A data processing method based on a self-service financial device

Abstract

The present application relates to the field of data processing, and more particularly to a data processing method based on a self-service financial device, comprising: determining an edge device end state based on the number of corresponding to-be-processed tasks and the proportion of complex tasks, and determining a task uploading mode as a task combination uploading processing mode or an end combination uploading processing mode according to the edge device end state; and performing to-be-processed task allocation based on a preset allocation strategy, wherein in the preset allocation strategy, the to-be-processed task allocation is performed according to server computing capability or according to cluster stability based on the comparison result of the server computing mean value and the preset server computing mean value; and the present application reduces the influence of the problems that the load adjustment of the cloud platform in the edge cloud collaborative computing applied to the bank self-service device in the prior art has delay, and the determination benchmark cannot adaptively change according to the actual situation, resulting in the problem that the cloud platform load cannot effectively meet the actual use demand.

Description

A data processing method based on self-service financial devices Technical Field

[0001] This invention relates to the field of data processing, and more particularly to a data processing method based on self-service financial equipment. Background Technology

[0002] With the transformation of bank branches towards lighter and smarter operations, self-service financial equipment (such as ATMs, CRS, VTMs, smart counters, and super counters) has become the core carrier for diverting counter business and extending services to lower levels. Self-service financial equipment often generates a large amount of data, which not only includes traditional transaction records but also adds multimodal information such as high-definition images, voice, and facial features. However, traditional centralized data processing is difficult to effectively cope with the generation of massive amounts of real-time data. Therefore, technical personnel usually manage data effectively through the collaboration of edge computing and cloud platforms.

[0003] Chinese Patent Publication No. CN105872109B discloses a cloud platform load operation method. The method includes: the control node of the cloud platform calculates the load balancing degree and business scheduling efficiency of the data server to select the optimal business scheduling strategy, thereby improving the throughput of the cloud platform data server and optimizing the external service performance of the data server. It can be seen that the above technical solution has the following problems: the load can only be adjusted according to the real-time load balancing degree and business scheduling efficiency, which has the delay of adjustment, and the single judgment criterion of load balancing degree and business scheduling efficiency cannot effectively meet the actual use needs. Summary of the Invention

[0004] To address this, the present invention provides a data processing method based on self-service financial equipment, which overcomes the problems in the prior art where the load adjustment of the cloud platform in edge cloud collaborative computing applied to bank self-service equipment is delayed and the judgment benchmark cannot be adaptively changed according to the actual situation, resulting in the cloud platform load not being able to effectively meet the actual usage needs.

[0005] To achieve the above objectives, the present invention provides a data processing method based on self-service financial equipment, comprising:

[0006] The edge device status is determined based on the number of pending tasks and the proportion of complex tasks, and the task upload method is determined as either task combination upload processing method or edge combination upload processing method based on the edge device status.

[0007] In the task combination upload processing method, a first extraction sequence and a second extraction sequence are constructed based on the task category of the task to be processed, and several task sets are constructed.

[0008] In the end-to-end combined upload processing method, the end set is constructed by combining and matching based on the number of tasks to be processed, and the tasks to be processed corresponding to the end set are assigned to the server.

[0009] Tasks to be processed are allocated based on a preset allocation strategy. The preset allocation strategy determines the allocation of tasks based on the server's computing power or the cluster's stability by comparing the server's average calculation value with the preset average calculation value.

[0010] Furthermore, the edge device status is determined based on the number of pending tasks and the proportion of complex tasks at the edge device. The edge status includes:

[0011] The states are: a first state where the number of pending tasks is greater than the preset number of pending tasks and the proportion of complex tasks is within the first preset range of complex task proportions; and a second state where the number of pending tasks is greater than the preset number of pending tasks and the proportion of complex tasks is within the second preset range of complex task proportions.

[0012] Furthermore, the task upload method is determined based on the corresponding endpoint status of the edge device;

[0013] If the edge device's corresponding endpoint status is the first endpoint status, then the task upload method is the task combination upload processing method;

[0014] If the edge device's corresponding endpoint status is the second endpoint status, then the task upload method is the endpoint combination upload processing method.

