Computer backup data monitoring system

Abstract

The invention relates to the field of computer data management, and discloses a computer backup data monitoring system which comprises a task state monitoring module, an anomaly prediction and propagation module, a global optimization scheduling module, a user interaction module and a system database. By collecting task states in real time, constructing an exception propagation model and dynamically generating a resource allocation and scheduling scheme based on an optimized objective function, task monitoring, exception prediction and suppression and efficient utilization of resources are realized. The user interaction module provides state display and scheduling adjustment functions, and the system database stores task historical states and optimization data and supports cooperative work of the modules. The system solves the problems of poor dynamic monitoring capability, abnormal response lagging and low resource allocation efficiency in the prior art, and has relatively high flexibility and expansibility.

Description

Technical Field

[0001] The invention relates to the technical field of computer data management, in particular to a computer backup data monitoring system. Background Art

[0002] In existing computer backup systems, task status monitoring and exception handling are mostly based on static rules or predefined fixed models. These methods are often designed to solve task management problems in a single scenario, but in actual operation, due to the dynamic nature of backup tasks and the unpredictability of exceptions, these traditional methods gradually expose many shortcomings.

[0003] First, in the existing technology, the monitoring of task status mainly relies on the periodic collection and storage of fixed data such as task completion rate and resource allocation ratio. However, this method lacks dynamic modeling of task status and cannot capture the changing trend of task status in real time. For example, when the task completion rate suddenly slows down, the traditional monitoring method cannot quickly warn. This lag directly leads to the system's insufficient response speed to potential anomalies, which may cause task accumulation or even system failure.

[0004] Second, existing anomaly prediction and processing methods are often based on simple single-point anomaly detection, such as threshold judgment of the completion time or resource allocation ratio of a single task. This approach ignores the complex relationship between tasks. For example, anomalies may spread from one task to other related tasks, and the scope of influence gradually expands. Traditional technology cannot describe the behavior and path of such anomaly propagation. Therefore, in processing multi-task systems, the spread of anomalies will cause difficult-to-control global problems.

[0005] In addition, in terms of resource scheduling, existing technologies usually adopt fixed resource allocation strategies or resource allocation methods based on static priorities. This method lacks the ability to adjust the dynamic priority of tasks in real time. Especially when resources are tight or abnormally high, fixed scheduling strategies can easily lead to high-priority tasks not getting enough resource support, while low-priority tasks may occupy too many resources. The result is inefficient resource allocation and the overall efficiency of task completion cannot be guaranteed.

[0006] Finally, in terms of user interaction, existing technologies usually only provide a simple task status display function, and users can only passively observe the progress of tasks. There is a lack of effective interactive means to support users to adjust priorities or limit resource allocation according to the real-time status of tasks. For example, when the abnormal intensity of a task increases rapidly, users are often unable to directly intervene in the optimization process of the system, thus missing the opportunity to reduce the impact of the abnormality. This lack of interactive design limits the flexibility of the system and reduces the management efficiency of the backup system.

[0007] Therefore, the present invention proposes a computer backup data monitoring system to solve the shortcomings of the prior art. Summary of the invention

[0008] In view of the shortcomings of the prior art, the present invention provides a computer backup data monitoring system that can realize dynamic monitoring of backup tasks, prediction and suppression of abnormal propagation paths, and global optimization scheduling of task resource allocation, effectively solving the problems of poor dynamic adaptability, delayed abnormal response, low resource utilization efficiency and insufficient user interactivity in the prior art.

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions: A computer backup data monitoring system, comprising:

[0010] A task status monitoring module is used to collect status information of backup tasks and generate a task status vector, wherein the status information includes task completion rate, resource allocation ratio and task priority;

[0011] The anomaly prediction and propagation module is used to simulate the propagation path and intensity of anomalies in backup tasks and generate anomaly intensity vectors based on the anomaly propagation results;

[0012] The global optimization scheduling module is used to calculate the optimal resource allocation scheme of the task based on the task state vector and the abnormal intensity vector, and adjust the scheduling priority of the task;

[0013] User interaction module, used to display backup task status, exception propagation information and optimization scheduling scheme in real time, and receive task weights and resource limits input by users;

[0014] The system database is used to store historical status information of tasks, exception records, scheduling plans and related parameters of optimization models.

[0015] Preferably, the task status monitoring module includes:

[0016] A data collection unit, used to collect task completion rate, resource allocation ratio and task priority from distributed backup tasks;

[0017] A state modeling unit is used to construct a task state vector based on the collected task state information, and the task state vector represents the dynamic change of the task completion rate;

[0018] The task trend analysis unit is used to predict the task completion time and possible delay based on the task state vector.

[0019] Preferably, the anomaly prediction and propagation module includes:

[0020] an abnormality detection unit, used to detect abnormal initial conditions of the backup task according to the task status information;

[0021] Anomaly propagation simulation unit, used to calculate the propagation path of anomalies based on anomaly diffusion model, which describes the spatiotemporal variation of anomaly intensity;

[0022] The exception feedback unit is used to pass the exception propagation results to the global optimization scheduling module to adjust the scheduling plan.

[0023] Preferably, the anomaly diffusion model is used to describe the diffusion behavior of anomaly intensity in the task network, and the diffusion behavior includes the change of anomaly intensity over time and the propagation range between task nodes.

