calls is huge, and the calling patterns are complex and changeable. However, traditional scheduling methods often rely on fixed algorithms and rules, and cannot adjust resource allocation in a timely and accurate manner, resulting in system performance degradation and poor user experience.

[0003] As a powerful machine learning method, deep neural networks have performed well in data processing, pattern recognition, and decision optimization. In recent years, deep neural networks have been widely used in various fields, such as natural language processing, image recognition, and intelligent recommendation.

[0004] In the related art, some neural network-based interface scheduling management methods have been proposed. However, most of them are based on static neural network models and are not adaptable to dynamically changing interface call patterns. When the interface call pattern changes significantly, the model cannot be adjusted in time, resulting in the failure of the scheduling strategy. Summary of the invention

[0005] In order to overcome the problems existing in the related art, the present disclosure provides an interface adaptive scheduling method, device, equipment and medium. The technical solution of the present disclosure is as follows: According to a first aspect of an embodiment of the present disclosure, there is provided an interface adaptive scheduling method, including: Receive business requests; Using a target neural network, predicting the future trend of interface calls to obtain a prediction result, wherein the target neural network is dynamically adjusted based on the real-time call status of the interface; By evaluating the prediction results, a target scheduling strategy is determined; the target scheduling strategy is used to adjust the calling conditions of each interface; The target scheduling strategy is used to call a corresponding interface for the service request.

[0006] Optionally, dynamically adjust the target neural network as follows: After executing the target scheduling strategy, each interface is monitored to obtain the real-time calling status of each interface; the real-time calling status of each interface includes the calling status of each interface and the performance data of each interface; the real-time calling status is used to evaluate the effectiveness of the target scheduling strategy; According to the real-time calling conditions of the various interfaces, the target neural network is adjusted to obtain an adjusted target neural network; wherein, adjusting the target neural network includes: fine-tuning the parameters of the target neural network, or retraining the target neural network.

[0007] Optionally, calling a corresponding interface for the service request through the target scheduling strategy includes: Through the target scheduling strategy, interface call management is performed; the interface call management includes: adjusting the load balance of each interface, adjusting the priority of each interface to use resources, and adjusting the calling order of each interface; By performing interface call management, an interface is called for the business request.

[0008] Optionally, when the interface call management is used to adjust the load balancing of each interface, calling the corresponding interface for the service request through the target scheduling strategy includes: Determine the predicted calling conditions of each interface; the calling conditions include high-concurrency requests and low-concurrency requests; For a target interface with high concurrent requests, determining a balancing interface corresponding to the target interface from interfaces with low concurrent requests; the balancing interface and the target interface can serve the same service request; Allocating requests corresponding to the target interface to the balancing interface; The balancing interface or the target interface is called for the service request.

[0009] Optionally, when the interface call management is used to adjust the priority of each interface in using resources, calling the corresponding interface for the service request through the target scheduling strategy includes: Determine the predicted call status of each interface and the importance of each interface; the call status includes high-concurrency requests and low-concurrency requests; When the calling condition of the interface is a high concurrent request, and / or when the priority of the interface is greater than or equal to a first threshold, according to the usage of the computing resources by the interface, computing resources are preferentially allocated to the interface; The interface after the computing resources are allocated is called for the business request.

[0010] Optionally, the interface call management is used to adjust the call order of each interface, and through the target scheduling strategy, the corresponding interface is called for the service request, including: Determine the priority of each service request corresponding to the interface; In the case where the priority of the service request indicates a priority process, calling the interface for the service request first; In the case where the priority of the service request does not indicate a priority process, the interface is called for the service request according to the time when the interface receives the service request.

[0011] Optionally, the target neural network is trained in the following manner, including: Collect historical data of interface calls; the historical data of interface calls includes interface name, call time, request parameters and response results; Preprocessing the historical data called by the interface to obtain a data set for training a neural network; Determine the initial neural network model; Customizing the initial neural network model according to the characteristics and requirements of the interface call to obtain a neural network model to be trained; Based on the purpose of learning the patterns and regularities of interface calls, the neural network model to be trained is trained through the data set to obtain a target neural network.