[0015] Furthermore, the task combination upload processing method includes:

[0016] A first extraction sequence including a first task and a second extraction sequence including a second task are constructed respectively. In the first extraction sequence, the first tasks are arranged in descending order of complexity coefficient, and in the second extraction sequence, the second tasks are arranged in ascending order of complexity coefficient.

[0017] Based on a preset extraction strategy, tasks to be processed are extracted from the first and second sequences to be extracted to construct several task sets, and the tasks to be processed in the task sets are assigned to the server.

[0018] Furthermore, the combined upload processing method includes:

[0019] Each edge device is combined and matched in descending order of the number of tasks to be processed to construct an edge set, and the tasks to be processed corresponding to the edge set are assigned to the server.

[0020] When performing combination matching for a single edge device, the edge device is recorded as the target edge device. Other edge devices with a processing smoothness lower than that of the target edge device and with the largest difference in processing smoothness are recorded as composable ends. The composable ends and the target edge device are recorded as a set of ends.

[0021] Furthermore, the processing smoothness at the edge device end is determined based on the touch response coefficient of each target device at the edge device end;

[0022] For any target device, the corresponding touch response coefficient is determined based on the function key execution reference value;

[0023] The touch-based thinking coefficient and the function key execution reference value are negatively correlated.

[0024] Furthermore, the task categories for tasks to be processed include:

[0025] The first task is one where the depth of the operation process is greater than the preset depth of the operation process or the amount of unstructured data to be uploaded is greater than the preset amount of unstructured data to be uploaded.

[0026] And a second task where the operation process depth is less than or equal to the preset operation process depth and the unstructured data upload requirement is less than or equal to the preset unstructured data upload requirement.

[0027] Furthermore, tasks to be processed are allocated based on a preset allocation strategy, which includes:

[0028] The server's average calculation value is checked. If the average calculation value of the server is less than the preset average calculation value of the server, the tasks to be processed are allocated according to the server's computing power.

[0029] If the server's calculated average is greater than or equal to the preset average, then the tasks to be processed will be allocated based on the cluster's stability.

[0030] Furthermore, the cluster stability is determined based on the fluctuation value of task uploads and iterations at the edge device corresponding to the server;

[0031] Cluster stability is positively correlated with the fluctuation value of task upload and iteration.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows: the technical solution of the present invention determines the task category of the task to be processed based on the depth of the operation process and the amount of unstructured data to be uploaded, thereby dividing the task to be processed into a first task and a second task, avoiding the problem of low analysis accuracy caused by unified analysis and processing of tasks. Furthermore, the end status is determined based on the number of tasks to be processed and the proportion of complex tasks corresponding to the edge device, and the end status effectively reflects the actual situation of the tasks to be processed on the edge device. The task upload method is determined based on the end status, making the task upload method more in line with the actual application scenario, thereby improving the effectiveness of the task upload method.

[0033] Furthermore, in the technical solution of the present invention, if the end state corresponding to the edge device is the first end state, the task upload method is the task combination upload processing method. The first end state reflects that the number of tasks to be processed is large and the distribution of the first task and the second task is balanced. Therefore, tasks to be processed with different difficulties are combined to improve the load balance of each server of the cloud platform when allocating tasks in the future, thereby improving data processing efficiency.

[0034] Furthermore, in the technical solution of the present invention, if the end state corresponding to the edge device is the second end state, the task upload method is the end combination upload processing method. The second end state reflects that the number of tasks to be processed is large and the first task accounts for a large proportion. Therefore, by evaluating the task generation speed of each edge device end through processing smoothness, the edge devices end are further combined so that the task generation speed of each combinable end tends to be balanced, thereby improving the load balance of each server of the cloud platform when allocating tasks in the future, and thus improving data processing efficiency.

[0035] Furthermore, in the technical solution of the present invention, the allocation of tasks to be processed is determined based on the comparison result between the average value of server calculation and the preset average value of server calculation, which makes the allocation method more in line with the actual application scenario and thus improves the data processing efficiency of the cloud platform. Attached Figure Description

[0036] Figure 1 is a schematic diagram of the data processing method based on self-service financial equipment according to the present invention;

[0037] Figure 2 is a flowchart illustrating the process of determining the task upload method based on the corresponding end status of the edge device according to the present invention.