[0024] Preferably, the global optimization scheduling module includes:

[0025] An optimization calculation unit is used to construct an objective function including task completion rate, resource allocation cost and abnormal intensity;

[0026] A scheduling generation unit is used to solve the optimal resource allocation solution based on the objective function generated by the optimization calculation unit and dynamically adjust the scheduling priority of the task;

[0027] The task adjustment unit is used to send resource allocation instructions to the task status monitoring module according to the optimal resource allocation plan.

[0028] Preferably, the objective function is used to balance the task completion rate, resource allocation cost and abnormal intensity, which is defined as the overall optimization goal in time and task space, and the goals include maximizing the task completion rate, minimizing resource consumption and minimizing the global impact of abnormal intensity.

[0029] Preferably, the user interaction module includes:

[0030] Status display unit, used to display task completion status, exception propagation path and resource allocation plan in real time;

[0031] A parameter adjustment unit, used to receive the task priority and resource limit parameters input by the user, and pass the adjusted parameters to the global optimization scheduling module;

[0032] Feedback unit, used to provide optimization results and abnormal response suggestions to users.

[0033] Preferably, the system database includes:

[0034] A historical status storage unit, used to store historical status information of backup tasks, including task completion rate and resource allocation records;

[0035] An exception record storage unit is used to store exception information that occurs during the backup task, including the initial conditions and propagation paths of the exception;

[0036] The optimization parameter storage unit is used to store the weight parameters of the global optimization objective function and the model calculation results.

[0037] Preferably, when the anomaly propagation result exceeds a preset threshold, the anomaly prediction and propagation module sends an anomaly feedback signal to the global optimization scheduling module. The global optimization scheduling module dynamically adjusts resource allocation to prioritize tasks in the anomaly area and simultaneously optimizes the scheduling order of remaining tasks.

[0038] Preferably, a computer backup data monitoring device comprises:

[0039] A processor, for executing program logic related to a task status monitoring module, an abnormality prediction and propagation module, and a global optimization scheduling module;

[0040] Memory, used to store task status information, exception records, optimization model parameters and scheduling plans;

[0041] A communication module is used to communicate with the task nodes of the distributed backup system to collect task status information and issue resource allocation instructions in real time;

[0042] The display module is used to display the task status, exception propagation path and scheduling optimization results.

[0043] The present invention provides a computer backup data monitoring system, which has the following beneficial effects:

[0044] 1. The present invention adopts a technical solution that combines a task status monitoring module with an anomaly prediction and propagation module. By real-time collection of status data such as task completion rate and resource allocation ratio, and dynamic modeling and prediction of anomaly propagation, the present invention achieves the technical effect of accurately predicting abnormal behavior and responding quickly. The principle is to describe the dynamic behavior of anomaly propagation through partial differential equations, combine the time series of task completion rate and resource allocation, and simulate the diffusion of anomaly intensity over time and task network. Compared with the technical solution that relies on static anomaly detection in the prior art, the present invention overcomes the shortcomings of lag and limitations in anomaly processing, and fundamentally improves the accuracy and real-time performance of anomaly prediction.

[0045] 2. The present invention integrates the task completion rate, resource allocation ratio and abnormal intensity into the optimization objective function through the global optimization scheduling module, and uses the variational method or gradient optimization method to calculate the optimal resource allocation plan, thereby achieving the technical effect of efficient resource utilization and dynamic optimization of task scheduling. The principle is to unify the maximization of task completion rate, minimization of resource allocation cost and minimization of abnormal impact into a mathematical optimization problem, and introduce constraints (such as total resource constraints and task priority restrictions) to ensure the feasibility of the optimization results. Compared with the technical solutions of fixed priority or static resource allocation in the prior art, the present invention solves the problems of uneven resource allocation and insufficient ability to suppress abnormal tasks, enabling the system to adapt to multi-task scheduling requirements in a dynamic environment.

[0046] 3. The present invention stores task status, exception records, optimized scheduling schemes and model parameters as time series data through the design and implementation of the system database, and supports real-time query and historical data analysis, achieving the technical effect of efficient data management and dynamic optimization support. The principle is to store task status data and exception propagation paths in a multi-dimensional structured manner, and provide a data basis for the generation of subsequent scheduling schemes by optimizing the historical evaluation of the objective function. Compared with the technical solution in the prior art that the database only stores static task status data, the present invention solves the shortcomings of data isolation and low query efficiency, and ensures data sharing and call efficiency when multiple modules collaborate.

[0047] 4. The present invention adopts a user interaction module, which provides users with the ability of real-time monitoring and parameter adjustment through intuitive task status display and dynamic adjustment functions, achieving the technical effect of efficient interaction between users and the system. The principle is to enable users to quickly understand the current system operation status by real-time display of task completion rate, resource allocation ratio, abnormal propagation path and scheduling optimization plan, and directly affect the scheduling optimization results by adjusting task priority and resource restrictions. Compared with the technical solutions in the prior art where user participation is limited and only static data can be observed, the present invention solves the shortcomings of users' difficulty in participating and adjusting in dynamic scheduling systems, and improves user experience and system flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Diagram of the architecture of the computer backup data monitoring system. DETAILED DESCRIPTION

[0049] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only 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.

[0050] Please see attached Figure 1 An embodiment of the present invention provides a computer backup data monitoring system. The following describes in detail each module of the system of the present invention.

