A monitoring management method and system for multiple wafer detection applications

Abstract

The present invention relates to a monitoring management method and system for multiple wafer detection applications, the method comprising the steps of: obtaining multiple wafer detection applications to be monitored through a monitoring management end, and the multiple wafer detection applications include offline and real-time applications; the monitoring management end obtains node information of several preset deployment nodes, and selects at least one deployment node based on the node information for monitoring and scheduling multiple wafer detection applications; the monitoring management end determines the monitoring point based on the application type of the wafer detection application; the monitoring management end monitors the monitoring point based on the deployment node and determines the operation status of the corresponding wafer detection application, and the operation status includes: abnormal operation and normal operation; when monitoring a wafer detection application with abnormal operation, the monitoring management end schedules and manages the corresponding wafer detection application based on the deployment node. The monitoring management method and system of the present invention can perform real-time monitoring, management and scheduling on various real-time or offline applications.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor manufacturing technology, and in particular to a monitoring management method and system for multiple wafer detection applications. Background Art

[0002] In the production of semiconductor wafers, there are a series of processes including single crystal pulling, slicing, grinding, polishing, layer addition, photolithography, doping, heat treatment, needle testing, and dicing. Chemical vapor deposition, optical development, and chemical mechanical polishing may cause defects on the wafer surface during this process, and defects on the wafer will directly affect the working life and reliability. Therefore, in the wafer production process, it is very important to perform regular or real-time defect detection on the wafers on the production line, and these defect detections are usually implemented through multiple wafer detection applications pre-configured by the user. Based on the analysis and processing results of each wafer detection application, the production line can be adjusted and maintained in real time to ensure the yield of the wafer.

[0003] In order to ensure the smooth progress of various wafer inspection applications, it is necessary to monitor their operating status in real time. However, the existing application scheduling software on the market is more focused on monitoring and scheduling offline scheduled applications, with relatively simple functions and unable to conduct in-depth monitoring and management of various applications. Summary of the invention

[0004] In order to partially alleviate or solve the above technical problems, the first aspect of the present invention provides a monitoring and management method for multiple wafer inspection applications, comprising the steps of:

[0005] Acquire multiple wafer detection applications to be monitored through a monitoring management terminal, wherein the multiple wafer detection applications include: offline applications and real-time applications;

[0006] The monitoring management end obtains node information of several preset deployment nodes, and selects at least one of the deployment nodes based on the node information to monitor and schedule the plurality of wafer inspection applications;

[0007] The monitoring management end determines a corresponding monitoring point based on the application type of the wafer detection application;

[0008] The monitoring management end monitors the monitoring point based on the deployment node and determines the operation status of the corresponding wafer detection application, where the operation status includes: abnormal operation and normal operation;

[0009] When the wafer detection application is monitored to have an abnormal operation, the monitoring management end performs scheduling management on the corresponding wafer detection application based on the deployment node;

[0010] The monitoring and management end is connected to the computing node through the deployment node to monitor and manage the wafer detection application on the computing node.

[0011] When the application type is an offline application, the monitoring point includes: the startup state of the offline application. Accordingly, judging the running state of the corresponding wafer inspection application includes the steps of: regularly monitoring the startup state of the offline application based on the expected startup time of the offline application, and when it is monitored that the startup state of the offline application within the expected startup time is not started, judging that the offline application is abnormal in operation;

[0012] When the application type is a real-time application, the monitoring points include: the online status of the real-time application, and the characteristics of the amount of data currently to be processed by the real-time application.

[0013] Furthermore, in some embodiments, determining the operating status of the corresponding wafer inspection application includes the steps of: when it is monitored that the online status of the real-time application is offline, determining that the real-time application is operating abnormally.

[0014] Further, in some embodiments, determining the operating status of the corresponding wafer inspection application includes the steps of: when it is monitored that the data volume characteristics exceed the processing capacity of the computing node where the real-time application is currently located, determining that the real-time application is operating abnormally.

[0015] Further, in some embodiments, determining the operation status of the corresponding wafer inspection application includes the steps of: when it is monitored that the data volume feature exceeds the historical safety index of the corresponding real-time application, determining that the real-time application is operating abnormally;

[0016] In some embodiments, the step of scheduling and managing the corresponding wafer inspection application by the deployment node includes:

[0017] The deployment node obtains node information of at least one of the computing nodes;

[0018] The deployment node selects an available computing node from at least one of the computing nodes based on the node information;

[0019] The deployment node allocates the wafer inspection application in which the abnormality occurs to the selected computing node;

[0020] In some embodiments, the step of scheduling and managing the corresponding wafer inspection application by the deployment node includes: the deployment node restarting the wafer inspection application.

[0021] In some embodiments, the node information includes: a safe load threshold, a health status, and an enabled status.

[0022] In some embodiments, the step of the deployment node selecting an available computing node from at least one of the computing nodes based on the node information includes:

[0023] When the wafer inspection application in which the abnormality occurs is an offline application, the corresponding deployment node obtains a first load indicator of the offline application, wherein the first load indicator is a regular data volume under normal operation of the offline task, and the first load indicator is preset by the staff when pre-configuring the offline task;

[0024] The deployment node selects the computing nodes whose safety load threshold is not less than the first load indicator as the available computing nodes from the computing nodes that are in a healthy state and are currently not enabled.

[0025] In some embodiments, the step of the deployment node selecting an available computing node from at least one of the computing nodes based on the node information includes:

[0026] When the wafer inspection application in which the abnormality occurs is a real-time application, the corresponding deployment node determines a second load indicator of the real-time application based on a data volume feature of the real-time application;

[0027] The deployment node selects the computing nodes whose safety load threshold is not less than the second load index as available computing nodes from the computing nodes that are in a healthy state and are currently not started.

[0028] In some embodiments, the data volume characteristic is the amount of data flowing in per unit time. Accordingly, the step of determining the second load indicator of the real-time application based on the data volume characteristic of the real-time application includes:

[0029] Acquire the amount of inflow data of the real-time application in the first time period, and determine the amount of inflow data per unit time of the real-time application in the first time period;

[0030] Determine the second load indicator based on the amount of data flowing in per unit time, the second load indicator being the amount of data processing per unit time to be achieved while maintaining the normal operation of the real-time application, and the safe load threshold being the amount of data processing per unit time that the computing node can achieve;

[0031] In some embodiments, the data volume characteristic is the current amount of data to be processed. Accordingly, the second load indicator is the maximum processing time required to complete processing of the data to be processed while ensuring the normal operation of the real-time application, and the safety load threshold is the processing time required for the computing node to complete processing the current data to be processed.

[0032] In some embodiments, the method further comprises the steps of:

[0033] After the wafer inspection application is allocated to the selected computing node, the wafer inspection application in which an abnormality occurs is destroyed or suspended.

[0034] In some embodiments, when an available computing node cannot be selected from the computing nodes, the method further includes the step of: the monitoring management terminal or the corresponding deployment node sends an alarm signal to a staff member indicating that human intervention is required.