[0038] Figure 3 is a flowchart of the preset allocation strategy of the present invention. Detailed Implementation

[0039] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0040] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0041] Please refer to Figures 1 to 3. This invention provides a data processing method based on self-service financial devices, comprising:

[0042] The edge device status is determined based on the number of pending tasks and the proportion of complex tasks, and the task upload method is determined as either task combination upload processing method or edge combination upload processing method based on the edge device status.

[0043] In the task combination upload processing method, a first extraction sequence and a second extraction sequence are constructed based on the task category of the task to be processed, and several task sets are constructed.

[0044] In the end-to-end combined upload processing method, the end set is constructed by combining and matching based on the number of tasks to be processed, and the tasks to be processed corresponding to the end set are assigned to the server.

[0045] Tasks to be processed are allocated based on a preset allocation strategy. The preset allocation strategy determines the allocation of tasks based on the server's computing power or the cluster's stability by comparing the server's average calculation value with the preset average calculation value.

[0046] This invention is applied to edge cloud collaborative computing in bank self-service equipment. The target equipment is a bank's self-service device, including but not limited to ATMs, CRSs, and VTMs. The self-service device acts as an edge computing terminal for business processing, including but not limited to transaction processing and identity registration / authentication. It periodically transmits the pending tasks (business data packets and work data packets) generated after business processing to a cloud platform for processing and storage. In this invention's technical solution, the pending tasks are the business data packets and work data packets obtained from the self-service device's most recent work cycle. For a single business data packet, it includes, but is not limited to, a single transaction. The corresponding single transaction timestamp, transaction type (withdrawal, deposit, transfer, inquiry, payment, etc.), transaction amount, transaction currency, counterparty account / card number, transaction result code, serial number, authorization method (PIN, fingerprint, facial recognition, etc.), on-site collected facial feature vectors or biometric templates, voiceprint feature vectors and image snapshots, etc., and working data packets, including but not limited to device ID, terminal number, branch number, firmware version, cash box / cash slot ID, cash box denomination and number of bills, cash box remaining capacity, maintenance log, on-site light intensity, noise decibels, customer abnormal behavior detection results (loitering, tailgating, obstructing camera, violent damage) and corresponding video clips, VPN tunnel status, TLS handshake time, abnormal connection events, intrusion detection alarms and data packet loss rate, are all contents that are already known to those skilled in the art, and will not be elaborated here.

[0047] This invention utilizes several historical records. Each historical record includes at least the number of pending tasks, function key execution reference values, operation process depth, unstructured data volume, and server-calculated average values ​​for a single historical process. Each historical record is also marked with a pass / fail flag, indicating whether the historical record meets the user's needs. The user, i.e., the data manager of the self-service financial device, can determine whether the historical record meets the user's needs based on the server's processing speed within a user-defined time period. Determining whether the historical record meets the user's needs through self-defined indicators is easily understood by those skilled in the art and will not be elaborated upon further.

[0048] Specifically, the edge device status is determined based on the number of pending tasks and the proportion of complex tasks. The edge status includes:

[0049] The states are categorized into three endpoints: a first endpoint where the number of pending tasks exceeds a preset number of pending tasks and the proportion of complex tasks falls within a first preset range of complex task proportions; a second endpoint where the number of pending tasks exceeds a preset number of pending tasks and the proportion of complex tasks falls within a second preset range of complex task proportions; and a third endpoint where the number of pending tasks is less than or equal to a preset number of pending tasks or the proportion of complex tasks falls within a third preset range of complex task proportions.

[0050] In this invention, the edge device end is a collection of target devices, and each edge device end includes several target devices. The specific target devices corresponding to a single edge device end are set by the user. It can be understood that the user can cluster the target devices that meet the user's distance requirements according to the geographical location and record them as a set, thereby facilitating subsequent management. This is content that is already known to those skilled in the art and will not be elaborated further.

[0051] For an edge device, the number of tasks to be processed is the total number of service data packets and work data packets generated by all target devices corresponding to the edge device in the most recent work cycle. The complexity ratio is equal to the number of first-class tasks in the tasks to be processed and the number of second-class tasks in the tasks to be processed. This invention uses a continuous cyclical work cycle. At the end of each work cycle, the end status is determined. The duration of a single work cycle is set by the user. It can be understood that the higher the generation rate of the total number of service data packets and work data packets per unit time of the edge device, the shorter the duration of the work cycle. A value for the duration of the work cycle is provided, which is 20 minutes.

