Scheduling System and Method for Heterogeneous Cryptographic Resource Pool

Abstract

The present invention relates to the technical field of information system security, and specifically discloses a scheduling system and method for a heterogeneous password resource pool. The system includes a service gateway, a general server, a learning database, a scheduling module, and a number of password devices. The general server is deployed with a virtual module and a load balancing module. The virtual module is used to virtualize the password devices into different types of resource clouds. The service gateway is also used to send service interface call data to the learning database for storage. The learning database is also used to obtain the device status data of the password devices. The scheduling module is used to train an AI model through the service interface call data and the device status data, and predict the resource load situation of the resource cloud in a future period of time. The load balancing module is used to compare and correct the preliminary result with the prediction result of the AI model when making a load balancing judgment, and determine the password device to be sent. By adopting the technical solution of the present invention, dynamic scheduling can be achieved, and the operation efficiency of the system can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of information system security, and in particular to a scheduling system and method for a heterogeneous cryptographic resource pool. Background Art

[0002] In the face of a complex network security environment, cryptographic technology has been widely used in various business links such as network security identity authentication, digital signatures, and data information encryption. As an important component of security application systems, cryptographic service systems have become increasingly important and complex. At present, cryptographic service systems are mainly networked and clustered with devices of the same manufacturer, type, and model. However, due to the non-uniform construction cycle of various cryptographic service systems, devices of different manufacturers, types, and models will be used, resulting in the need to develop and deploy multiple sets of cryptographic service systems in a targeted manner, which has poor scalability and cannot meet the continuous expansion of business development scale. At the same time, as the business volume increases, it is easy to see a sharp increase in access volume and data traffic, making the processing capacity of a single cryptographic device unable to withstand the current business, resulting in high energy consumption, easy damage, and difficult operation and maintenance of cryptographic devices.

[0003] Therefore, a scheduling system and method for a heterogeneous cryptographic resource pool is needed that can realize dynamic scheduling and improve system operation efficiency. Summary of the invention

[0004] One of the purposes of the present invention is to provide a scheduling system for a heterogeneous cryptographic resource pool, which can realize dynamic scheduling and improve system operation efficiency.

[0005] In order to solve the above technical problems, this application provides the following technical solutions:

[0006] A scheduling system for a heterogeneous cryptographic resource pool, including a service gateway, a synchronization module, a general server, a learning database, a scheduling module, and several cryptographic devices;

[0007] The synchronization module is used to synchronize key resources of cryptographic devices of the same type after the cryptographic devices are connected to the network;

[0008] The general server is deployed with a virtual module and a load balancing module; the load balancing module is used to connect the cryptographic device after key resource synchronization; the virtual module is used to virtualize the physical resources of the cryptographic device into different types of resource clouds through virtualization technology;

[0009] The service gateway is used to obtain the call application of the third-party application service interface; it is also used to parse the call application and determine the service request for different resource clouds based on the parsed results;

[0010] The service gateway is also used to send the service interface call data to the learning database for storage; the learning database is also used to obtain the device status data of the cryptographic device;

[0011] The scheduling module pre-stores an AI model. The scheduling module is used to train the AI model by invoking data and device status data through a service interface; the scheduling module is also used to predict the resource cloud resource load situation within a certain period of time in the future through the trained AI model;

[0012] When the load balancing module is also used for load balancing judgment, it obtains a preliminary result according to the load balancing algorithm preset for different types of cryptographic devices, and then compares and corrects the preliminary result with the prediction result obtained by the AI model to determine the cryptographic device to be sent.

[0013] The principle and beneficial effects of the basic solution are as follows:

[0014] In this solution, data and device status data are pre-invoked through a service interface, that is, long-term data such as service interface invocation situations and device working conditions are used to train the AI model, so that the trained AI model can predict the resource cloud resource load situation within a certain period of time in the future, such as predicting the system resource usage trend and what specific operations users will perform when reaching the critical points of different trends, etc., as the basis for load balancing judgment, and finally generate a decision flow. This decision flow can adjust resource allocation more accurately and finely, realize dynamic scheduling, and improve the system operation efficiency.

[0015] Furthermore, the scheduling module is also used to correct the AI model according to the difference between the actual result and the prediction result after the training data reaches the set quantity.

[0016] The prediction accuracy rate of the AI model can be improved.

[0017] Furthermore, the learning database is also used to classify and store device status data. The classified device status data includes device load data and device fault data.