[0051] Task status monitoring module

[0052] The task status monitoring module in the present invention is one of the core basic modules of the backup data monitoring system. It mainly realizes the real-time collection, modeling and trend analysis of the distributed backup task status, generates the task status vector and provides accurate data support for subsequent modules (such as the abnormal prediction and propagation module and the global optimization scheduling module). The task status monitoring module dynamically generates an overall description of the task status by collecting multi-dimensional status information such as the completion rate, resource allocation ratio, and task priority of the backup task. At the same time, through trend analysis, the changing trend of the task status is predicted to provide basic information for abnormal prediction. In general, the task status monitoring module is also linked with the system database to store historical status data to provide support for subsequent analysis and optimization.

[0053] In the system architecture of the present invention, the task status monitoring module is highly connected with other modules. First, the task state vector generated by the module is an important input for the operation of the abnormal prediction and propagation module, and its integrity and accuracy directly determine the effect of the abnormal propagation simulation. Secondly, the data output of the task status monitoring module is also an important basis for the global optimization scheduling module to perform resource allocation and scheduling optimization. In addition, while the task status monitoring module updates the status data in real time, it also displays the trend analysis results in conjunction with the user interaction module to enhance the transparency and user participation of the system.

[0054] In this embodiment, the specific implementation of the task status monitoring module is as follows:

[0055] In a possible implementation, the task status monitoring module is composed of the following three units: a data acquisition unit, a status modeling unit, and a task trend analysis unit. The units work together to achieve comprehensive monitoring of the backup task status.

[0056] Implementation of data acquisition unit

[0057] In general, the data acquisition unit communicates with the interface of the distributed backup system to obtain multi-dimensional status data of the backup task in real time, including the following:

[0058] Task completion rate data: The task completion rate is the core indicator that describes the progress of task execution. It is defined as the ratio of the amount of completed backup data to the total amount of data. The calculation formula is:

[0059]

[0060] Among them, xi (t): represents the completion rate of task i at time t; D backup (i,t): indicates the amount of data that has been backed up by task i at time t; D total (i): represents the total data size of task i.

[0061] The real-time update of the completion rate is achieved by periodically collecting the backup progress. As an option, the data collection unit can sample and filter the task completion rate data to remove the abnormally fluctuating instantaneous values ​​and improve the data quality.

[0062] Resource allocation ratio data: The resource allocation ratio of a task is used to describe the distribution of system resources (such as bandwidth, CPU, and memory). The calculation formula is as follows:

[0063]

[0064] in:

[0065] u i (t): represents the resource allocation ratio of task i at time t;

[0066] R alocated (it): represents the amount of resources allocated to task i at time t;

[0067] R total : Indicates the total available resources of the system.

[0068] In a specific implementation, the resource allocation ratio data is directly provided by the system task manager, and the data collection unit records it in real time.

[0069] Task priority data: Priority is used to distinguish the importance of completing different tasks. The priority value can be preset by the user or dynamically adjusted according to the data importance of the task or business needs. Priority data is represented by an integer, and the larger the value, the higher the priority.

[0070] In some embodiments, the data collection unit may also collect other operating parameters of the distributed backup nodes, such as storage load, network delay, and error rate. These parameters may be used as extended data input to provide more comprehensive information for the abnormal prediction module.

[0071] Implementation of state modeling unit

[0072] Specifically, the state modeling unit generates the system's task state vector based on the collected task state data, which is used to describe the overall operation of the backup task. The definition of the task state vector is as follows:

[0073] X(t) = [x 1 (t),x 2 (t),…,xn (t)]

[0074] Where, X(t): represents the task state vector at time t; x i (t): represents the status data of task i, including the completion rate x i (t), resource allocation ratio u i (t) and priority p i (t).

[0075] In a possible implementation, in order to enhance the robustness of the model, the state modeling unit normalizes the task state data. The normalization formula is:

[0076]

[0077] Among them, max(x): represents the maximum value of all task completion rates; max(u): represents the maximum value of all task resource allocation ratios; max(p): represents the maximum value of all task priorities.

[0078] After normalization, each dimension of the task state vector is mapped to a uniform numerical range (such as [0, 1]), which is convenient for calculation and processing in subsequent modules.

[0079] In addition, the state modeling unit also supports the calculation of the state change rate to reflect the dynamic change trend of the task state. The formula for the state change rate is:

[0080]

[0081] in, Represents the rate of change of the completion rate of task i at time t.

[0082] The state change rate calculation results can be used to identify abnormal behavior of the task, such as progress stagnation or abnormal decline.

[0083] Implementation of the Task Trend Analysis Unit

[0084] The task potential analysis unit predicts the task completion time and possible delay risk by analyzing the time series data of the task state vector.

[0085] The estimated completion time of a task is calculated as:

[0086]

[0087] Among them, T i,fnish : indicates the estimated completion time of task i; t: current time; 1-x i (t): indicates the proportion of unfinished tasks; Indicates the current completion rate of the task.

[0088] If T i,finish Exceeding the task deadline T deadline , a delay warning signal is generated and passed to the anomaly prediction and propagation module.

[0089] In another implementation, the task trend analysis unit can also analyze the changing trend of resource allocation ratio and priority. i When (t) continues to decrease, the module will record the task as a resource risk task and prompt the global optimization scheduling module to reallocate resources.

[0090] In some embodiments, the task status monitoring module can also integrate an adaptive data collection mechanism to adjust the collection frequency according to the dynamic changes of task nodes. For example, for tasks with slow completion rate changes, the collection frequency can be reduced to reduce system overhead; while for areas with high abnormality, the collection frequency can be increased to ensure real-time data. This mechanism achieves dynamic adjustment of the collection frequency by calculating the standard deviation of the task status change rate.