[0035] In some embodiments, the step of periodically monitoring the startup status of the offline application based on the expected startup time of the offline application includes:

[0036] Obtaining a set startup time of the offline application, wherein the set startup time is: (T1, T2), wherein T1 is the start time and T2 is the stop time;

[0037] Determine the expected startup time based on the set startup time, wherein the expected startup time is: (T1-ΔT, T2+ΔT'), wherein ΔT is a first preset threshold, and ΔT' is a second preset threshold;

[0038] The startup status of the offline task is monitored within the expected startup time.

[0039] The second aspect of the present invention is to provide a monitoring and management system for multiple wafer detection applications, including: a monitoring and management end; wherein the monitoring and management end includes:

[0040] A target acquisition unit is configured to acquire a plurality of wafer detection applications to be monitored, wherein the plurality of wafer detection applications include: offline applications and real-time applications;

[0041] A deployment unit, configured to obtain node information of a plurality of deployment nodes, and select at least one of the deployment nodes based on the node information, so as to monitor and schedule a plurality of the wafer inspection applications;

[0042] A monitoring point determination unit, configured to determine a corresponding monitoring point based on an application type of the wafer inspection application;

[0043] The operation status monitoring unit is configured to monitor the monitoring point based on the deployment node and determine the operation status of the corresponding wafer inspection application, wherein the operation status includes: abnormal operation and normal operation; wherein,

[0044] When the application type is an offline application, the monitoring point includes: the startup state of the offline application. Accordingly, the running state monitoring unit periodically monitors the startup state of the offline application based on the expected startup time of the offline application, and when it is monitored that the startup state of the offline application within the expected startup time is not started, it is determined that the offline application is abnormally running;

[0045] When the application type is a real-time application, the monitoring points include: the online state of the real-time application, and the data volume characteristics currently to be processed by the real-time application. Accordingly, when the running state monitoring unit monitors that the online state of the real-time application is offline, and / or when the running state monitoring unit monitors that the data volume characteristics exceed the processing capacity of the computing node where the real-time application is currently located, and / or when the running state monitoring unit monitors that the data volume characteristics exceed the corresponding historical safety indicators of the real-time application, the real-time application is judged to be operating abnormally;

[0046] A scheduling management unit, configured to, when the wafer inspection application is monitored to have an abnormal operation, schedule and manage the corresponding wafer inspection application based on the deployment node;

[0047] The monitoring and management end is connected to the computing node through the deployment node to monitor and manage the wafer detection application on the computing node.

[0048] Beneficial Effects

[0049] In the production process of wafers, the production line of wafers often works 24 hours a day. In order to ensure the normal operation of the wafer production line, it is necessary to monitor the operation of each processing step such as the corresponding machine in a timely manner (such as monitoring whether the machine fails). Therefore, it is usually necessary to quickly scan the defects of the wafers obtained in each processing stage, and timely analyze the obtained defect scanning data (or defect detection data) by setting the corresponding wafer detection application, so as to feed back the analysis results (including relevant information such as failure) to the current production line, so that the production line can make timely adjustments or maintenance to each processing technology on the production line based on the analysis results. With the gradual development of chip manufacturing technology and high-end electronic products, higher requirements are also put forward for the output and quality of wafers (smaller scale and higher precision). Correspondingly, under such high-yield and high-standard (small scale, high precision) production requirements, the number of wafer detection applications that need to be arranged in the wafer processing process is also increasing, and the types of wafer detection applications and the layout network are becoming more and more complex. Therefore, the monitoring demand for such a complex application network is more urgent.

[0050] The monitoring and management method provided by the present invention monitors the operating status of different wafer detection applications based on monitoring key monitoring points, and realizes the status monitoring of wafer detection applications with a relatively small monitoring load; at the same time, the present invention also preferably adopts distributed technology to schedule and manage a large number of wafer detection applications monitored (specifically, by arranging a number of computing nodes to run wafer detection applications), so as to quickly schedule and manage abnormal wafer detection applications, and ensure that abnormal wafer detection applications can continue to run. In other words, the present invention selects appropriate monitoring points and uses distributed technology to jointly realize timely monitoring and scheduling management of a large number of complex wafer detection applications. Therefore, the monitoring and management method provided by the present invention can meet the application monitoring needs in a high-yield and high-standard wafer processing environment.

[0051] Furthermore, the present invention effectively reduces the operating cost (i.e., implantation cost) of the entire monitoring solution by selecting monitoring points and utilizing distributed technology. That is, the economic cost required for the monitoring solution proposed by the present invention in the actual equipment configuration process is very small, which is conducive to promotion in wafer processing plants of different sizes. Moreover, the monitoring solution is not complicated and has high feasibility in actual application. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale. Obviously, the drawings described below are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without paying creative labor.

[0053] Figure 1 It is a flowchart of a monitoring management method for multiple wafer inspection applications according to an exemplary embodiment of the present invention;

[0054] Figure 2 It is a schematic diagram of an implementation process of a monitoring management method for multiple wafer detection applications in a specific embodiment of the present invention;

[0055] Figure 3 It is a schematic diagram of the device structure of a monitoring and management system for multiple wafer inspection applications in an exemplary embodiment of the present invention;

[0056] Figure 4 is a structural block diagram of an electronic device in an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] Herein, suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present invention, and have no specific meanings by themselves. Therefore, "module", "component" or "unit" can be used mixedly.

[0059] In this article, "wafer inspection application", also referred to as "application", includes: offline application and real-time application. Among them, "offline application" refers to the application of data analysis for historical wafer inspection data (such as defect scanning data, roughness data, etc.) that has already occurred. For example, the task of analyzing the inspection data of one or more previous processes (flow steps) that have been completed, such as the defect result file processing task. "Real-time application" refers to the application of data analysis for data that is occurring in real time (or the interval between the occurrence time and the current time is short, such as data that only occurred for a few seconds or minutes), or data with strong timeliness that needs to be processed in time. For example, applications that analyze the inspection data of the current process step or the process step that has just been completed, such as Defect Cluster real-time calculation tasks and ADC defect automatic classification tasks.

[0060] Embodiment 1

[0061] See also Figure 1 , is a monitoring and management method for multiple wafer inspection applications according to an exemplary embodiment of the present invention, comprising the steps of:

[0062] S2: obtaining multiple wafer detection applications to be monitored through the monitoring management terminal, and the multiple wafer detection applications include: offline applications and real-time applications;

[0063] S4: The monitoring management end obtains node information of several preset deployment nodes, and selects at least one deployment node based on the node information to monitor and schedule multiple wafer inspection applications;

[0064] S6: The monitoring management end determines the corresponding monitoring point based on the application type of the wafer detection application;

[0065] S8: The monitoring management end monitors the monitoring point based on the deployment node and determines the operation status of the corresponding wafer inspection application, where the operation status includes: abnormal operation and normal operation;

[0066] S10: When a wafer inspection application with abnormal operation is monitored, the monitoring management end schedules and manages the corresponding wafer inspection application based on the deployment node.