[0012] According to a second aspect of an embodiment of the present disclosure, there is provided an interface adaptive scheduling device, including: A receiving module, used for receiving a service request; A prediction module is used to use a target neural network to predict the future trend of interface calls and obtain a prediction result. The target neural network is dynamically adjusted based on the real-time call status of the interface. A determination module, used to determine a target scheduling strategy by evaluating the prediction results; the target scheduling strategy is used to adjust the calling conditions of each interface; The calling module is used to call the corresponding interface for the business request through the target scheduling strategy.

[0013] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of the interface adaptive scheduling method described in the first aspect are implemented.

[0014] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the interface adaptive scheduling method described in the first aspect are implemented.

[0015] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps of the interface adaptive scheduling method described in the first aspect are implemented.

[0016] The present disclosure introduces a target neural network that is dynamically adjusted based on the real-time call status of the interface, which can capture the dynamic changes of interface calls in real time, and can respond to the problem of changes in interface call patterns in a timely manner, thereby maintaining accurate predictions of future trends of interface calls, and has high flexibility and robustness. Through accurate predictions of future trends of interface calls, it is possible to use appropriate target scheduling strategies to manage the scheduling of each interface, optimize resource allocation for interface calls, reduce idle resources and waste, and thus improve the operating efficiency of the entire system. The present disclosure realizes the intelligentization of interface adaptive scheduling management by introducing advanced machine learning methods such as deep neural networks, and can more accurately capture the laws and trends of interface calls, so as to use more reasonable target scheduling strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for use in the description of the embodiments of the present disclosure will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 is a schematic diagram of the steps of an interface adaptive scheduling method shown in an embodiment of the present disclosure; Figure 2 is a schematic diagram of the steps of an interface adaptive scheduling method shown in an embodiment of the present disclosure; Figure 3 is a block diagram of an interface adaptive scheduling device shown in an embodiment of the present disclosure; Figure 4 It is a schematic diagram of an electronic device proposed in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0020] The terms "first", "second", etc. in the specification and claims of the present disclosure are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable when appropriate, so that the embodiments of the present disclosure can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally of the same type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0021] In order to solve the problems existing in the related art, the present disclosure proposes an interface adaptive scheduling method. The method can combine the adaptive characteristics of deep neural networks to improve the resource utilization rate and task response speed of the system.

[0022] Figure 1 Schematic diagram of the steps of an interface adaptive scheduling method shown in an embodiment of the present disclosure. Figure 1 As shown, the method may specifically include the following steps: Step S11: receiving a service request.

[0023] The system receives business requests from external users or other systems. Business requests include request parameters, request time, interface name, call frequency, user identity, priority, and other information.

[0024] The received business request information can be preprocessed, including data cleaning, format conversion, error detection, feature extraction, etc.

[0025] Step S12: Use the target neural network to predict the future trend of the interface call and obtain a prediction result. The target neural network is dynamically adjusted based on the real-time call status of the interface.

[0026] The target neural network is pre-trained using historical data of interface calls. When the target neural network is used to predict future trends of interface calls, it will be dynamically adjusted according to the real-time calling situation of the interface to adapt to changes in the interface calling pattern.

[0027] Use the trained neural network model to predict the trend of future interface calls and obtain prediction results. The prediction results include the frequency of interface calls, possible resource requirements, response time, error rate, etc. Possible resource requirements include CPU, memory, bandwidth, etc.

[0028] Step S13: Determine a target scheduling strategy by evaluating the prediction results; the target scheduling strategy is used to adjust the calling conditions of each interface.

[0029] Based on the prediction results output by the neural network model, further evaluation is performed to analyze the future resource requirements, call frequency, and possible load of each interface to obtain the evaluation results. Through the evaluation results, it can be determined which interfaces will become bottlenecks in the future and which interfaces may have excess resources.