[0091] To ensure the security of task status data, the data acquisition unit in this embodiment adopts end-to-end encryption transmission technology. During the transmission process, all status data are protected by encryption algorithms to prevent data leakage or tampering.

[0092] Anomaly prediction and propagation module

[0093] The anomaly prediction and propagation module in the present invention is a key module for realizing abnormal state analysis, propagation path simulation and response optimization of backup tasks. Based on the real-time task state vector provided by the task state monitoring module, this module combines the initial conditions of the anomaly, establishes an anomaly propagation model, dynamically simulates the propagation behavior of the anomaly in the task network, and generates an anomaly intensity vector. The core functions of the anomaly prediction and propagation module include anomaly detection, propagation path simulation and anomaly feedback. It can capture the occurrence of anomalies in real time, predict the impact of anomalies on global tasks, and provide key input information to the global optimization scheduling module to optimize resource allocation solutions.

[0094] The anomaly prediction and propagation module is closely connected with other modules. On the one hand, it relies on the task completion rate, resource allocation ratio, task priority and other data provided by the task status monitoring module to initialize the abnormal state; on the other hand, the anomaly propagation results it generates are directly input into the global optimization scheduling module for dynamic adjustment of resource allocation and priority. In addition, the anomaly prediction and propagation module is also linked with the user interaction module to intuitively present the anomaly propagation path and the status of the affected tasks to the user, so that the user can make auxiliary decisions. The following is a detailed description of the anomaly prediction and propagation module in combination with the module functions and specific technical implementation content.

[0095] In this embodiment, the specific implementation of the abnormal prediction and propagation module is as follows:

[0096] In a possible implementation, the anomaly prediction and propagation module includes an anomaly detection unit, an anomaly propagation simulation unit, and an anomaly feedback unit, and each unit collaborates to complete anomaly detection, propagation simulation, and response.

[0097] Implementation of the anomaly detection unit

[0098] In general, the anomaly detection unit obtains the task completion rate x from the task status monitoring module. i (t), resource allocation ratio u i (t) and task priority p i (t) and other state data, and identify abnormal states through preset rules or calculation methods. The goal of the anomaly detection unit is to provide initial conditions for the anomaly propagation model, that is, the initial value of the anomaly intensity e i (0).

[0099] Specifically, the anomaly detection unit detects task anomalies through the following conditions:

[0100] Task delay: If the rate of change of task completion rate Below the threshold ∈ 1 , then task i is marked as a delayed task;

[0101] Insufficient resources: If the task resource allocation ratio u i (t) is less than the threshold ∈ 2 , then task i is marked as a resource-constrained task;

[0102] Deadline risk: If the predicted completion time of a task is T i,finish Exceeding the task deadline T deadline , then task i is marked as a high-risk task.

[0103] In a specific implementation, the anomaly detection unit generates an initial value of anomaly intensity e according to the above conditions. i (0), the formula is:

[0104]

[0105] Among them, e i (0) is the initial abnormal intensity of task i; x i (t) is the current completion rate of task i, u i (t) is the current resource allocation ratio of task i; T i,finish is the predicted completion time of task i; T deadline is the deadline of task i; α 1 ,α2 ,α 3 is the weight factor, which reflects the contribution of different conditions to the abnormal intensity.

[0106] In some embodiments, the anomaly detection unit may also combine historical anomaly data and task dependencies and utilize a machine learning model (such as a random forest or a support vector machine) to identify anomaly patterns, thereby more accurately generating anomaly initial conditions.

[0107] Implementation of the exception propagation simulation unit

[0108] The core of the anomaly propagation simulation unit is the anomaly diffusion model, which describes the propagation behavior of anomaly intensity in the task network.

[0109] The anomaly propagation simulation unit is based on the anomaly initial condition e i (0) and the task network structure are used to simulate the changes of abnormal states in time t and space x.

[0110] Anomalous Diffusion Model

[0111] The anomaly propagation simulation unit uses partial differential equations to describe the propagation path and range of anomaly intensity. The core formula is:

[0112]

[0113] Where, e(x,t): the distribution of anomaly intensity in the task network, which changes with time t and task node position x; D: anomaly diffusion coefficient, which indicates the speed of anomaly propagation, which is usually determined by the correlation between tasks; The Laplace operator of the anomaly is used to describe the spatial propagation of the anomaly between task nodes; R(e): The damping term of the anomaly, which represents the natural attenuation of the anomaly, is defined as:

[0114] R(e)=γe(x,t)

[0115] Among them, γ>0 is the anomaly decay rate. The larger the value, the faster the anomaly decays.

[0116] In the task network, the diffusion coefficient D can be dynamically adjusted according to the data dependencies between tasks.

[0117] For example, if the data dependency between task i and task j is strong, then D ij A larger value of indicates that the anomaly propagates faster.

[0118] Numerical solution method

[0119] The anomaly propagation simulation unit discretizes the anomaly diffusion equation using the finite difference method and obtains the anomaly propagation path through iterative calculation.

[0120] The time discretization uses the step size Δt, and the space discretization uses the spacing between task nodes Δx. The discretized difference equation is:

[0121]

[0122] Among them, represents the abnormal intensity of task i at the kth time step; represents the abnormal intensity of task i at the k+1th time step.