[0067] The monitoring and management method proposed in this embodiment does not directly monitor each machine (or in other words, does not directly monitor the various output data of the machine), but selects a wafer detection application (for example, an auxiliary APP for monitoring a specific production indicator) that analyzes and processes the data output by the machine as the monitoring object. In other words, the monitoring method selected by this method is: to monitor the operating status of applications (such as various APPs) used to assist the healthy operation of the wafer production line, thereby indirectly monitoring the ecological environment of the entire wafer production, so as to quickly monitor the abnormalities that may appear or are occurring in the wafer ecological environment. It can be understood that this method strikes a balance between the monitoring scale and the amount of monitoring data, so as to monitor various possible abnormalities in the wafer ecological environment in a timely and comprehensive manner under the condition of a relatively small amount of monitoring data.

[0068] In this embodiment, the above-mentioned offline application and real-time application are preferably run on a pre-created computing node, wherein the computing node is created at the application processing end (ie, a server for executing the application).

[0069] Preferably, in some embodiments, when executing step S8, when the application type is an offline application, the monitoring point includes: the startup state of the offline application, and accordingly, judging the running state of the corresponding wafer inspection application includes the steps of: regularly monitoring the startup state of the offline application based on the expected startup time of the offline application, and when it is monitored that the startup state of the offline application within the expected startup time is not started, judging that the offline application is abnormal in operation;

[0070] Preferably, in some embodiments, when executing step S8, when the application type is a real-time application, the monitoring points include: the online status of the real-time application, and the characteristics of the amount of data currently to be processed by the real-time application.

[0071] Furthermore, in some embodiments, determining the operating status of the corresponding wafer inspection application includes the steps of: when it is monitored that the online status of the real-time application is offline, determining that the real-time application is operating abnormally.

[0072] Further, in some embodiments, determining the operating status of the corresponding wafer inspection application includes the steps of: when it is monitored that the data volume characteristics exceed the processing capacity of the computing node where the real-time application is currently located, determining that the real-time application is operating abnormally.

[0073] For example, in some embodiments, processing capacity refers to a safe load threshold for a computing node.

[0074] Further, in some embodiments, determining the operating status of the corresponding wafer inspection application includes the steps of: when the monitored data volume characteristics exceed the historical safety index of the corresponding real-time application, determining that the real-time application is operating abnormally.

[0075] The monitoring management method in this embodiment can monitor a large number of running offline applications or real-time applications.

[0076] The key monitoring points selected by the monitoring management method in this embodiment can reflect the common fault conditions of different types of wafer inspection applications. Based on the selection of the above monitoring points, by monitoring a portion of the data points of each type of application, it is possible to achieve a more comprehensive and accurate monitoring of the operating status of real-time applications or offline applications.

[0077] Furthermore, in some embodiments, the method can also complete the deployment of each offline application or real-time application to be started. Specifically, before determining the corresponding monitoring point, it also includes the steps of:

[0078] Get the application packages for multiple wafer inspection applications to be monitored (such as Figure 2 As shown, in some embodiments, the staff uploads the application package through the monitoring management terminal);

[0079] The application package is split based on the distributed file system to obtain a plurality of corresponding sub-program packages, wherein the application package includes configuration information set by the user for a plurality of wafer detection applications, and the sub-program package includes: configuration information required for the normal operation of a single wafer detection application;

[0080] The deployment node dispatches the corresponding wafer inspection application to each computing node based on the configuration information of the sub-program package.

[0081] Preferably, in some embodiments, in order to ensure the failover capability of the above-mentioned monitoring and management method, multiple deployment nodes are pre-set for application startup and scheduling. When entering the working phase of monitoring and scheduling, one or more deployment nodes are preferably selected for application monitoring and scheduling management, that is, one or more deployment nodes are set as active nodes (i.e., active nodes), and the remaining deployment nodes are used as standby deployment nodes. Once the current active node fails, the operating data of the current active node can be transferred to one or more standby deployment nodes, that is, a new active node is selected from the standby deployment nodes. Therefore, in the actual monitoring process of the wafer production line, once a sudden problem (such as a deployment node is damaged, or a failure occurs during operation) is encountered, the method has a good failover capability.

[0082] Specifically, in some embodiments, a deployment node may be selected to implement monitoring and scheduling management of multiple applications. In this embodiment, one-to-many application monitoring is implemented through a single deployment node, and the arrangement relationship of the monitoring management system is relatively simple.

[0083] Of course, in other embodiments, when the number of applications is larger or the application network is more complex, two or more deployment nodes may be selected to respectively implement monitoring and scheduling management of multiple applications. In this example, many-to-many application monitoring management is implemented through two or more deployment nodes to reduce the load of a single deployment node. And because the monitoring and scheduling work between each deployment node is independent of each other, when a single deployment node fails, it will not affect the operation of the entire monitoring management system.

[0084] In some embodiments, the node information of the deployment node includes: a safe load threshold, a health status, and an enabled status; wherein the safe load threshold is the computing space available for the deployment node under safe operation, the health status includes: a healthy state and an abnormal state, and the enabled status includes: enabled and not enabled.

[0085] Specifically, in some embodiments, based on step S2, multiple wafer inspection applications to be monitored may be obtained. For example, a list of applications including wafer inspection applications to be monitored may be obtained (e.g., Figure 2 As shown, the staff can set a list of (applications) to be monitored based on the monitoring management terminal). And for those skilled in the art, it is feasible to determine the required computing space to complete the corresponding monitoring scheduling task based on the currently acquired application list. Alternatively, the staff can directly set the required computing space to complete the monitoring task of a specific monitoring scenario (such as monitoring of different production lines or batches) based on historical experience. Therefore, the appropriate deployment node can be selected as the active node based on the required computing space determined above.

[0086] Of course, in other embodiments, for some more conventional monitoring scenarios, the actual needs of the monitoring tasks have been taken into consideration when setting up the deployment nodes, so that the safety load thresholds of the set deployment nodes can basically meet the monitoring needs of each stage.

[0087] Furthermore, in some embodiments, the step of deploying nodes to perform scheduling management on corresponding wafer inspection applications includes:

[0088] Deploy the node to restart the wafer inspection application.

[0089] For example, in some embodiments, when the computing node where the wafer inspection application is located is not damaged, or when the abnormality of the wafer inspection application can be predicted and the abnormality can be resolved by restarting, it is preferred to restart the wafer inspection application by deploying the node so that the wafer inspection application can proceed normally.

[0090] For another example, in some embodiments, when it is detected that a real-time application or an offline application is not online (ie, not started), the application is preferably restarted based on the deployment node.