[0052] The values ​​within the first preset range of complex task proportions are all greater than the first complex task proportion and less than the second complex task proportion. The values ​​within the second preset range of complex task proportions are all greater than or equal to the second complex task proportion. The values ​​within the third preset range of complex task proportions are all less than or equal to the first complex task proportion, wherein the first complex task proportion is less than the second complex task proportion. Users can set the values ​​of the first and second complex task proportions according to their actual scenarios. It can be understood that the first state reflects a balanced distribution of the number of first and second tasks. Therefore, the greater the user's demand for task balance in the first state, the larger the first complex task proportion should be and the closer it is to 50%, and the smaller the second complex task proportion should be and the closer it is to 50%. One specific implementation example is a first complex task proportion of 40% and a second complex task proportion of 60%.

[0053] The preset number of pending tasks can be set by the user according to actual needs. It can be understood that the larger the number of pending tasks, the greater the data processing demand. Therefore, the greater the user's tolerance for data processing pressure, the larger the preset number of pending tasks. A method for setting this number is provided: extract the number of pending tasks corresponding to the historical records that meet the user's needs, remove outliers, and record the average number of pending tasks after removing outliers as the preset number of pending tasks. The method for removing outliers can be, but is not limited to, the 3σ criterion or the IQR method.

[0054] Specifically, the task upload method is determined based on the corresponding endpoint status of the edge device;

[0055] If the edge device's corresponding endpoint status is the first endpoint status, then the task upload method is the task combination upload processing method;

[0056] If the edge device's corresponding endpoint status is the second endpoint status, then the task upload method is the endpoint combination upload processing method.

[0057] It is understandable that different edge devices may have different endpoint states, so different edge devices may have different task upload methods.

[0058] Specifically, the task combination upload processing method includes:

[0059] A first extraction sequence including a first task and a second extraction sequence including a second task are constructed respectively. In the first extraction sequence, the first tasks are arranged in descending order of complexity coefficient, and in the second extraction sequence, the second tasks are arranged in ascending order of complexity coefficient.

[0060] Based on a preset extraction strategy, tasks to be processed are extracted from the first and second sequences to be extracted to construct several task sets, and the tasks to be processed in the task sets are assigned to the server.

[0061] Preset extraction strategies include:

[0062] S1, extract the first task that is not currently included in the task set and has the largest order in the first sequence to be extracted;

[0063] S2, extract the second task that is currently not included in the task set and has the largest order in the second sequence to be extracted;

[0064] S3, the first and second tasks extracted from S1 and S2 are recorded as a task set;

[0065] S4, if there is a first task not included in the task set in the first sequence to be extracted and a second task not included in the task set in the second sequence to be extracted, then proceed to step S1; if there is no task to be processed in either the first sequence to be extracted or the second sequence to be extracted, then proceed to step S5; if there is no first task not included in the task set in the first sequence to be extracted and no second task not included in the task set in the second sequence to be extracted, then proceed to step S6.

[0066] S5, Record the tasks in the sequence that do not have any pending tasks as a separate task set;

[0067] S6, End.

[0068] For any task to be processed, its complexity coefficient = depth of operation process of the task to be processed / preset depth of operation process + amount of unstructured data to be uploaded by the task to be processed / preset amount of unstructured data to be uploaded.

[0069] Specifically, the combined upload processing method includes:

[0070] Each edge device is combined and matched in descending order of the number of tasks to be processed to construct an edge set, and the tasks to be processed corresponding to the edge set are assigned to the server.

[0071] When performing combination matching for a single edge device, the edge device is designated as the target edge device. Other edge devices with a processing smoothness lower than that of the target edge device and the largest difference in processing smoothness are designated as composable ends. The composable ends and the target edge device are designated as a set of ends. The difference in processing smoothness is the absolute value of the difference between the processing smoothness of the edge device and the target edge device.

[0072] Specifically, the processing smoothness of the edge device is determined based on the touch response coefficient of each target device on the edge device.