[0018] Furthermore, the AI model includes a load balancing configuration model; the scheduling module is also used to train the load balancing configuration model by invoking data and device load data through a service interface, so that the load balancing configuration model obtains the resource trend situation of the cryptographic device, and predicts the resource cloud resource load situation and device resource load situation within a certain period of time in the future through the trained load balancing configuration model.

[0019] Furthermore, the AI model also includes an interface call prediction model;

[0020] The scheduling module is used to train the interface call prediction model through the service interface call data in the learning database, so that the interface call prediction model obtains the call trend situation of the service interface according to the time distribution, and predicts the interface call trend within a certain period of time in the future through the trained interface call prediction model;

[0021] The scheduling module is also used to analyze according to the prediction results of the interface call prediction model, so as to obtain the interface service call frequency, interface concurrency, data traffic in each time period of the whole day, and to give early warnings for peak periods.

[0022] Furthermore, the AI model further includes a device status prediction model;

[0023] The scheduling module is also used to train the device status prediction model through device load data, so that the device status prediction model obtains the resource trend of the cryptographic device, and predicts the device resource load in the future for a period of time through the trained device status prediction model;

[0024] The scheduling module is also used to analyze according to the prediction results of the device status prediction model, so as to obtain the operation conditions of the cryptographic device throughout the day, including device resource load, peak periods, idle periods, CPU utilization in each period, memory occupancy in each period, and TPS.

[0025] Realize early warning for peak periods.

[0026] Furthermore, the AI model further includes a fault and problem handling model;

[0027] The scheduling module is also used to train the fault and problem handling model through device fault data, so that the fault and problem handling model obtains the fault trend of the cryptographic device, and predicts the device fault probability in the future for a period of time through the trained fault and problem handling model;

[0028] The scheduling module is also used to judge whether the device fault probability is higher than the threshold according to the prediction results of the fault and problem handling model. If it is higher than the threshold, an abnormal alarm is given.

[0029] It can give an abnormal alarm before a system error occurs.

[0030] The second object of the present invention is to provide a scheduling method for a heterogeneous cryptographic resource pool, including the following steps:

[0031] S1. The service gateway receives a call application from a third-party application through a service interface;

[0032] S2. Parse the call application, determine the resource cloud to be called according to the parsed result, and forward the service request of the third-party application to the corresponding resource cloud;

[0033] S3. The load balancing module obtains a preliminary result according to the preset load balancing algorithm for different types of cryptographic devices, and then compares and corrects the preliminary result with the prediction result obtained by the load balancing configuration model to determine the distributed cryptographic device;

[0034] S4. Distribute service requests according to the determined cryptographic device;

[0035] S5. After the cryptographic device receives a service request, process the service request;

[0036] S6. The device status prediction model outputs the prediction result of the processing completion time and sends the prediction result to the load balancing module;

[0037] S7. After the load balancing module receives the prediction result, establish a connection with the service gateway in advance;

[0038] S8. After the cryptographic device finishes processing the service request, send the processing result to the resource cloud;

[0039] S9. The load balancing module reads the processing result from the resource cloud and sends it to the service gateway;

[0040] S10. The service gateway sends the processing result to the third-party application through the service interface.

[0041] Further, in step S2, after the service request is forwarded to the resource cloud, the connection of the service request is disconnected.

[0042] Resources can be released for other service requests. Description of the Drawings

[0043] Figure 1 It is the logical block diagram of the scheduling system of the heterogeneous cryptographic resource pool in the first embodiment;

[0044] Figure 2 It is the flowchart of the scheduling method of the heterogeneous cryptographic resource pool in the first embodiment. Detailed Embodiments

[0045] The following is further detailed through specific embodiments:

[0046] The First Embodiment

[0047] As Figure 1 shown, the scheduling system of the heterogeneous cryptographic resource pool in this embodiment includes a service gateway, a synchronization module, a general server, a learning database, a scheduling module, and several cryptographic devices.

[0048] The synchronization module is used to synchronize key resources of the same type of cryptographic devices after the cryptographic devices are connected to the network. For example, for encryption machines, cryptographic cards, etc., key synchronization is involved; for timestamps, signature verification, etc., certificate synchronization is involved, and for other cryptographic devices, the key resources to be synchronized are determined according to the nature of the devices.