[0091] Further, in some embodiments, when a direct restart cannot solve the abnormal problem of the application, for example, the computing node where the application is located fails, or the current processing data volume is too large and the computing node cannot meet the computing requirements. The step of the deployment node in S10 scheduling and managing the corresponding wafer inspection application includes:

[0092] The deployment node obtains node information of at least one preset computing node;

[0093] The deployment node selects an available computing node from at least one computing node based on the node information;

[0094] The deployment node distributes the wafer inspection application with abnormality to the selected computing node.

[0095] Preferably, in this embodiment, multiple computing nodes are pre-set at the application processing end (e.g., one or more servers for running applications). Among them, some computing nodes are pre-configured to process corresponding real-time applications or offline applications under normal operation, and some computing nodes are also reserved as standby computing nodes. When an abnormality occurs in a running computing node or an abnormality occurs in an application on a computing node, the standby computing node can be called in time to execute the corresponding application, further ensuring the failover capability of the method.

[0096] In some embodiments, the node information includes: a safety load threshold, a health state, and an enabled state, wherein the health state includes: a healthy state and an abnormal state, and the enabled state includes: enabled and not enabled.

[0097] In some embodiments, the safe load threshold is the computing space available for the computing node under safe operation.

[0098] Preferably, in this embodiment, based on the operation characteristics of offline applications in the field of wafer inspection, the focus is on monitoring whether the offline applications are started and ended normally within the expected startup time.

[0099] Since the amount of data processed by the offline application used to perform wafer inspection and analysis tasks is usually relatively stable, the present embodiment preferably monitors whether the offline application is started or ended normally within the expected time to determine whether the offline application is operating normally.

[0100] For example, the more common offline applications include: defect result file processing tasks, which include the coordinates of defect points, wafer information, and analysis of scanned photos; defect result files are transmitted through Scan or Review machines and are the data basis for all defect analysis statistics. Therefore, the expected scheduling and stable operation of defect result file parsing tasks are the most important links. In this embodiment, the defect result file processing tasks can be monitored by monitoring whether they are started within the expected time and ended within the expected time.

[0101] Furthermore, in some embodiments, when it is monitored that the offline application is not started normally within the expected time, the offline application can be directly restarted through the deployment node. Figure 2 As shown in , when the decision step of "Has the application been started?" is "No", the application is started based on the deployment node, and if the offline application can run normally after restarting, it is continuously monitored whether it can be completed within the expected time. Figure 2 As shown, when the result is yes in the decision step of "Timeout?", the staff can be notified via email to solve the problem with manual assistance.

[0102] Of course, in other embodiments, if the offline application cannot be restored to normal by restarting, the application can be deployed to other computing nodes based on the deployment node.

[0103] Alternatively, when the above operations fail to restore the offline application to normal, the staff can be notified to perform manual intervention.

[0104] For example, in some embodiments, the step of the deployment node selecting an available computing node from at least one computing node based on the node information in step S10 includes:

[0105] When the wafer inspection application in which the abnormality occurs is an offline application, the corresponding deployment node obtains a first load indicator of the offline application, wherein the first load indicator is a regular data volume under normal operation of the offline task, and the first load indicator is preset by the staff when pre-configuring the offline task;

[0106] The deployment node selects computing nodes whose safety load threshold is not less than the first load indicator as available nodes from computing nodes that are in a healthy state and are currently not enabled, wherein the safety load threshold is the computing space available for the deployment node under safe operation.

[0107] It is understandable that, since offline applications involving wafer inspection are usually pre-configured by staff based on inspection and analysis requirements, a preliminary assessment can be made of the amount of data to be processed by the offline application, thereby estimating the first load indicator of the offline application. When the offline application is started normally, the amount of data to be processed by the offline application usually does not exceed the first load indicator.

[0108] Of course, in other embodiments, in order to improve the accuracy of the monitoring and management method, the specific status of the offline application (including the amount of data to be processed, etc.) can also be monitored, and its monitoring and management method can refer to the monitoring and management method of the real-time application.

[0109] For real-time applications, in some embodiments, the step of the deployment node selecting an available computing node from at least one computing node based on the node information in step S10 includes:

[0110] When the wafer inspection application in which the abnormality occurs is a real-time application, the corresponding deployment node determines a second load indicator of the real-time application based on a data volume feature of the real-time application;

[0111] The deployment node selects a computing node whose safety load threshold is not less than a second load indicator as an available computing node from computing nodes that are in a healthy state and are currently not started.

[0112] Further, in some embodiments, the data volume feature is the amount of data flowing in per unit time. Accordingly, the step of determining the second load indicator of the real-time application based on the data volume feature of the real-time application includes:

[0113] Acquire the amount of inflow data of the real-time application in the first time period, and determine the amount of inflow data per unit time of the real-time application in the first time period;

[0114] The second load indicator is determined based on the amount of data flowing in per unit time. The second load indicator is the amount of data processing per unit time required to keep the real-time application running normally, and the safety load threshold is the amount of data processing per unit time that the computing node can achieve.

[0115] For example, in some embodiments, when the amount of data flowing into a real-time application within a unit time (such as within 1 minute) is 10G, in order to avoid time delays or failures, it is necessary to complete the processing of the incoming data at a certain processing speed. For example, it takes 1 minute to process 5G of data, and when the safety load threshold of the computing node can only process 4G of data within 1 minute, that is, the processing capacity of the computing node is limited, it is pre-determined that the current computing node cannot execute the real-time application normally, so that the real-time application can be transferred to other available computing nodes in advance. Alternatively, the staff can be notified to intervene in the processing.

[0116] For another example, in some embodiments, if it is evaluated that a computing node can process 100 Records of data in 1 Second, but now 1,000,000 Records of data suddenly flow into the window, it can be determined that the computing node cannot complete the real-time application within the specified time.

[0117] Alternatively, in some other embodiments, the data volume feature is the current amount of data to be processed. Accordingly, the second load indicator is the maximum processing time I1 required to complete the processing of the data to be processed while ensuring the normal operation of the real-time application. At this time, the safe load threshold is the processing time I2 required for the computing node to complete the processing of the current data to be processed (that is, the fastest processing time). Figure 2 As shown, when the maximum processing time I1 is less than the required processing time I2, then in the decision step of evaluating the inability to process?, a yes determination is made, and resources are temporarily adjusted and restarted (ie, a standby computing node is called and restarted on the standby computing node).

[0118] In some embodiments, when the data volume characteristic is the amount of data flowing in per unit time, the processing capacity (ie, the safety load threshold of the current computing node) is the expected amount of data processed per unit time of the corresponding computing node.

[0119] Alternatively, in other embodiments, the data volume characteristic is the amount of data to be processed, and the processing capacity (ie, the safety load threshold of the current computing node) is the processing time required for the corresponding computing node to complete processing the current data to be processed.

[0120] Alternatively, in other embodiments, the data volume characteristic is the amount of data to be processed, and the processing capacity (ie, the safe load threshold of the current computing node) is the maximum amount of data that the corresponding computing node can process within the safe operating time.