[0030] Based on the evaluation results, the target scheduling strategy is determined by predicting the interface call trend. The scheduling strategy adjusts the interface call situation from multiple dimensions. For example, the interface call situation can be adjusted from the perspective of the interface call sequence. This can meet the goals of improving system performance, optimizing resource utilization, and ensuring the response speed of key interfaces.

[0031] The target scheduling strategy is preset, and according to the evaluation result, it is determined that the current interface call mode is the target scheduling strategy to be used among the preset scheduling strategies.

[0032] Step S14: calling a corresponding interface for the service request through the target scheduling strategy.

[0033] Execute the target scheduling strategy. When receiving a business request, determine the preset settings of the target scheduling strategy for each dimension. By matching the target scheduling strategy, call the corresponding interface for the business request.

[0034] By using the embodiments of the present disclosure, the system can prepare resources in advance and optimize the load distribution of the interface by predicting the trend of future interface calls, avoiding the situation where the interface is overloaded at certain times and the resources are idle at other times, and can significantly improve the overall efficiency of interface calls. The target neural network is dynamically adjusted based on the real-time call status of the interface, so the system can learn and adapt to changes in the interface call pattern in real time, and can flexibly respond to sudden traffic, seasonal changes, or changes in user behavior patterns.

[0035] Among them, in an optional embodiment, the target neural network is dynamically adjusted in the following manner: after executing the target scheduling strategy, each interface is monitored to obtain the real-time calling status of each interface; the real-time calling status of each interface includes the calling status of each interface and the performance data of each interface; the real-time calling status is used to evaluate the effectiveness of the target scheduling strategy; according to the real-time calling status of each interface, the target neural network is adjusted to obtain an adjusted target neural network; wherein, adjusting the target neural network includes: fine-tuning the parameters of the target neural network, or retraining the target neural network.

[0036] According to the interface call trend predicted by the target neural network, the corresponding target scheduling strategy is implemented.

[0037] After executing the scheduling strategy, each interface can be monitored to determine the real-time call status of each interface. The monitoring content should include the call status of the interface, such as success, failure, timeout, etc., and performance data, such as response time, throughput, error rate, etc. The real-time call status can be determined based on the data corresponding to each indicator included in the monitoring content.

[0038] By using the real-time call status obtained through monitoring, we can analyze whether the scheduling strategy has achieved the expected effect and obtain the analysis results. For example, we can check whether the response time is shortened, the error rate is reduced, and the throughput is increased.

[0039] If the analysis results determine that the scheduling strategy is effective, then the target neural network is judged to be able to accurately predict the future trend of interface calls, and the target neural network does not need to be adjusted at this time. If the analysis results determine that the scheduling strategy is invalid, it is necessary to use the data corresponding to the various indicators obtained from monitoring to fine-tune the target neural network or retrain the target neural network. Among them, if there is only a slight performance fluctuation or fine-tuning requirement, you can choose to fine-tune the parameters of the target neural network. If the scheduling strategy effect is significantly different from expectations, or the interface call pattern has changed significantly, the target neural network needs to be retrained.

[0040] If you choose the fine-tuning method, adjust the weights, biases and other parameters of the neural network based on the latest interface call data and performance data. The fine-tuning process can be carried out step by step, adjusting a small part of the parameters each time and observing the effect of the adjustment. If the prediction results of the fine-tuned target neural network are still inaccurate, you need to further adjust the neural network or reconsider the scheduling strategy.

[0041] If you choose to retrain, you need to collect a new training data set. The data set includes the latest interface call data and corresponding performance data. Use the new data set to train the neural network until the expected accuracy or convergence state is achieved.

[0042] By adopting the embodiments of the present disclosure, after executing the target scheduling strategy, by monitoring the real-time calling conditions of each interface, including calling status and performance data, the effectiveness of the scheduling strategy can be accurately evaluated, so that the system can continuously optimize the target neural network and improve the efficiency and performance of interface calling. The introduction of the neural network prediction and dynamic adjustment mechanism can dynamically adapt to the changes in the interface calling mode, so that the neural network model can continuously learn and improve its prediction accuracy, thereby better guiding interface scheduling.