[0049] The general server is deployed with a virtual module and a load balancing module. The load balancing module is used to connect to the password device after key resource synchronization. The virtual module is used to pool the physical resources of the password device through virtualization technology to form a resource pool, and then virtualize the resource pool into different types of resource clouds to shield the underlying implementation. The resource clouds include one or more of an encryption service cloud, a signature verification cloud, and a dynamic password cloud. In this embodiment, it includes an encryption service cloud, a signature verification cloud, and a dynamic password cloud, etc. In this embodiment, the load balancing module runs on the general server. In other embodiments, a load balancing cluster can also be constructed according to actual situations for hot standby and performance expansion. The load balancing module adopts pipeline technology and cooperates with a caching mechanism to improve the concurrency of the system.

[0050] The service gateway serves as a service entry, used to obtain the call application of the third-party application service interface; it is also used to parse the call application and determine the internal interface to be called according to the parsed result, that is, the service request for different resource clouds. Specifically, the service gateway determines whether the call application is an authorization request according to the call application. If it is a non-authorization request, it rejects the call application to implement call permission control, that is, it allows authorized requests to call and does not allow unauthorized calls.

[0051] The service gateway is also used to send the service interface call data to the learning database for storage to achieve bypass drainage, avoiding increasing the burden on the system when obtaining call data from the main data path.

[0052] The learning database is also used to obtain the device status data of the password device.

[0053] The scheduling module pre-stores an AI model. The scheduling module is used to train the AI model through the service interface call data and the device status data, so that the AI model obtains the resource trend of the password device, so as to be able to predict the resource load of the resource cloud and the device resource load within a certain period of time in the future to optimize the load balancing configuration. The scheduling module is also used to correct the AI model according to the difference between the actual result and the predicted result after the training data reaches the set quantity to improve the prediction accuracy of the AI model. In this embodiment, the set quantity is at the 100,000 level.

[0054] The load balancing module is also used to perform load balancing judgment after receiving the service request: obtain a preliminary result according to the load balancing algorithm preset for different types of password devices, and then compare and correct the preliminary result with the predicted result obtained by the AI model to obtain the optimal result. The optimal result is the smallest resource overhead, meeting the maximum demand, and obtaining the most satisfactory result. In other words, it is the greatest benefit.

[0055] In this embodiment, the load balancing algorithm obtains the monitoring results of the cryptographic device resources and specifically uses the least pressure algorithm to obtain preliminary results. Taking a server cryptographic machine as an example, at 12:00 on the Xth day of X month, the system received a service request. At this time, the load balancing module obtained through the least pressure algorithm that the current pressure of Device A is the smallest, and the preliminary result is to forward the request to Device A; however, according to the prediction of the AI model, since Device A has a serious shortage of resources, it will automatically shut down and go offline through the dynamic resource pool mechanism in the future. Therefore, after comparison, this request will be transferred to other devices; for another example, although the pressure of Device A is small at this time, the required requests exceed the capacity of Device A. Although the pressure of Device B is not the smallest at this time, but it is predicted that the resources of Device B can be released at the next moment, then the system will still forward the request to Device B at this time.

[0056] The cryptographic device is used to respond to the service request after receiving it and send the response result to the cache of the resource cloud. The load balancing module is also used to send the response result in the cache to the service gateway; the service gateway sends the response result to the corresponding third-party application through the service interface. In this embodiment, the purpose of setting the cache is to improve the request response speed of the system and increase the TPS (the number of transactions processed by the server per second) value.

[0057] Specifically, the learning database is also used to classify and store the device status data. In this embodiment, the classified device status data includes device load data and device fault data.

[0058] The AI model includes several types. In this embodiment, it includes an interface call prediction model, a device status prediction model, a fault and problem handling model, a load balancing configuration model, etc. In this embodiment, the above models are constructed by combining the clustering analysis model and the time series model.

[0059] The scheduling module is used to train the interface call prediction model through the service interface call data in the learning database, so that the interface call prediction model obtains the call trend of the service interface according to the time distribution, so as to be able to predict the interface call trend in the future for a period of time.

[0060] The scheduling module is also used to analyze according to the prediction results of the interface call prediction model to obtain the interface service call frequency, interface concurrency, data traffic in each time period of the whole day, and to give early warnings for peak periods. It is also used to alarm the IP or third-party application whose abnormal call status times exceed the set value.