[0121] Since real-time applications are performed on wafer inspection data output in real time, considering the characteristics of wafer production lines, the amount of inspection data flowing into real-time applications may increase dramatically within a certain period of time. In order to avoid the sudden increase in inspection data causing abnormal jams in real-time applications, it is necessary to complete the analysis and processing of the inspection data within a safe time so as not to delay the calculation needs of subsequent inspection data. In this embodiment, the maximum processing time I1 to ensure the normal advancement of real-time applications is evaluated by the current amount of data to be processed, and the maximum processing time is compared with the actual processing time I2 required by the computing node, so that it can be pre-determined whether the current computing node has the corresponding processing capacity. When it is determined that the current computing node has limited processing capacity, scheduling preparations can be made for the current real-time application in advance to avoid failure delays.

[0122] It is understandable that the specific data targeted by the monitoring points of different real-time applications are different. For example, if the real-time application involves wafer map identification, then the real-time application involves processing wafers one by one, and the corresponding monitoring points mainly involve monitoring the number of wafers. In other words, the data volume characteristic is the current amount of wafers to be processed, or the amount of wafers flowing in per unit time. For another example, when the real-time application involves defect scanning statistics, the monitored data volume characteristic is the data volume of the current defect point (defect point) to be processed or the amount of defect point data flowing in per unit time.

[0123] Alternatively, in other embodiments, when the monitored data volume characteristics exceed the corresponding historical safety indicators of the real-time application, the real-time application is judged to be operating abnormally, wherein the historical safety indicators can be set by the staff based on historical generation experience.

[0124] The real-time applications of the wafer usually continuously and dynamically receive new data for processing. In some embodiments, the amount of data processed in history can be analyzed and counted, such as the historical total average amount of data processed by the real-time application in the past, or the average amount of data processed in a certain period of time (equivalent to the above-mentioned historical safety index), etc. When the processing data flow far exceeds the average value, it can be determined that the real-time application is abnormal, and scheduling processing or alarming can be performed to eliminate the abnormality.

[0125] Specifically, in some embodiments, real-time applications with a relatively high running frequency include: Defect Cluster real-time calculation tasks, ADC defect automatic classification tasks, wafer defect detection of wafers, and the like.

[0126] Among them, since clustered point defects on wafers are usually caused by process problems (such as equipment failure) or human errors, such defects are highly repeatable and will appear continuously in a batch of wafers. Therefore. Information about cluster size, geometry, and spatial position is very valuable for process engineers seeking to identify potential production problems. The real-time and stable operation of the Defect Cluster real-time calculation task ensures that engineers can understand the abnormal situation of wafer production in a timely and accurate manner and recover losses. The ADC defect automatic classification task uses an artificial intelligence algorithm to classify scanned defects in real time, greatly reducing labor, improving classification efficiency, and accelerating the statistical analysis steps of defect types.

[0127] As mentioned above, both real-time applications and offline applications are very important for the entire production process of wafers. Through the monitoring and management method in the above embodiment, different real-time applications or offline applications can be monitored and scheduled in a targeted manner to ensure that the running status of real-time applications and offline applications can be stably monitored, and applications that may have abnormalities can be predicted to troubleshoot in advance, or applications that have currently had abnormalities can be monitored and scheduled in a timely manner to ensure that the application will not be delayed in execution. The monitoring and management method in this embodiment can realize real-time or scheduled monitoring of each application under limited load (only monitoring of part of the data corresponding to the monitoring point), and has good failover capabilities in the face of various types of predicted abnormalities or sudden abnormalities.

[0128] In this embodiment, the monitoring management end monitors the node information of each computing node in real time through the deployment node. Among them, the deployment node can dynamically combine the running status of the current application and the usage of the computing node to reasonably manage and plan the computing node (that is, computing resources) to achieve the coordination between the allocation of abnormal applications and computing nodes. In this way, it is ensured that the application with the fault can be smoothly promoted without major delays.

[0129] It is understandable that the deployment node and the computing node can be set on the same server or on different servers.

[0130] Preferably, in some embodiments, the deployment node monitors the node information of the computing node through heartbeat detection. In this embodiment, the deployment node can monitor the activation status and health status of the computing node in real time through heartbeat detection to facilitate subsequent scheduling management.

[0131] Preferably, in some embodiments, the deployment node monitors the relevant information of the monitoring points of the offline application or the real-time application through heartbeat detection.

[0132] For example, in some embodiments, the deployment node determines the node information of the computing node by receiving a heartbeat packet from the application processing end (specifically, each computing node) to select an available computing node.

[0133] For another example, in some embodiments, the deployment node determines the running status (such as whether it is started, etc.) of the corresponding wafer inspection application by receiving heartbeat packets from each computing node.

[0134] For example, in some instances, the deployment node determines the node information of the computing node by receiving a heartbeat packet from the computing node. Specifically, the multiple computing nodes include: computing node a, computing node b... computing node n. If the heartbeat packet from computing node a is not received within the set time I, the health status of computing node a is considered to be abnormal, and the timing starts from the receipt of the last heartbeat packet of computing node a. If the heartbeat packet from computing node a is received within the set time II, the computing node a is considered to be in a healthy state (or restored to normal). If the timing starts from the receipt of the last heartbeat packet of computing node a, if the heartbeat packet from computing node a is not received within the set time III, the computing node a is abnormal and sends an instruction to the deployment node or the application processing end indicating the activation of a new computing node (for example, computing node b). Specifically, when there are spare computing nodes on the application processing end, one of the spare computing nodes is enabled to continue processing the application originally executed on computing node a.

[0135] Further, in some embodiments, the sending interval of the heartbeat packet is preset, for example, once every 3 seconds. When the deployment node or the monitoring management end does not receive the heartbeat packet from the computing node for more than the set time 1, such as 6 seconds, it is judged that the corresponding computing node is abnormal, and the timing starts from the receipt of the last heartbeat packet from the computing node. Of course, the deployment node or the monitoring management end may not be able to know in time what problems actually occur in the computing node. Therefore, in order to normally advance the processing process, only a few preparations can be made.

[0136] If the computing node is only temporarily abnormal and logs in again within the set time II, such as 6 seconds, the corresponding application will be restored to the computing node, that is, the deployment node can restart the application.

[0137] If a computing node has a short-term unrepairable fault, the deployment node will distribute the application executed by the computing node to other available computing nodes within the set time II. Furthermore, in order to ensure processing efficiency, the computing node with the abnormality is forced to go offline, that is, the computing process of the computing node is forced to stop, to prevent abnormal applications from occupying too much computing space.

[0138] It is understandable that setting time I, setting time II, and setting time III can all be flexibly set based on specific application scenarios and monitoring time requirements.

[0139] Preferably, in some embodiments, information with high accuracy and real-time requirements (such as heartbeat packets) can be transmitted using long connections to avoid information loss and ensure the real-time and accuracy of monitoring. Of course, information with relatively low accuracy and real-time requirements (such as information related to monitoring points of offline applications) can also be transmitted using short connections, thereby reducing the amount of data transmission and avoiding resource waste.