[0043] Among them, in an optional embodiment, the target neural network is trained in the following manner, including: collecting historical data of interface calls; the historical data of interface calls include interface name, call time, request parameters and response results; preprocessing the historical data of interface calls to obtain a data set for training the neural network; determining an initial neural network model; customizing the initial neural network model according to the characteristics and requirements of the interface call to obtain a neural network model to be trained; based on the purpose of learning the patterns and rules of interface calls, the neural network model to be trained is trained through the data set to obtain a target neural network.

[0044] Collect historical data of interface calls from sources such as log systems, monitoring systems, or databases. Ensure that the data covers a sufficient time range to reflect long-term trends and patterns of interface calls.

[0045] Remove duplicate, invalid or abnormal data to ensure the accuracy and consistency of the data set. Check and process missing values ​​in the data, such as using interpolation, mean filling or deleting missing values. Convert data into a format suitable for neural network training, such as converting timestamps into time format, converting request parameters and response results into vector or matrix form, etc. Extract useful features from the raw data, which will serve as input for the deep neural network. Feature extraction can include statistical features, text features, image features, etc. Normalize or standardize numerical features to eliminate the impact of different dimensions on neural network training. Divide the preprocessed data set into training set, validation set and test set for subsequent model training and evaluation.

[0046] According to the characteristics and requirements of the interface call, select the appropriate neural network model type, such as long short-term memory network, convolutional neural network or recurrent neural network, etc. According to the requirements of the interface call, customize the structure of the deep neural network, such as determining the number of layers of the neural network, the number of neurons in each layer, the activation function, the optimization algorithm and other key parameters, which directly affect the performance and training effect of the neural network.

[0047] During the training process, the model parameters and structure are adjusted to minimize the prediction error and improve the accuracy of the model. During the training process, strategies such as adaptive learning rate and regularization are also used to prevent the model from overfitting. The adaptive learning rate can dynamically adjust the learning rate according to the training situation of the model to improve the training efficiency and accuracy. Regularization limits the complexity of the model by introducing additional constraints to prevent the model from overfitting during the training process.

[0048] By adopting the embodiments of the present disclosure, by collecting historical data of interface calls, including interface name, call time, request parameters and response results, a rich information basis is provided for the training of neural networks. Preprocessing the historical data of interface calls, such as data cleaning, formatting, normalization, etc., can ensure the quality and consistency of the data set, help improve the efficiency and accuracy of neural network training, and reduce noise interference in the training process. According to the characteristics and requirements of the interface call, the initial neural network model is customized to obtain a neural network model to be trained that is more suitable for practical applications, so that the neural network can better adapt to the complexity and diversity of interface calls. The method is based on neural network training and prediction, so it has strong scalability and maintainability. With the development of business and the increase in the number of interfaces, the neural network model can be easily expanded or the training strategy can be adjusted to meet new needs.

[0049] Among them, in an optional embodiment, the corresponding interface is called for the business request through the target scheduling strategy, including: interface call management is performed through the target scheduling strategy; the interface call management includes: load balancing adjustment of each interface, adjustment of the priority of resource use of each interface and adjustment of the calling order of each interface; through interface call management, the interface is called for the business request.

[0050] After determining the target scheduling strategy required for the current interface call mode, interface call management is performed using the target scheduling strategy.

[0051] Interface call management requires adjusting the load balancing of each interface, adjusting the priority of each interface in using resources, and adjusting the calling sequence of each interface.

[0052] After executing the interface call management, a suitable interface that will not be overloaded is called for the service request.

[0053] By adopting the embodiments of the present disclosure, by adjusting the load balance of each interface, it is ensured that the system can run smoothly under high concurrent requests, avoiding response delays or crashes of certain interfaces due to excessive load, and ensuring the effective use of resources. By adjusting the calling order of the interface, the business process can be optimized, unnecessary waiting time can be reduced, and the user experience can be improved, so that users can experience faster response speeds and smoother operation processes. Through the target scheduling strategy, the system can dynamically allocate resources to each interface according to current needs and resource conditions, avoiding waste and idleness of resources and improving resource utilization.