[0123] The anomaly propagation result is output as an anomaly intensity vector E(t), which is defined as:

[0124] E(t)=[e 1 (t),e 2 (t),…,e n (t)]

[0125] Among them, e i (t) represents the abnormal intensity of task i at time t.

[0126] Implementation of abnormal feedback unit

[0127] The anomaly feedback unit analyzes the output results of the anomaly propagation simulation unit and generates an anomaly warning signal according to the anomaly intensity vector E(t).

[0128] In general, the abnormal feedback unit will generate a high-risk mark for tasks whose abnormal intensity exceeds the preset threshold. The specific rules are as follows:

[0129] If e i (t)>∈ 3 , then task i is marked as a high-risk task.

[0130] The exception feedback unit passes the high-risk task list to the global optimization scheduling module for recalculating the resource allocation plan. In addition, the exception feedback unit also passes the exception propagation path and the affected task list to the user interaction module for intuitive display of the exception status.

[0131] As an option, the anomaly feedback unit can perform path aggregation analysis on the anomaly propagation path, identify key task nodes that may affect multiple high-priority tasks, and prioritize resource allocation to suppress the spread of anomalies.

[0132] In some embodiments, the anomaly propagation simulation unit can also introduce dynamic task weights and adjust the diffusion coefficient D through task weights to more accurately describe the propagation behavior of anomalies in the task network. For example, for high-priority tasks, the weight value is higher, thereby reducing the impact of anomalies on them. In addition, the anomaly feedback unit can generate an anomaly propagation probability map in combination with historical data to assist in resource allocation of the global optimization scheduling module.

[0133] Through the disclosure of the above technical contents, the abnormal prediction and propagation module of the present invention realizes the accurate modeling and propagation simulation of abnormal states, provides important data support for global optimization scheduling, and improves the real-time and stability of the system in dealing with abnormal situations.

[0134] Global optimization scheduling module

[0135] The global optimization scheduling module of the present invention is one of the key modules of the backup data monitoring system, which is used to realize the global optimization and dynamic scheduling of backup task resource allocation. Through the comprehensive analysis of the task state vector and the abnormal intensity vector, the module can dynamically adjust the resource allocation scheme and scheduling strategy of the task according to the current state of the system, the task priority and the abnormal distribution, so as to maximize the system efficiency while meeting the resource constraints, reduce the impact of the abnormality on the task execution, and ensure that the high priority tasks can be completed first. The global optimization scheduling module generates the optimal scheduling scheme and applies it to the system in real time by constructing the objective function and combining the mathematical optimization method or the numerical calculation technology.

[0136] The global optimization scheduling module is highly coordinated with other modules. The task status monitoring module provides real-time updated task status vectors as the basic input data for scheduling. The anomaly prediction and propagation module provides anomaly intensity vectors to identify high-risk areas of tasks, thereby providing constraints and guidance for scheduling optimization. At the same time, the solution generated by the scheduling will be returned to the task status monitoring module, triggering a new round of data collection and status updates, forming a closed-loop optimization feedback mechanism. In addition, the optimization results are displayed to users through the user interaction module so that users can understand the dynamic changes in resource allocation and exception handling in real time. The following is a detailed disclosure of the global optimization scheduling module in combination with the module's functions and technical implementations.

[0137] In this embodiment, the specific implementation of the global optimization scheduling module is as follows:

[0138] In a possible implementation, the global optimization scheduling module includes an optimization calculation unit, a scheduling generation unit and a task adjustment unit, which work together to complete optimization calculation, scheduling generation and dynamic adjustment.

[0139] Optimize the implementation of computing units

[0140] The optimization calculation unit is the core unit of the global optimization scheduling module, which is used to construct and solve the optimization objective function. Generally speaking, the optimization objective function comprehensively considers the task completion rate, resource allocation cost and abnormal intensity, aiming to maximize the efficiency of task completion while reducing resource waste and inhibiting abnormal propagation.

[0141] Definition of objective function

[0142] The global optimization scheduling module aims to maximize the system task completion rate, minimize the impact of abnormalities and minimize resource consumption, and construct the objective function J:

[0143]

[0144] Among them, J: objective function value, used to measure the global scheduling effect of the system, the smaller the better; x(x,t): completion rate of task x at time t, the value range is [0,1], indicating the ratio of the amount of data backed up by the task to the total amount of data; u(x,t): resource allocation ratio of task x at time t, defined as the ratio of the amount of resources obtained by the task to the total amount of resources in the system; e(x,t): abnormal intensity of task x at time t, indicating the abnormal state of the task, the value range is [0,∞); α 1 ,α 2 ,α 3 : Weight coefficient, which represents the influence weight of task completion rate, resource allocation cost and abnormal intensity in the objective function, α 1 ,α 2 ,α 3 >0;t 0 ,t f : The start time and end time of task scheduling; Ω: Task space, which represents the set of all tasks.

[0145] Optimizing the setting of constraints

[0146] In order to ensure the feasibility of optimization, the optimization calculation unit introduces the following constraints:

[0147] Total resource constraints

[0148] ∫ Ω u(x,t)dx≤R total

[0149] Among them, R total : The total available resources of the system, which indicates the total amount of resources that can be allocated by the system during the scheduling cycle.

[0150] Task completion rate range constraints

[0151] x(xt)∈[0,1]

[0152] Where x(x,t) is the completion rate of task x at time t, and its value must be between 0 (not completed) and 1 (completely completed).