[0140] Specifically, in some embodiments, the deployment node and the computing node for processing offline applications are connected through long connection transmission to monitor the running status of the corresponding real-time application; the deployment node and the computing node for processing real-time applications are connected through short connection transmission to monitor the running status of the corresponding offline application.

[0141] Of course, in other embodiments, since computing nodes are reused, a computing node may deploy real-time applications or offline applications. Therefore, in order to simplify the monitoring solution, no matter whether the computing node is used to process real-time applications or offline applications, the computing node preferably communicates with the deployment node through a heartbeat packet, so that the deployment node can understand the health status of the computing node in real time.

[0142] Preferably, in some embodiments, the heartbeat packet includes computing node information, such as: the health status, enabled status, and safe load threshold of the computing node.

[0143] Specifically, in some embodiments, the heartbeat packet also includes one or more types of information such as the host name of the current computing node, heartbeat time, deployed application information, node used memory, used CPU, remaining resources, etc.

[0144] Preferably, in other embodiments, in addition to monitoring the health status of the computing nodes through heartbeat detection, the deployment node can also maintain periodic or real-time communication connections with the computing nodes through other communication methods (for example, connecting through Zigbee or some other wired or wireless communication methods) to obtain node information such as the enabling status and safe load threshold of the computing nodes.

[0145] Furthermore, in some embodiments, when an available computing node cannot be selected from the computing nodes, or when the current scheduling management cannot completely solve the fault problem, the following steps are further included:

[0146] The monitoring management terminal or the deployment node sends an alarm signal indicating that human intervention is required to the staff (such as through email notification).

[0147] Furthermore, in some embodiments, in order to prevent a failed application or computing node (or computing nodes) from occupying computing resources and affecting the operating efficiency of the monitoring system and the wafer detection and analysis system, the steps are also included:

[0148] After the wafer inspection application is allocated to the selected computing node, the wafer inspection application in which the abnormality occurs is destroyed or suspended.

[0149] Specifically, in some embodiments, a timed monitoring strategy is mainly used for offline applications. Accordingly, the steps of regularly monitoring the startup status of offline applications based on the expected startup time of offline applications include:

[0150] Get the set startup time of the offline application, where the set startup time is: (T1, T2), where T1 is the start time and T2 is the stop time;

[0151] Determine an expected startup time based on the set startup time, wherein the expected startup time is: (T1-△T, T2+△T'), wherein △T is a first preset threshold value, and △T' is a second preset threshold value;

[0152] Monitor the startup status of offline tasks within the expected startup time.

[0153] The first preset threshold and the second preset threshold may be preset by a staff member when configuring an offline application.

[0154] In some embodiments, the deployment nodes and the computing nodes may be simultaneously provided at the application processing end.

[0155] For example, in some embodiments, the deployment nodes and the computing nodes may be provided on the same server.

[0156] Alternatively, in other embodiments, the deployment nodes and the computing nodes may be set in different servers.

[0157] The method in this embodiment is based on the characteristics of real-time applications and offline applications in the field of wafer detection, and selects specific monitoring points to monitor the running status of real-time applications and offline applications respectively, so as to predict whether the application will be abnormal or monitor the sudden abnormal state. Specifically, the method monitors the selected representative key monitoring points to monitor multiple wafer detection applications more comprehensively when the monitoring load is relatively small. And the method also realizes a good fault switching capability for the monitoring scheme through the setting of deployment nodes and the application of distributed technology. On the one hand, it can pre-schedule the applications that may be abnormal based on the prediction to prevent sudden abnormalities in the application, so as to avoid delays caused by abnormalities in the application or adverse effects on the entire application operating environment. On the other hand, it can also handle the current sudden abnormal application in a timely manner, such as restarting or rescheduling it based on the deployment node, to ensure that the sudden abnormal application can be monitored and processed in time.

[0158] Embodiment 2

[0159] Based on the monitoring and management method in the above-mentioned embodiment 1, the present invention also provides a monitoring and management system for multiple wafer detection applications.

[0160] In some embodiments, Figure 3 As shown, the monitoring management system includes: a monitoring management terminal 1; wherein the monitoring management terminal 1 includes:

[0161] The target acquisition unit 12 is configured to acquire a plurality of wafer detection applications to be monitored, and the plurality of wafer detection applications include: offline applications and real-time applications;

[0162] A deployment unit 14 is configured to obtain node information of a plurality of deployment nodes, and select at least one of the deployment nodes based on the node information to monitor and schedule the plurality of wafer inspection applications;

[0163] A monitoring point determination unit 16 is configured to determine a corresponding monitoring point based on an application type of the wafer inspection application;

[0164] An operation status monitoring unit 18 is configured to monitor the monitoring point based on the deployment node and determine the operation status of the corresponding wafer inspection application, wherein the operation status includes: abnormal operation and normal operation;

[0165] The scheduling management unit 19 is configured to schedule and manage the corresponding wafer inspection application based on the deployment node when the wafer inspection application with abnormal operation is monitored by the monitoring management end;

[0166] The monitoring management terminal 1 is connected to the computing node through the deployment node to monitor and manage the wafer detection application on the computing node.

[0167] In some embodiments, when the application type is an offline application, the monitoring points include: the startup status of the offline application. Accordingly, the running status monitoring unit periodically monitors the startup status of the offline application based on the expected startup time of the offline application, and when it is monitored that the startup status of the offline application within the expected startup time is not started, the offline application is judged to be running abnormally.

[0168] In some embodiments, when the application type is a real-time application, the monitoring points include: the online status of the real-time application, and the characteristics of the amount of data currently to be processed by the real-time application.

[0169] Furthermore, in some embodiments, when the running status monitoring unit monitors that the online status of the real-time application is offline, it is determined that the real-time application is running abnormally.

[0170] Furthermore, in some embodiments, when the running status monitoring unit monitors that the data volume characteristic exceeds the processing capacity of the computing node where the real-time application is currently located, it is determined that the real-time application is running abnormally.

[0171] Furthermore, in some embodiments, when it is monitored that the data volume characteristic exceeds the corresponding historical safety index of the real-time application, the real-time application is determined to be operating abnormally.

[0172] In some embodiments, the computing node is set at the application processing end 2. The application processing end 2 is provided with a plurality of computing nodes, some of which are used to preferentially execute the assigned real-time applications or offline applications, that is, the application processing units 22 used to process the applications, and the other computing nodes can be used as standby computing nodes 24, which can be immediately activated when an application fails.

[0173] In some embodiments, the monitoring management end 1 is connected to the application processing end 2 to monitor and manage the wafer inspection application running on the application processing end 2 and monitor the node information of the computing node in real time.

[0174] In this embodiment, by selecting key monitoring points of different application types, offline applications and real-time applications can be monitored in a targeted manner during the operation of the application. That is, the monitoring management system in this embodiment can fully realize the timing or real-time monitoring of different types of applications based on the operation characteristics of offline applications and real-time applications.