[0054] Among them, in an optional embodiment, when the interface call management is used to adjust the load balancing of each interface, the corresponding interface is called for the business request through the target scheduling strategy, including: determining the predicted call status of each interface; the call status includes high concurrency requests and low concurrency requests; for the target interface of high concurrency requests, determining the balancing interface corresponding to the target interface from the interfaces of low concurrency requests; the balancing interface and the target interface can serve the same business request; allocating the request corresponding to the target interface to the balancing interface; calling the balancing interface or the target interface for the business request.

[0055] You can use a load balancer or related tools to monitor the load of each interface in real time, including CPU usage, memory usage, network bandwidth and other load-related data.

[0056] You can set high concurrent request thresholds based on system performance and business requirements, such as the number of requests per second, response time, etc. The interfaces that exceed the threshold in the prediction results are identified as target interfaces for high concurrent requests, which may face greater pressure in the future.

[0057] Interfaces that can process the same service requests as high-concurrency interfaces are selected from interfaces with low-concurrency requests as candidate balancing interfaces. The candidate balancing interfaces have sufficient processing capabilities and remaining resources.

[0058] Evaluate the capabilities of candidate balancing interfaces, including response time, error rate, resource utilization, etc., to ensure that they can handle the additional requests. Based on the evaluation results, select the most suitable candidate interface as the balancing interface. The evaluation results will include evaluation indicators in multiple dimensions, such as the response time, error rate, and resource utilization balance of the interface.

[0059] Design a request allocation strategy that can dynamically adjust the allocation ratio of requests to the target interface and the balancing interface according to the real-time load situation. Implement the request allocation strategy in the system to ensure that business requests can be correctly allocated to the target interface or the balancing interface according to the strategy.

[0060] Monitor the load of the target interface and the balancing interface in real time, including the request volume, response time, etc. The target scheduling strategy can be adjusted dynamically according to the real-time load situation, thereby dynamically adjusting the request allocation strategy to ensure that the load of the target interface and the balancing interface remains balanced. When a service request is received, it is allocated to the target interface or the balancing interface for processing according to the current request allocation strategy.

[0061] By adopting the embodiments of the present disclosure, by allocating part of the requests of the target interface with high concurrency requests to the balancing interface with low concurrency requests, the balanced distribution of the system load is achieved, and the performance bottleneck that may occur due to excessive load on a single interface is avoided, thereby optimizing the overall performance of the system. The interface with low concurrency requests is fully utilized under the balanced load strategy, and the originally idle resources are effectively utilized, which improves the overall utilization rate of resources and reduces resource waste. By balancing the load, the risk of overload and crash of the target interface caused by high concurrency requests is avoided, the stability and reliability of the system are enhanced, and the continuity of the service is ensured.

[0062] Among them, in an optional embodiment, when the interface call management is used to adjust the priority of resource usage by each interface, the corresponding interface is called for the business request through the target scheduling strategy, including: determining the predicted call status of each interface and the importance of each interface; the call status includes high concurrency requests and low concurrency requests; when the call status of the interface is a high concurrency request, and / or when the priority of the interface is greater than or equal to a first threshold, computing resources are preferentially allocated to the interface based on the usage of computing resources by the interface; and calling the interface after the computing resources are allocated for the business request.

[0063] According to system performance and business requirements, a first threshold can be set, and the first threshold is used to determine which interfaces need to be allocated computing resources first. The priority scores of the interfaces are compared with the first threshold to determine which interfaces have higher priorities.

[0064] Check the predicted interface calls and identify the interfaces with high concurrent requests.

[0065] Develop a computing resource allocation plan based on interface calls, importance, and resource requirements.

[0066] When the interface call situation is a high concurrent request, or the interface priority is greater than or equal to the first threshold, computing resources are allocated to these interfaces in priority according to the allocation plan. During the resource allocation process, the allocation strategy is dynamically adjusted according to the real-time load and resource usage of the interface to ensure the effective use of resources.