[0073] For any target device, the corresponding touch response coefficient is determined based on the function key execution reference value;

[0074] The touch-based thinking coefficient and the function key execution reference value are negatively correlated.

[0075] The processing smoothness of a single edge device is equal to the average of the touch-thinking coefficients of all target devices corresponding to the edge device. For a single target device, its touch-thinking coefficient is equal to the preset function key execution reference value / function key execution reference value. The function key execution reference value is confirmed by obtaining the sub-click duration of each business processed in the most recent work cycle of the target device. For a single business, the sub-click duration is the average of the interval duration between the click times of adjacent function keys in the click sequence during the processing. It can be understood that manual operation is required on the target device during business processing, but the operation process requires clicking the function keys (e.g., "OK", "Start", "Complete", etc.) on the display interface of the target device. All of the above are already known to those skilled in the art and will not be elaborated here.

[0076] The preset function key execution reference value can be set by the user according to actual needs. It can be understood that the smaller the function key execution reference value, the greater the average task processing efficiency in the actual scenario. This can be understood as the presence of staff assistance, resulting in a faster generation speed of the corresponding tasks. Therefore, the greater the user's acceptance of the task generation speed, the smaller the preset function key execution reference value. One method is to extract the function key execution reference values ​​corresponding to the historical records that meet the user's needs, remove outliers, and record the average value of the function key execution reference values ​​after removing outliers as the preset function key execution reference value.

[0077] Specifically, the task categories for tasks to be processed include:

[0078] The first task is one where the depth of the operation process is greater than the preset depth of the operation process or the amount of unstructured data to be uploaded is greater than the preset amount of unstructured data to be uploaded.

[0079] And a second task where the operation process depth is less than or equal to the preset operation process depth and the unstructured data upload requirement is less than or equal to the preset unstructured data upload requirement.

[0080] The operation process depth is the total number of function buttons clicked during the business processing of the task to be processed; the unstructured data upload requirement is the amount of image and video data in the task to be processed, in MB.

[0081] Users can set the preset operation process depth and preset unstructured data upload requirements. It's understood that a greater operation process depth results in a larger unstructured data upload requirement and a greater processing difficulty for the task. Therefore, the greater the user's tolerance for the processing difficulty, the larger the preset operation process depth and preset unstructured data upload requirements will be. One method for setting these values ​​is to extract the operation process depth and unstructured data upload requirements from historical records that meet the user's needs, remove outliers from both values, and record the average values ​​of the removed values ​​as the preset operation process depth and preset unstructured data upload requirements, respectively.

[0082] Specifically, tasks to be processed are allocated based on a preset allocation strategy, which includes:

[0083] The server's average calculation value is checked. If the average calculation value of the server is less than the preset average calculation value of the server, the tasks to be processed are allocated according to the server's computing power.

[0084] If the server's calculated average is greater than or equal to the preset server's calculated average, then the tasks to be processed will be allocated based on the cluster's stability.

[0085] Server computing average = the average computing power of each server on the cloud platform. This can be understood as the current computing power of a server being represented by the percentage of CPU unused. The preset server computing average value can be determined by the user based on their actual needs. Generally, the greater the user's demand for server computing power, the larger the preset server computing average value. One method for determining this value is to extract the server computing average value corresponding to historical records that meet the user's needs, remove outliers, and calculate the average value to record as the preset server computing average value.

[0086] If the task upload method is a task combination upload processing method, the task sets will be allocated in a random order. If the task upload method is an end combination upload processing method, the combinable ends will be allocated in a random order. When allocating tasks to be processed based on server computing power, priority will be given to the server with the largest current computing power. When allocating tasks to be processed based on cluster stability, priority will be given to the server with the largest current cluster stability.

[0087] Specifically, the cluster stability is determined based on the fluctuation value of task uploads and iterations on the edge devices corresponding to the server.

[0088] Cluster stability is positively correlated with the fluctuation value of task upload and iteration.

[0089] For a single server, its corresponding cluster stability is equal to the average value of the task upload iteration fluctuation of each edge device currently assigned to that server. The method for confirming the task upload iteration fluctuation of a single edge device is to detect the end status of the edge device in the last five monitoring periods and record the number of different end statuses in each adjacent period as the task upload iteration fluctuation value.