[0175] Therefore, in this embodiment, there is no need to perform comprehensive monitoring of each real-time application or offline application. Instead, each application can be monitored in a targeted manner based on some selected monitoring points. This allows monitoring and scheduling of various types of applications to be achieved while keeping the monitoring load small (i.e., the amount of monitored data small).

[0176] Furthermore, in some embodiments, the scheduling management unit 19 includes:

[0177] A first information acquisition subunit is configured to acquire node information of at least one of the computing nodes (for example, acquiring node information by deploying a node);

[0178] A first node selection subunit is configured to select an available computing node from at least one of the computing nodes based on the node information;

[0179] The first application allocation subunit is configured to allocate the wafer inspection application in which an abnormality occurs to the selected computing node.

[0180] It is understandable that, in other embodiments, the scheduling management unit 19 does not necessarily need to be set at the monitoring management end, but can be set on the deployment node.

[0181] In some embodiments, the deployment node is set at the monitoring management end. For example, in some embodiments, the function of the scheduling management unit 19 can be implemented by the deployment node.

[0182] Furthermore, in some embodiments, the scheduling management unit 19 further includes: a restart subunit configured to restart the wafer detection application.

[0183] Further, in some embodiments, the node information includes: a safe load threshold, a health status, and an enabled status.

[0184] Further, in some embodiments, the first node selection subunit is configured to obtain a first load indicator of the offline application when the wafer inspection application in which the exception occurs is an offline application, wherein the first load indicator is the regular data volume under normal operation of the offline task, and the first load indicator is preset by the staff when pre-configuring the offline task; and from the computing nodes that are in a healthy state and are not currently enabled, the computing nodes whose safety load threshold is not less than the first load indicator are selected as the available computing nodes.

[0185] Further, in some embodiments, the first node selection subunit is configured to determine a second load indicator of the real-time application based on a data volume characteristic of the real-time application when the wafer inspection application in which the abnormality occurs is a real-time application; and select the computing nodes whose safety load threshold is not less than the second load indicator as available computing nodes from the computing nodes that are in a healthy state and are not currently started.

[0186] Furthermore, in some embodiments, the data volume feature is the amount of data flowing in per unit time, and accordingly, the first node selection subunit further includes: a first calculation subunit for determining a second load index, configured to obtain the amount of data flowing in the real-time application in the first time period, and determine the amount of data flowing in per unit time of the real-time application in the first time period; determine the second load index based on the amount of data flowing in per unit time, and the second load index is the amount of data processing per unit time to be achieved to keep the real-time application running normally. Wherein, the safe load threshold is the amount of data processing per unit time that the computing node can achieve.

[0187] In some embodiments, the data volume characteristic is the current amount of data to be processed. Accordingly, the first node selection subunit further includes: a second computing subunit, configured to determine a second load indicator, the second load indicator being the maximum processing time required to complete processing of the data to be processed while ensuring that the real-time application runs normally, and the safe load threshold being the processing time required for the computing node to complete processing the current data to be processed.

[0188] Furthermore, in some embodiments, the system further includes: a resource releasing unit configured to destroy or suspend the wafer inspection application that has an abnormality after allocating the wafer inspection application to the selected computing node.

[0189] Furthermore, in some embodiments, the system also includes: an alarm unit, configured to send an alarm signal indicating that human intervention is required to a staff member based on the monitoring management terminal or the corresponding deployment node when an available computing node cannot be selected from the computing nodes.

[0190] Embodiment 3

[0191] In a third aspect of the present invention, an electronic device is provided, comprising a memory 502, a processor 501, and a computer program stored in the memory 502 and executable on the processor 501, wherein the processor 501 implements the steps of the method described above when executing the program. For ease of explanation, only the parts related to the embodiments of this specification are shown. For specific technical details not disclosed, please refer to the method section of the embodiments of this specification. The electronic device may include various electronic devices, PCs, network cloud servers, or even mobile phones, tablet computers, PDAs (Personal Digital Assistants), POS (Point of Sales), vehicle-mounted computers, desktop computers, and other electronic devices.

[0192] Specifically, Figure 4The electronic device structure block diagram related to the technical solution provided in the embodiment of this specification is shown. The bus 500 may include any number of interconnected buses and bridges, which link various circuits including one or more processors represented by the processor 501 and the memory represented by the memory 502. The bus 500 can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits together, which are all well known in the art, so they are not further described herein. The communication interface 503 provides an interface between the bus 500 and the receiver and / or transmitter 504, which can be a separate independent receiver or transmitter or the same element such as a transceiver, providing a unit for communicating with various other devices on a transmission medium. The processor 501 is responsible for managing the bus 500 and general processing, and the memory 502 can be used to store data used by the processor 501 when performing operations.

[0193] The computer readable storage medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by an instruction execution system, an apparatus, or a device or used in combination with it. The program code contained on the readable storage medium may be transmitted with any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0194] Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0195] The above-mentioned computer-readable medium carries one or more programs. When the above-mentioned one or more programs are executed by a device, the computer-readable medium realizes the following functions: a plurality of wafer detection applications to be monitored are obtained through a monitoring management end, and the plurality of wafer detection applications include: offline applications and real-time applications; the monitoring management end obtains node information of several preset deployment nodes, and selects at least one deployment node based on the node information for monitoring and scheduling the plurality of wafer detection applications; the monitoring management end determines the corresponding monitoring point based on the application type of the wafer detection application; the monitoring management end monitors the monitoring point based on the deployment node and determines the corresponding running status of the wafer detection application, and the running status includes: abnormal operation and normal operation; wherein, when the application type is an offline application, the monitoring point includes: the startup status of the offline application, and accordingly, the running status of the corresponding wafer detection application is determined. The method comprises the following steps: based on the expected startup time of the offline application, the startup state of the offline application is periodically monitored, and when it is monitored that the startup state of the offline application within the expected startup time is not started, the offline application is judged to be operating abnormally; when the application type is a real-time application, the monitoring points include: the online state of the real-time application, and the data volume characteristics currently to be processed by the real-time application. Accordingly, judging the operating state of the corresponding wafer detection application comprises the following steps: when it is monitored that the online state of the real-time application is offline, and / or when it is monitored that the data volume characteristics exceed the processing capacity of the computing node where the real-time application is currently located, and / or when it is monitored that the data volume characteristics exceed the historical safety indicators of the corresponding real-time application, the real-time application is judged to be operating abnormally; when the wafer detection application with operating abnormality is monitored, the monitoring management end schedules and manages the corresponding wafer detection application based on the deployment node.

[0196] Those skilled in the art will appreciate that the above modules can be distributed in the device according to the description of the embodiment, or can be changed accordingly and only used in one or more devices different from the embodiment. The modules of the above embodiments can be combined into one module, or further divided into multiple sub-modules.