[0153] Abnormal strength non-negative constraints

[0154] e(xt)≥0

[0155] Where, e(x,t) is the abnormal intensity of task x at time t, which cannot be a negative value.

[0156] Task priority constraints

[0157] u(xt)≥βp(x)

[0158] Where p(x) is the priority of task x, ranging from 0 to 1. The larger the value, the higher the priority. β is the priority resource allocation factor, which indicates the minimum proportion of resource allocation for high-priority tasks.

[0159] Objective function solution

[0160] The optimization calculation unit solves the above objective function by using the variational method or numerical optimization method. Taking the variational method as an example, the optimization calculation unit constructs the objective functional:

[0161]

[0162] Solve the optimal resource allocation solution through the Euler-Lagrange equation:

[0163]

[0164] Solve the optimal resource allocation solution u (*) (x,t).

[0165] In a specific implementation, the optimization calculation unit uses a gradient descent method for iterative solution. The iterative update formula is:

[0166]

[0167] Among them, u (k) (x, t): resource allocation plan for the kth iteration; η: learning rate, used to control the iteration step size; The partial derivative of the objective function with respect to the resource allocation u.

[0168] Implementation of the scheduling generation unit

[0169] The scheduling generation unit generates a scheduling strategy for the task, including a resource allocation vector U(t) and a priority adjustment vector P(t), according to the resource allocation ratio and abnormal intensity output by the optimization calculation unit.

[0170] Resource Allocation Vector

[0171] The resource allocation vector U(t) describes the resource allocation ratio of all tasks at time t, which is defined as:

[0172] U(t)=[u 1 (t),u 2 (t),…,u n (t)]

[0173] Among them, u i(t) represents the resource allocation ratio of task i at time t.

[0174] The scheduling generation unit will prioritize the allocation of resources in high-risk areas based on the results of the anomaly propagation simulation. i (t) exceeds the threshold ∈ 3 , the scheduling generation unit will allocate more resources to the task.

[0175] Priority Adjustment Vector

[0176] The priority adjustment vector is defined as:

[0177] P(t)=[p 1 (t),p 2 (t),…,p n (t)]

[0178] Among them, p i (t): dynamic priority of task i at time t.

[0179] The scheduling generation unit will prioritize resource allocation for tasks in high-risk areas based on the anomaly propagation simulation results. For example, when the anomaly intensity e of task i is i (t) exceeds the threshold ∈ 3 When , the scheduling generation unit will add resource support for task i.

[0180] The priority adjustment vector P(t) describes the priority change of the task and is used to guide the reallocation of resources. Generally, the priority adjustment vector is determined by the task completion rate and the abnormal intensity.

[0181] The task adjustment unit is responsible for applying the task scheduling results generated by the scheduling generation unit to the system and monitoring the scheduling execution in real time.

[0182] In a possible implementation, the task adjustment unit sends updated resource allocation data to the task status monitoring module through an interface and receives task status feedback. If the actual task status deviates greatly from the optimization result, the task adjustment unit triggers recalculation.

[0183] To improve the efficiency of resource adjustment, the task adjustment unit also supports a gradual adjustment mechanism. In the gradual adjustment, the change of resource allocation ratio is gradually realized by linear interpolation method to avoid the instability of task execution caused by sudden changes in resources.

[0184] In some embodiments, the global optimization scheduling module may also consider the dependencies of tasks. For example, if task A depends on the completion of task B, the module will prioritize the allocation of resources to accelerate the execution of task B. In addition, the scheduling generation unit may dynamically adjust the scheduling cycle in combination with historical data to adapt to the complexity of tasks and changes in system load.

[0185] Through the above implementation, the global optimization scheduling module of the present invention realizes the global optimization of resource allocation and task scheduling, and provides a reliable technical guarantee for the efficient execution of backup tasks and abnormal response.

[0186] User interaction module

[0187] The user interaction module of the present invention is used to provide users with a visual interface for status monitoring, abnormal propagation feedback and scheduling optimization results of backup tasks, and supports users to input adjustment parameters and dynamically adjust resource allocation strategies. The user interaction module works closely with the task status monitoring module, the abnormal prediction and propagation module and the global optimization scheduling module. The core functions of the module include status display, parameter input and feedback response, forming an interactive closed loop between the user and the system.

[0188] Through the user interaction module, users can view the task status (such as completion rate, resource allocation ratio and abnormal propagation path) in real time, and modify the task priority and resource limit when necessary. The module also has an alarm function to remind users to pay attention to abnormal tasks or high-risk areas and assist them in task adjustment operations. In addition, the module is linked with the global optimization scheduling module to pass the user input parameters to the scheduling calculation unit, generate updated resource allocation and scheduling plans in real time, and feed back the optimization results to the user interface.

[0189] The following is a comprehensive description of the user interaction module in combination with the specific functions and technical implementation of the module.

[0190] In this embodiment, the specific implementation of the user interaction module is as follows:

[0191] In a possible implementation, the user interaction module is composed of a state display unit, a parameter input unit, and a feedback response unit. Each unit independently implements different functions, but works together through a data sharing mechanism.

[0192] Implementation of the status display unit

[0193] The status display unit is used to present the output data of the task status monitoring module, the anomaly prediction and propagation module, and the global optimization scheduling module to the user in an intuitive way. Generally, this unit displays the task completion rate, resource allocation ratio, anomaly propagation path, and scheduling optimization results through multi-dimensional dynamic data visualization.

[0194] Task status display

[0195] Task status includes task completion rate x i (t), resource allocation ratio u i (t) and priority p i (t).