[0067] After allocating computing resources, you can also select the appropriate resource-allocated interface for processing based on the type and priority of the business request. Call the selected interface to process the business request and monitor its execution and response time.

[0068] Continuously monitor the resource usage and performance of each interface. According to the monitoring results, the target scheduling strategy can be dynamically adjusted, so as to timely adjust the allocation strategy of computing resources to ensure that the system is always in the best state.

[0069] By adopting the embodiments of the present disclosure, by giving priority to allocating computing resources to high-concurrency requests or important interfaces, business requests can be responded to more quickly, thereby improving the response speed of the overall system. The system can dynamically adjust resource allocation according to the priority and call status of the interface, ensuring that key business interfaces receive sufficient resource support when there is high concurrency, while also avoiding the situation where resources are over-allocated to low-priority or low-concurrency interfaces, thereby optimizing resource utilization.

[0070] Among them, in an optional embodiment, the interface call management is used to adjust the calling order of each interface, and through the target scheduling strategy, the corresponding interface is called for the business request, including: determining the priority of each business request corresponding to the interface; when the priority of the business request indicates priority processing, the interface is called for the business request in priority; when the priority of the business request does not indicate priority processing, the interface is called for the business request according to the time when the interface receives the business request.

[0071] The priority of each business request can be evaluated based on business logic, user needs or system policies. The evaluation result can be a comprehensive consideration of factors such as the importance, urgency, and resource requirements of the request. The priority information obtained from the evaluation is recorded in the record of each business request.

[0072] From all pending business requests, select requests with a priority higher than or equal to the preset threshold as high-priority requests. According to the characteristics and requirements of high-priority requests, assign appropriate interfaces to them for calling. According to the assigned interface and request parameters, execute the interface call operation and process the return results.

[0073] Business requests that are not filtered as high-priority requests are considered non-priority requests. Sort by receiving time: Sort non-priority requests in the order of receiving time. Based on the sorting results, interfaces are allocated and called for non-priority requests in turn.

[0074] Monitor interface calls: Monitor the calls of each interface in real time, including call times, response time, success rate and other indicators. According to the monitoring results, dynamically adjust the target scheduling strategy, thereby dynamically adjusting the priority evaluation strategy of business requests, interface allocation strategy, etc., to optimize system performance and resource utilization.

[0075] By adopting the embodiments of the present disclosure, high-priority requests can be processed first, meeting the user's rapid response to key business needs. Interface calls are made according to the priority and reception time of the request, avoiding excessive occupation and waste of resources. Through real-time monitoring and dynamic adjustment of strategies, the system can better cope with emergencies and resource pressures, ensuring the stable operation of the system.

[0076] Figure 2 Schematic diagram of the steps of an interface adaptive scheduling method shown in an embodiment of the present disclosure. Figure 2 As shown, firstly, data collection and data processing are performed; the target neural network is trained based on the collected data; the future trend of interface calls is predicted using the trained target neural network, and the prediction results are evaluated; according to the evaluation results, the target scheduling strategy is determined and executed; under the target scheduling strategy, each interface is monitored to obtain the real-time calling status, and the target neural network is adjusted according to the real-time calling status.

[0077] Based on the same technical concept, the present disclosure proposes an interface adaptive scheduling device. Figure 3 is a block diagram of an interface adaptive scheduling device shown in an embodiment of the present disclosure. Figure 3 As shown, the device comprises: A receiving module 310, configured to receive a service request; A prediction module 320 is used to predict the future trend of interface calls using a target neural network to obtain a prediction result, wherein the target neural network is dynamically adjusted based on the real-time call status of the interface; The determination module 330 is used to determine a target scheduling strategy by evaluating the prediction result; the target scheduling strategy is used to adjust the calling conditions of each interface; The calling module 340 is used to call the corresponding interface for the service request through the target scheduling strategy.