[0197] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a disk, an optical disk, a CD-ROM, a U disk, a mobile hard disk, etc.), and includes a number of instructions for a computer terminal (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0198] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0199] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A monitoring and management method for multiple wafer inspection applications, characterized in that: Includes steps: Acquire multiple wafer detection applications to be monitored through a monitoring management terminal, wherein the multiple wafer detection applications include: offline applications and real-time applications; The monitoring management end obtains node information of several preset deployment nodes, and selects at least one of the deployment nodes based on the node information to monitor and schedule the plurality of wafer inspection applications; The monitoring management end determines a corresponding monitoring point based on the application type of the wafer detection application; The monitoring management end monitors the monitoring point based on the deployment node and determines the corresponding operation status of the wafer detection application, and the operation status includes: abnormal operation and normal operation; wherein, When the application type is an offline application, the monitoring point includes: the startup state of the offline application. Accordingly, judging the running state of the corresponding wafer inspection application includes the steps of: regularly monitoring the startup state of the offline application based on the expected startup time of the offline application, and when it is monitored that the startup state of the offline application within the expected startup time is not started, judging that the offline application is abnormal in operation; When the application type is a real-time application, the monitoring points include: the online state of the real-time application, and the data volume characteristics currently to be processed by the real-time application. Accordingly, judging the running state of the corresponding wafer inspection application includes the steps of: when it is monitored that the online state of the real-time application is offline, and / or when it is monitored that the data volume characteristics exceed the processing capacity of the computing node where the real-time application is currently located, and / or when it is monitored that the data volume characteristics exceed the corresponding historical safety indicators of the real-time application, then judging that the real-time application is running abnormally; When the wafer detection application is monitored to have an abnormal operation, the monitoring management end performs scheduling management on the corresponding wafer detection application based on the deployment node; The monitoring and management end is connected to the computing node through the deployment node to monitor and manage the wafer detection application on the computing node.

2. The monitoring management method according to claim 1, characterized in that: The step of the deployment node scheduling and managing the corresponding wafer inspection application includes: The deployment node obtains node information of at least one of the computing nodes; The deployment node selects an available computing node from at least one of the computing nodes based on the node information; The deployment node distributes the wafer inspection application in which the abnormality occurs to the selected computing node; Alternatively, the step of the deployment node scheduling and managing the corresponding wafer detection application includes: the deployment node restarting the wafer detection application.

3. The monitoring management method according to claim 2, characterized in that: The node information includes: safe load threshold, health status and enabled status.

4. The monitoring management method according to claim 3, characterized in that: The step of the deployment node selecting an available computing node from at least one computing node based on the node information includes: When the wafer inspection application in which the abnormality occurs is an offline application, the corresponding deployment node obtains a first load indicator of the offline application, wherein the first load indicator is a regular data volume under normal operation of the offline task, and the first load indicator is preset by the staff when pre-configuring the offline task; The deployment node selects the computing nodes whose safety load threshold is not less than the first load indicator as the available computing nodes from the computing nodes that are in a healthy state and are currently not enabled.

5. The monitoring management method according to claim 3, characterized in that: The step of the deployment node selecting an available computing node from at least one computing node based on the node information includes: When the wafer inspection application in which the abnormality occurs is a real-time application, the corresponding deployment node determines a second load indicator of the real-time application based on a data volume feature of the real-time application; The deployment node selects the computing nodes whose safety load threshold is not less than the second load index as available computing nodes from the computing nodes that are in a healthy state and are currently not started.

6. The monitoring management method according to claim 5, characterized in that: The data volume characteristic is the amount of data flowing in per unit time. Accordingly, the step of determining the second load indicator of the real-time application based on the data volume characteristic of the real-time application includes: Acquire the amount of inflow data of the real-time application in the first time period, and determine the amount of inflow data per unit time of the real-time application in the first time period; Determine the second load indicator based on the amount of data flowing in per unit time, the second load indicator being the amount of data processing per unit time to be achieved while maintaining the normal operation of the real-time application, and the safe load threshold being the amount of data processing per unit time that the computing node can achieve; and / or, The data volume characteristic is the current amount of data to be processed. Correspondingly, the second load indicator is the maximum processing time required to complete the processing of the data to be processed while ensuring the normal operation of the real-time application. The safe load threshold is the processing time required for the computing node to complete processing the current data to be processed.

7. The monitoring management method according to claim 2, characterized in that: The method further comprises the steps of: After the wafer inspection application is allocated to the selected computing node, the wafer inspection application in which an abnormality occurs is destroyed or suspended.

8. The monitoring management method according to claim 2, characterized in that: When an available computing node cannot be selected from the computing nodes, the method further comprises the step of: the monitoring management terminal or the corresponding deployment node sending an alarm signal indicating that human intervention is required to a staff member.

9. The monitoring management method according to claim 4, characterized in that: The step of regularly monitoring the startup state of the offline application based on the expected startup time of the offline application includes: Obtaining a set startup time of the offline application, wherein the set startup time is: (T1, T2), wherein T1 is the start time and T2 is the stop time; Determining the expected startup time based on the set startup time, wherein the expected startup time is: (T1-ΔT, T2+ΔT'), wherein ΔT is a first preset threshold, and ΔT' is a second preset threshold; The startup status of the offline task is monitored within the expected startup time.

10. A monitoring and management system for multiple wafer inspection applications, characterized in that: include: Monitoring and management terminal; wherein the monitoring and management terminal includes: A target acquisition unit is configured to acquire a plurality of wafer detection applications to be monitored, wherein the plurality of wafer detection applications include: offline applications and real-time applications; A deployment unit, configured to obtain node information of a plurality of deployment nodes, and select at least one of the deployment nodes based on the node information, so as to monitor and schedule a plurality of the wafer inspection applications; A monitoring point determination unit, configured to determine a corresponding monitoring point based on an application type of the wafer inspection application; The operation status monitoring unit is configured to monitor the monitoring point based on the deployment node and determine the operation status of the corresponding wafer inspection application, wherein the operation status includes: abnormal operation and normal operation; wherein, When the application type is an offline application, the monitoring point includes: the startup state of the offline application. Accordingly, the running state monitoring unit periodically monitors the startup state of the offline application based on the expected startup time of the offline application, and when it is monitored that the startup state of the offline application within the expected startup time is not started, it is determined that the offline application is abnormally running; When the application type is a real-time application, the monitoring points include: the online state of the real-time application, and the data volume characteristics currently to be processed by the real-time application. Accordingly, when the running state monitoring unit monitors that the online state of the real-time application is offline, and / or when the running state monitoring unit monitors that the data volume characteristics exceed the processing capacity of the computing node where the real-time application is currently located, and / or when the running state monitoring unit monitors that the data volume characteristics exceed the corresponding historical safety indicators of the real-time application, the real-time application is judged to be operating abnormally; A scheduling management unit, configured to, when the wafer inspection application is monitored to have an abnormal operation, schedule and manage the corresponding wafer inspection application based on the deployment node; The monitoring and management end is connected to the computing node through the deployment node to monitor and manage the wafer detection application on the computing node.

Images