[0196] The task completion rate is usually displayed in the form of a real-time updated progress bar. The length of the progress bar indicates the proportion of task completion. The calculation formula is as follows:

[0197]

[0198] Among them, x i (t): completion rate of task i at time t; D backup (i,t): the amount of data backed up by task i at time t; D total (i): The total data volume of task i.

[0199] Resource allocation ratio u i (t) is usually displayed in the form of a bar chart or pie chart. The specific calculation formula is:

[0200]

[0201] Among them, u i (t): resource allocation ratio of task i at time t; R allocated (i,t): the amount of resources allocated to task i at time t; R total : The total available resources of the system.

[0202] Priority p i (t) is usually distinguished by highlighting in the task list, and is calculated as follows:

[0203]

[0204] Among them, p i (t): the priority of task i at time t; w i : The weight value of task i; n: The total number of tasks in the system.

[0205] Abnormal propagation path display:

[0206] The anomaly propagation path is based on the anomaly intensity data generated by the anomaly prediction and propagation module. The anomaly intensity is represented by the depth of color, and the propagation path shows the relationship between tasks through dynamic connections. The anomaly intensity calculation formula is:

[0207]

[0208] Among them, e i (t): the anomaly intensity of task i at time t; γ: the initial intensity of anomaly propagation; λ i : The abnormal attenuation coefficient of task i. The task nodes in the propagation path are represented by different colors to indicate the abnormal risk level, and the color depth is related to the abnormal intensity e i (t) is proportional to.

[0209] Scheduling optimization results display

[0210] The resource allocation scheme U(t) and task priority adjustment vector P(t) generated by the global optimization scheduling module are presented in the form of a trend graph in the status display unit.

[0211] The resource allocation vector is defined as:

[0212] U(t)=[u 1 (t),u 2 (t),…,u n (t)]

[0213] The priority adjustment vector is defined as:

[0214] P(t)=[p 1 (t),p 2 (t),…,p n (t)]

[0215] The trend display of resource allocation schemes helps users understand the dynamic changes of resource allocation, while the real-time display of priority adjustment vectors allows users to quickly grasp the changes in scheduling strategies.

[0216] Implementation of parameter input unit

[0217] The parameter input unit is used to receive the user's adjustments to task priorities and resource constraints, and pass the adjusted parameters to the global optimization scheduling module in real time to trigger the recalculation of the resource allocation plan.

[0218] Priority adjustment function

[0219] The user can adjust the priority of the task through the slider or input box. The value of the adjusted priority p i (t is redefined as:

[0220]

[0221] The user can adjust w i The value of changes the priority weight of the task, which will directly affect the resource allocation calculation of the global optimization scheduling module.

[0222] Resource limit adjustment function

[0223] Users can set the upper limit of resource allocation for tasks max,i , which is used to limit the resource allocation ratio of a specific task. The formula is as follows:

[0224] u i (t)≤u max,i

[0225] Among them, u max,i: The upper limit of resource allocation for task i, input by the user; u i (t): Resource allocation ratio of task i.

[0226] Resource limit adjustment can be completed through the numerical input box or by dragging the slider, and the adjusted parameters are passed to the optimization calculation unit in real time.

[0227] Implementation of feedback response unit

[0228] The feedback response unit is used to provide users with real-time feedback on the adjustment results of the scheduling plan and abnormal status information. Generally, the unit will display the impact of user input on the scheduling results in a dynamically updated graphical manner.

[0229] Scheduling feedback

[0230] The update plan generated by the global optimization scheduling module will be displayed in an intuitive form through the feedback response unit. For example, the changing trend of resource allocation ratio is presented in a line graph, and tasks in abnormally high-risk areas will be highlighted.

[0231] Abnormal alarm feedback

[0232] When the abnormal intensity e i (t) exceeds the preset threshold ∈ 3 When a message is received, the feedback response unit will prompt the user to pay attention through a pop-up window, sound alarm or highlighted color. For example:

[0233]

[0234] Among them, ∈ 3 : Anomaly intensity threshold.

[0235] Decision-making suggestion feedback

[0236] The feedback response unit can provide suggestions to users based on the current task status and the propagation of anomalies. For example, when the intensity of anomalies in a task continues to increase, the system will prompt the user to increase the priority of the task.

[0237] In some embodiments, the user interaction module can be integrated with external devices (such as mobile phones or tablets) to achieve remote monitoring and operation. In addition, the user interaction module supports historical data playback, and users can view historical changes in task status, abnormal propagation, and resource allocation by selecting a specific time point.

[0238] Through the above detailed disclosure, the user interaction module of the present invention realizes a comprehensive display of task status, exception propagation, resource allocation and scheduling optimization, while supporting users to make real-time adjustments to task priorities and resource constraints, forming a closed-loop interactive optimization system, which provides important support for efficient task completion.

[0239] System database

[0240] The system database of the present invention is one of the core support modules of the computer backup data monitoring system, which is used to store and manage task status data, abnormal record data, scheduling plan data and optimization parameter data, and provide real-time data support and historical data query services for each module. The database combines structured and unstructured storage to ensure the real-time, integrity and efficiency of data, and supports parallel access and dynamic update of large-scale task data. The database works closely with the task status monitoring module, abnormal prediction and propagation module, global optimization scheduling module and user interaction module, not only providing key historical data for the system, but also supporting real-time data input and output. The design of the system database is highly scalable, and its performance can be further enhanced through distributed deployment.