[0061] The scheduling module is also used to read device status data from the learning database. Specifically, the device status prediction model is trained with device load data, enabling the device status prediction model to obtain the resource trend of the cryptographic device, so as to predict the device resource load situation in a future period of time.

[0062] The scheduling module is also used to analyze according to the prediction results of the device status prediction model to obtain the full-day operation situation of the cryptographic device, including device resource load situation, peak hours, idle hours, CPU utilization rate in each period, memory occupancy in each period, and TPS. Thus, early warnings can be issued for peak hours.

[0063] The scheduling module is also used to train the fault and problem handling model with device fault data, enabling the fault and problem handling model to obtain the fault trend of the cryptographic device, so as to predict the device fault probability in a future period of time.

[0064] The scheduling module is also used to judge whether the device fault probability is higher than the threshold based on the prediction results of the fault and problem handling model, that is, the device fault probability in a future period of time. If it is higher than the threshold, an abnormal alarm is issued before a system error occurs.

[0065] The scheduling module is also used to train the load balancing configuration model with data called through the service interface and device load data, enabling the load balancing configuration model to obtain the resource trend of the cryptographic device, so as to predict the resource cloud resource load situation and device resource load situation in a future period of time. The future period of time can be set according to the actual situation, such as 3 hours, 6 hours, 12 hours, 24 hours, 48 hours, 72 hours, 168 hours, etc.

[0066] The load balancing module is also used to dynamically configure and adjust the load balancing scheme of the resource cloud according to the prediction results of the load balancing configuration model to achieve the optimization of resource utilization.

[0067] As Figure 2 shown, for the scheduling system based on the heterogeneous cryptographic resource pool, this embodiment also provides a scheduling method for the heterogeneous cryptographic resource pool, including the following steps:

[0068] S1. The service gateway receives a call application from a third-party application through a unified service interface;

[0069] S2. Parse the call application, determine the internal interface to be called according to the parsed result, and forward the service request of the third-party application to the resource cloud corresponding to the internal interface; after the service request is forwarded to the resource cloud, the service request immediately disconnects and releases resources (the disconnection in this embodiment is based on the pipeline technology, and the connections in the pre-established connection pool will not be released, the same below), for other service requests.

[0070] S3. The load balancing module obtains a preliminary result according to the load balancing algorithms preset for different types of cryptographic devices, and then compares and corrects the preliminary result with the prediction result obtained from the load balancing configuration model to determine the distributed cryptographic device;

[0071] S4. Distribute the service requests according to the determined cryptographic device; after the service requests are distributed, disconnect the connection and release the resources;

[0072] S5. After receiving the service request, the cryptographic device processes the service request;

[0073] S6. The device status prediction model outputs the prediction result of the processing completion time and sends the prediction result to the load balancing module;

[0074] S7. After receiving the prediction result, the load balancing module establishes a connection with the service gateway in advance;

[0075] S8. After the cryptographic device finishes processing the service request, it sends the processing result to the cache of the resource cloud;

[0076] S9. The load balancing module reads the processing result from the cache and sends it to the service gateway;

[0077] S10. The service gateway sends the processing result to the third-party application through the service interface.

[0078] The solution of this embodiment integrates different manufacturers and different types of cryptographic devices internally. Based on virtualization technology, the cryptographic device resources are combined into a heterogeneous resource cloud; externally, it provides a unified interface service and provides a resource cloud by type. The cryptographic devices can be networked flexibly, which is convenient for reusing existing resources, does not waste existing resources, saves costs, shields the details of the composition of the cryptographic device resources, can improve the operation efficiency, and reduce the operation and maintenance complexity.

[0079] Through the AI model, load balancing is realized to achieve the purpose of intelligent dynamic scheduling, improve the system reliability, extend the device life, reduce the operation and maintenance time, and reduce the device energy consumption.

[0080] Embodiment 2

[0081] The difference between this embodiment and Embodiment 1 is that in this embodiment, the scheduling module is also used to generate a disposal plan according to the prediction result of the AI model.

[0082] For example, before a system error occurs, an abnormal alarm is given, and then the abnormal point information is displayed, and the disposal plan for the current abnormality is determined from several pre-stored disposal plans. According to the actual situation, the disposal plan can be executed manually or automatically by the system. Through the solution in this embodiment, semi-automatic and fully automatic operation and maintenance can be provided in the operation and maintenance link, reducing manual intervention.