[0078] Optionally, the device further comprises a feedback module, wherein the feedback module is configured to: After executing the target scheduling strategy, each interface is monitored to obtain the real-time calling status of each interface; the real-time calling status of each interface includes the calling status of each interface and the performance data of each interface; the real-time calling status is used to evaluate the effectiveness of the target scheduling strategy; According to the real-time calling conditions of the various interfaces, the target neural network is adjusted to obtain an adjusted target neural network; wherein, adjusting the target neural network includes: fine-tuning the parameters of the target neural network, or retraining the target neural network.

[0079] Optionally, the calling module is specifically used to execute: Through the target scheduling strategy, interface call management is performed; the interface call management includes: adjusting the load balance of each interface, adjusting the priority of each interface to use resources, and adjusting the calling order of each interface; By performing interface call management, an interface is called for the business request.

[0080] Optionally, when the interface call management is used to adjust the load balancing of each interface, the calling module is specifically used to execute: Determine the predicted calling conditions of each interface; the calling conditions include high-concurrency requests and low-concurrency requests; For a target interface with high concurrent requests, determining a balancing interface corresponding to the target interface from interfaces with low concurrent requests; the balancing interface and the target interface can serve the same service request; Allocating requests corresponding to the target interface to the balancing interface; The balancing interface or the target interface is called for the service request.

[0081] Optionally, when the interface call management is used to adjust the priority of each interface using resources, the calling module is specifically used to execute: Determine the predicted call status of each interface and the importance of each interface; the call status includes high-concurrency requests and low-concurrency requests; When the calling condition of the interface is a high concurrent request, and / or when the priority of the interface is greater than or equal to a first threshold, according to the usage of the computing resources by the interface, computing resources are preferentially allocated to the interface; The interface after the computing resources are allocated is called for the business request.

[0082] Optionally, the interface call management is used to adjust the call order of each interface, and the call module is specifically used to execute: Determine the priority of each service request corresponding to the interface; In the case where the priority of the service request indicates a priority process, calling the interface for the service request first; In the case where the priority of the service request does not indicate a priority process, the interface is called for the service request according to the time when the interface receives the service request.

[0083] Optionally, the feedback module is specifically configured to execute: Collect historical data of interface calls; the historical data of interface calls includes interface name, call time, request parameters and response results; Preprocessing the historical data called by the interface to obtain a data set for training a neural network; Determine the initial neural network model; Customizing the initial neural network model according to the characteristics and requirements of the interface call to obtain a neural network model to be trained; Based on the purpose of learning the patterns and regularities of interface calls, the neural network model to be trained is trained through the data set to obtain a target neural network.

[0084] The present disclosure also provides an electronic device, referring to Figure 4 , Figure 4 is a schematic diagram of an electronic device proposed in an embodiment of the present disclosure. Figure 4 As shown, the electronic device 400 includes: a memory 410 and a processor 420. The memory 410 and the processor 420 are connected via a bus communication. A computer program is stored in the memory 410. The computer program can be run on the processor 420 to implement the steps in the interface adaptive scheduling method disclosed in the embodiment of the present disclosure.

[0085] The embodiment of the present disclosure further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the interface adaptive scheduling method disclosed in the embodiment of the present disclosure are implemented.

[0086] The embodiment of the present disclosure further provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps in the interface adaptive scheduling method disclosed in the embodiment of the present disclosure are implemented.

[0087] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0088] It should be understood by those skilled in the art that the embodiments of the present disclosure may be provided as methods, devices or computer program products. Therefore, the embodiments of the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0089] The embodiments of the present disclosure are described with reference to the flowcharts and / or block diagrams of the methods, apparatuses, electronic devices, and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0090] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0092] Although some embodiments of the present disclosure have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiment and all changes and modifications that fall within the scope of the present disclosure.

[0093] The above is a detailed introduction to an interface adaptive scheduling method, device, equipment and medium provided by the present disclosure. Specific examples are used in this article to illustrate the principles and implementation methods of the present disclosure. The description of the above embodiments is only used to help understand the method of the present disclosure and its core idea. At the same time, for general technical personnel in this field, according to the ideas of the present disclosure, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present disclosure.