[0090] In addition, when the terminal state is the third terminal state, where the number of pending tasks is less than or equal to the preset number of pending tasks or the proportion of complex tasks is within the third preset range of complex task proportion, there is no need to determine the task upload method. The pending tasks will be randomly assigned in order, and priority will be given to assigning them to the server with the largest current computing power.

[0091] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0092] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A data processing method based on self-service financial equipment, characterized in that, include: The edge device's state is determined based on the number of pending tasks and the proportion of complex tasks. The task upload method is then determined as either a task combination upload or an edge combination upload based on the edge device's state. In the task combination upload method, a first extraction sequence including a first task and a second extraction sequence including a second task are constructed. The first tasks in the first extraction sequence are arranged in descending order of complexity coefficient, and the second tasks in the second extraction sequence are arranged in ascending order of complexity coefficient. Tasks to be processed are extracted from the first and second extraction sequences based on a preset extraction strategy to construct several task sets, and the tasks in these sets are then assigned to the server. In the edge combination upload method, each edge device... Tasks are combined and matched sequentially in descending order of the number of tasks to be processed to construct an edge set, and the tasks corresponding to the edge set are assigned to the server. Tasks are assigned based on a preset allocation strategy, which determines whether to allocate tasks based on server computing power or cluster stability, based on a comparison between the server's average calculation value and a preset server average calculation value. Cluster stability is determined based on the task upload iteration fluctuation value of the edge device corresponding to the server. Cluster stability and task upload iteration fluctuation value are positively correlated. The method for confirming the task upload iteration fluctuation value of a single edge device is to detect the edge status of the edge device in the last five monitoring periods and record the number of different edge statuses in each adjacent period as the task upload iteration fluctuation value.

2. The data processing method based on self-service financial equipment according to claim 1, characterized in that, The edge device's state is determined based on the number of pending tasks and the proportion of complex tasks. The edge state includes: a first edge state where the number of pending tasks is greater than a preset number of pending tasks and the proportion of complex tasks is within a first preset range of complex task proportions; and a second edge state where the number of pending tasks is greater than a preset number of pending tasks and the proportion of complex tasks is within a second preset range of complex task proportions.

3. The data processing method based on self-service financial equipment according to claim 2, characterized in that, The task upload method is determined based on the terminal status of the edge device. If the terminal status of the edge device is the first terminal status, the task upload method is the task combination upload processing method. If the terminal status of the edge device is the second terminal status, the task upload method is the terminal combination upload processing method.

4. The data processing method based on self-service financial equipment according to claim 3, characterized in that, The edge device combination upload processing method includes: when performing combination matching for a single edge device, the edge device is recorded as the target edge device, and other edge devices with a processing smoothness lower than that of the target edge device and with the largest difference in processing smoothness are recorded as composable ends. The composable ends and the target edge device are recorded as an end set. The processing smoothness of the edge device is determined based on the touch thinking coefficient of each target device corresponding to the edge device. For any target device, its corresponding touch thinking coefficient is determined based on the function key execution reference value. The touch thinking coefficient and the function key execution reference value are negatively correlated. The function key execution reference value is the average of the sub-click duration of each business processed in the most recent work cycle of the target device. For a single business, the sub-click duration is the average of the interval duration between the click times of adjacent function keys in the click sequence during the processing.

5. The data processing method based on self-service financial equipment according to claim 4, characterized in that, The task categories for pending tasks include: the first task, where the operation process depth is greater than the preset operation process depth or the unstructured data upload requirement is greater than the preset unstructured data upload requirement; and the second task, where the operation process depth is less than or equal to the preset operation process depth and the unstructured data upload requirement is less than or equal to the preset unstructured data upload requirement. The operation process depth is the total number of function buttons clicked during the business processing corresponding to the pending task.

6. The data processing method based on self-service financial equipment according to claim 5, characterized in that, Tasks to be processed are allocated based on a preset allocation strategy, which includes: detecting the average computing power of the servers; if the average computing power of the servers is less than the preset average computing power of the servers, then tasks to be processed are allocated according to the computing power of the servers; if the average computing power of the servers is greater than or equal to the preset average computing power of the servers, then tasks to be processed are allocated according to the stability of the cluster.

Images