[0083] Embodiment 3

[0084] The difference between this embodiment and the first embodiment is that, in this embodiment, in the initial stage of platform construction and operation, after the key resources of the same type of cryptographic devices are synchronized, a resource pool is formed through virtual modules.

[0085] During the operation of the system, the resource consumption of different cryptographic devices and the overall system resource usage are obtained through device load data; in this embodiment, the overall system resource usage refers to the summary of the resource consumption of the same type of cryptographic devices.

[0086] The scheduling module performs deep learning on the load balancing configuration model through such data, and predicts the device resource load and the overall system resource usage of different cryptographic devices in a future period through the load balancing configuration model.

[0087] If the difference obtained by subtracting the product of the current overall system resource and the system security redundancy factor from the predicted overall system resource is greater than the resource capacity of a certain cryptographic device, the virtual module is also used to set this cryptographic device as a spare device, and the load balancing module is used to transfer the service requests on this cryptographic device to other cryptographic devices, so as to shut down this device, saving energy and device wear; the system security redundancy factor is used to ensure that the resources will not reach 100% utilization rate, and the initial value is 1.2 - 1.3.

[0088] If the difference obtained by subtracting the predicted overall system resource used in a future period from the quotient of the current overall system resource divided by the system security redundancy factor is less than 0, the virtual module is also used to issue a resource shortage warning and automatically start the cryptographic devices in the off state to increase system resources.

[0089] In summary, through the prediction results of the load balancing configuration model, the automatic start / stop of cryptographic devices is realized, achieving the purpose of dynamically constructing a resource pool.

[0090] Embodiment Four

[0091] The difference between this embodiment and the third embodiment is that, in this embodiment, the scheduling module is also used to determine whether it is necessary to expand cryptographic devices according to the predicted overall system resource usage in a future period; for example, if the overall system resource usage has been in a full-load state, it is determined that cryptographic devices need to be expanded. Scoring is only performed after it is predicted that expansion is needed, which can make the scoring most reflect the current device status.

[0092] If it is necessary to expand the cryptographic device, determine the order of the scored cryptographic devices according to the predicted device resource load conditions of different cryptographic devices within the next 3 hours, 6 hours, 12 hours, 24 hours, 48 hours, 72 hours, and 168 hours. In this embodiment, the sorting is performed in descending order of the device resource load conditions. It is possible to score a single cryptographic device or multiple cryptographic devices simultaneously. The scheduling module is also used to mark the cryptographic devices being scored. The load balancing module is also used to, after receiving a service request, preferentially distribute the service request to the cryptographic devices being scored. When the cryptographic devices being scored cannot handle the service request, then determine the cryptographic device to be sent through load balancing judgment.

[0093] The scheduling module is also used to score the cryptographic devices according to the device failure data of each cryptographic device, and generate a device recommendation list according to the descending order of the scores of the same type of cryptographic devices. When expanding, it can be used as a reference for relevant personnel to select devices of appropriate manufacturers and models. For example, the failure score X = -(a * λ_1) - (b * λ_2) - (c * λ_3) + (d * λ_4), where a is the total number of failures per unit working hour, b is the total repair time of failures during the unit working hour, c is the number of failures during peak hours per unit working hour, d is the accuracy rate of the device failure probability prediction by the failure and problem handling model; λ_1, λ_2, λ_3, and λ_4 are all weight coefficients and can be set separately according to the actual situation; in this embodiment, the unit working hour is 500 hours.

[0094] The above are only embodiments of the present invention. The invention is not limited to the fields involved in this embodiment case. Common knowledge such as the specific structures and characteristics known in the art is not described in detail herein. Those of ordinary skill in the art know all the common technical knowledge in the technical field to which the invention belongs before the filing date or the priority date, can know all the existing technologies in this field, and have the ability to apply the conventional experimental means before this date. Those of ordinary skill in the art can, under the inspiration given in this application, combine their own abilities to improve and implement this solution. Some typical well-known structures or well-known methods should not become an obstacle for those of ordinary skill in the art to implement this application. It should be noted that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can still be made, and these should also be regarded as the protection scope of the present invention, and these will not affect the implementation effect of the present invention and the practicality of the patent. The protection scope required by this application should be based on the content of its claims, and the specific implementation manners described in the specification can be used to explain the content of the claims.