Claims

1. An interface adaptive scheduling method, characterized in that: include: Receive business requests; Using a target neural network, predicting the future trend of interface calls to obtain a prediction result, wherein the target neural network is dynamically adjusted based on the real-time call status of the interface; Determining a target scheduling strategy by evaluating the prediction results; The target scheduling strategy is used to adjust the calling conditions of each interface; The target scheduling strategy is used to call a corresponding interface for the service request.

2. The method according to claim 1, characterized in that The target neural network is dynamically adjusted as follows: After executing the target scheduling strategy, each interface is monitored to obtain the real-time calling status of each interface; the real-time calling status of each interface includes the calling status of each interface and the performance data of each interface; the real-time calling status is used to evaluate the effectiveness of the target scheduling strategy; According to the real-time calling conditions of the various interfaces, the target neural network is adjusted to obtain an adjusted target neural network; wherein, adjusting the target neural network includes: fine-tuning the parameters of the target neural network, or retraining the target neural network.

3. The method according to claim 1, characterized in that By using the target scheduling strategy, calling a corresponding interface for the service request includes: Through the target scheduling strategy, interface call management is performed; the interface call management includes: adjusting the load balance of each interface, adjusting the priority of each interface to use resources, and adjusting the calling order of each interface; By performing interface call management, an interface is called for the business request.

4. The method according to claim 3, characterized in that In the case where the interface call management is used to adjust the load balancing of each interface, calling the corresponding interface for the service request through the target scheduling strategy includes: Determine the predicted calling conditions of each interface; the calling conditions include high-concurrency requests and low-concurrency requests; For a target interface with high concurrent requests, determining a balancing interface corresponding to the target interface from interfaces with low concurrent requests; the balancing interface and the target interface can serve the same service request; Allocating requests corresponding to the target interface to the balancing interface; The balancing interface or the target interface is called for the service request.

5. The method according to claim 3, characterized in that: In the case where the interface call management is used to adjust the priority of each interface in using resources, calling the corresponding interface for the service request through the target scheduling strategy includes: Determine the predicted call status of each interface and the importance of each interface; the call status includes high-concurrency requests and low-concurrency requests; When the calling condition of the interface is a high concurrent request, and / or when the priority of the interface is greater than or equal to a first threshold, according to the usage of the computing resources by the interface, computing resources are preferentially allocated to the interface; The interface after the computing resources are allocated is called for the business request.

6. The method according to claim 3, characterized in that The interface call management is used to adjust the call order of each interface, and through the target scheduling strategy, the corresponding interface is called for the service request, including: Determine the priority of each service request corresponding to the interface; In the case where the priority of the service request indicates a priority process, calling the interface for the service request first; In the case where the priority of the service request does not indicate a priority process, the interface is called for the service request according to the time when the interface receives the service request.

7. The method according to claim 2, characterized in that: The target neural network is trained as follows, including: Collect historical data of interface calls; the historical data of interface calls includes interface name, call time, request parameters and response results; Preprocessing the historical data called by the interface to obtain a data set for training a neural network; Determine the initial neural network model; Customizing the initial neural network model according to the characteristics and requirements of the interface call to obtain a neural network model to be trained; Based on the purpose of learning the patterns and regularities of interface calls, the neural network model to be trained is trained through the data set to obtain a target neural network.

8. An interface adaptive scheduling device, characterized in that: include: A receiving module, used for receiving a service request; A prediction module is used to use a target neural network to predict the future trend of interface calls and obtain a prediction result. The target neural network is dynamically adjusted based on the real-time call status of the interface. A determination module, used to determine a target scheduling strategy by evaluating the prediction results; The target scheduling strategy is used to adjust the calling conditions of each interface; The calling module is used to call the corresponding interface for the business request through the target scheduling strategy.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of the interface adaptive scheduling method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the interface adaptive scheduling method according to any one of claims 1 to 7 are implemented.

Images