Claims

1. A scheduling system for a heterogeneous cryptographic resource pool, characterized in that, It includes a service gateway, a synchronization module, a general server, a learning database, a scheduling module, and several cryptographic devices; The synchronization module is used to synchronize the critical resources of the same type of cryptographic devices after the cryptographic devices are connected to the network; The general server is deployed with a virtual module and a load balancing module; The load balancing module is used to connect the cryptographic devices after the critical resource synchronization; the virtual module is used to virtualize the physical resources of the cryptographic devices into different types of resource clouds through virtualization technology; The service gateway is used to obtain the call application of the third-party application service interface; it is also used to parse the call application and determine the service request for different resource clouds according to the parsed result; The service gateway is also used to send the service interface call data to the learning database for storage; the learning database is also used to obtain the device status data of the cryptographic devices; The scheduling module pre-stores an AI model. The scheduling module is used to train the AI model through the service interface call data and the device status data; the scheduling module is also used to predict the resource load situation of the resource cloud in a future period of time through the trained AI model; When the load balancing module is also used for load balancing judgment, it obtains a preliminary result according to the preset load balancing algorithm of different types of cryptographic devices, and then compares and corrects the preliminary result with the prediction result obtained by the AI model to determine the cryptographic device to be sent; The learning database is also used to classify and store the device status data. The classified device status data includes device load data and device fault data; The AI model includes a load balancing configuration model; the scheduling module is also used to train the load balancing configuration model through the service interface call data and the device load data, so that the load balancing configuration model obtains the resource trend of the cryptographic devices, and predicts the resource load situation of the resource cloud and the device resource load situation in a future period of time through the trained load balancing configuration model; The AI model also includes an interface call prediction model; The scheduling module is used to train the interface call prediction model through the service interface call data in the learning database, so that the interface call prediction model obtains the call trend of the service interface according to the time distribution, and predicts the interface call trend in a future period of time through the trained interface call prediction model; The scheduling module is also used to analyze according to the prediction result of the interface call prediction model to obtain the interface service call frequency, interface concurrency, data traffic, and peak period warning throughout the day; The AI model also includes a device status prediction model; The scheduling module is also used to train the device status prediction model from the device load data, so that the device status prediction model obtains the resource trend of the cryptographic devices, and predicts the device resource load situation in a future period of time through the trained device status prediction model; The scheduling module is also used to analyze according to the prediction result of the device status prediction model to obtain the daily operation of the cryptographic devices, including the device resource load, peak period, idle period, CPU utilization rate of each period, memory occupancy of each period, and TPS; The AI model also includes a fault and problem handling model; The scheduling module is also used to train the fault and problem handling model with device fault data, so that the fault and problem handling model obtains the fault trend of the cryptographic device, and predicts the device fault probability within a certain period in the future through the trained fault and problem handling model; The scheduling module is also used to judge whether the device fault probability is higher than the threshold according to the prediction result of the fault and problem handling model. If it is higher than the threshold, an abnormal alarm is issued.

2. The scheduling system for the heterogeneous cryptographic resource pool according to claim 1, wherein: The scheduling module is also used to correct the AI model according to the difference between the actual result and the prediction result after the training data reaches the set quantity.

3. A scheduling method for a heterogeneous cryptographic resource pool, using the system according to any one of claims 1-2, characterized in that, It includes the following steps: S1. The service gateway receives the call application of the third-party application through the service interface; S2. Parse the call application, determine the resource cloud to be called according to the parsed result, and forward the service request of the third-party application to the corresponding resource cloud; S3. The load balancing module obtains the preliminary result according to the preset load balancing algorithm for different types of cryptographic devices, and then compares and corrects the preliminary result with the prediction result obtained by the load balancing configuration model to determine the distributed cryptographic device; S4. Distribute the service request according to the determined cryptographic device; S5. After receiving the service request, the cryptographic device processes the service request; S6. The device status prediction model outputs the prediction result of the processing completion time and sends the prediction result to the load balancing module; S7. After receiving the prediction result, the load balancing module establishes a connection with the service gateway in advance; S8. After the cryptographic device finishes processing the service request, it sends the processing result to the resource cloud; S9. The load balancing module reads the processing result from the resource cloud and sends it to the service gateway; S10. The service gateway sends the processing result to the third-party application through the service interface.

4. The scheduling method for a heterogeneous cryptographic resource pool according to claim 3, wherein: In step S2, after the service request is forwarded to the resource cloud, the service request disconnects.

Images