[0241] The database not only stores data such as task completion rate, resource allocation ratio, abnormal propagation path, and optimization model parameters, but also provides time series storage and query functions for these data. In addition, the database is connected to the user interaction module through the interface mechanism between modules, supporting the user's playback of historical data and analysis of scheduling strategies. The following combines functional and technical details to fully explain the system database and define the formulas and parameters in detail.

[0242] In this embodiment, the specific implementation of the system database is as follows:

[0243] In a possible implementation, the system database includes a task status storage unit, an exception record storage unit, a scheduling solution storage unit, and an optimization parameter storage unit, and each unit achieves efficient collaboration through a unified data interface.

[0244] Implementation of task status storage unit

[0245] The task status storage unit is used to store the real-time task status data provided by the task status monitoring module, including key data such as task completion rate, resource allocation ratio, task priority, and task completion time prediction. These data are stored in the form of time series, providing input data support for other modules and providing a basis for historical data analysis.

[0246] Storage of task completion rate

[0247] Task completion rate x i (t) is the core indicator of the task running status, and the calculation formula is:

[0248]

[0249] Among them, x i (t): completion rate of task i at time t; D backup(i,t): the amount of data backed up by task i at time t; D total (i): The total data volume of task i.

[0250] The database stores task completion rate data in a table format, with fields including task identifier ID i , timestamp t, task completion rate x i (t).

[0251] The storage of completion rate data supports fast queries based on task identifiers or time ranges.

[0252] Storage of resource allocation ratios

[0253] Task resource allocation ratio u i (t) reflects the proportion of tasks in system resources, and the calculation formula is:

[0254]

[0255] Among them, u i (t): resource allocation ratio of task i at time t; R allocated (i,t): the amount of resources allocated to task i at time t; R total : The total available resources of the system.

[0256] Resource allocation ratios are stored in the form of time series, and each record includes a task identifier, a timestamp, and a resource allocation ratio.

[0257] Storage of task priorities

[0258] Task priority p i (t) determines the priority of resource allocation, and its value is calculated based on the importance of the task, and the formula is:

[0259]

[0260] Among them, p i (t): the priority of task i at time t; w i : The weight value of task i; n: The total number of tasks in the system.

[0261] Priority data is stored in the same structure as task status data, with fields including task identifier, timestamp, and priority value.

[0262] Prediction of task completion time

[0263] Task completion time T i,finish Based on the current completion rate x i (t) and completion rate Calculation, the formula is:

[0264]

[0265] Among them, the estimated completion time of task i; t: current time; 1-x i (t): proportion of unfinished tasks; The rate of change of task completion rate.

[0266] The completion time prediction data storage format is similar to the task status data and can be used for anomaly detection and optimized scheduling.

[0267] Implementation of abnormal record storage unit

[0268] The anomaly record storage unit is used to store the anomaly intensity data, propagation path and classification records generated by the anomaly prediction and propagation module.

[0269] Storage of abnormal intensity data

[0270] Abnormal intensity i (t) represents the abnormal state of the task, and its calculation formula is:

[0271]

[0272] Among them, e i (t): the anomaly intensity of task i at time t; γ: the initial intensity of anomaly propagation; λ i : Abnormal attenuation coefficient.

[0273] The abnormal intensity data storage fields include task identifier, timestamp, abnormal intensity, i (t), and supports query by task and time range.

[0274] Storage of exception propagation paths

[0275] The exception propagation path is stored in a graph structure, where nodes represent tasks and edges represent the exception propagation relationship between tasks. ij represents the abnormal propagation strength between task i and task j, which is defined as:

[0276]

[0277] Among them, C ij : The communication intensity between task i and task j; T ij The communication delay between task i and task j.

[0278] The graph structure is stored in the form of an adjacency matrix, and the dynamic changes of node and edge weights are recorded as time series.

[0279] Storage of abnormal classification records

[0280] Exception records are stored by exception type (such as task delay, insufficient resources, dependency failure, etc.), and each type of exception contains exception intensity and propagation path information.

[0281] Implementation of the Scheduling Scheme Storage Unit

[0282] The scheduling scheme storage unit is used to record the resource allocation scheme and priority adjustment results generated by the global optimization scheduling module.

[0283] Storage of resource allocation plans

[0284] The resource allocation scheme is stored in the form of a vector and is defined as:

[0285] U(t)=[u 1 (t),u 2 (t),…,u n (t)]

[0286] Among them, u i (t): Resource allocation ratio of task i at time t.

[0287] Resource allocation plans are stored by task and timestamp, and support dynamic updates.

[0288] Storage of priority adjustment results

[0289] The priority adjustment result is stored in the form of a vector, defined as:

[0290] P(t)=[p 1 (t),p 2 (t),…,p n (t)]

[0291] Priority adjustment records are used to analyze the impact of resource allocation on task completion.

[0292] Optimize the implementation of parameter storage unit

[0293] The optimization parameter storage unit is used to store the objective function weights and intermediate calculation results of scheduling optimization.

[0294] Storage of objective function weights Objective function weights α 1 ,α 2 ,α 3 The definition is as follows:

[0295]

[0296] Among them, α 1 : The weight of task completion rate; α 2 : The weight of resource allocation cost; α 3 : The weight of the anomaly intensity.

[0297] Weight data is recorded by task and can be adjusted dynamically.

[0298] In some embodiments, the database uses distributed storage technology to store task slices in multiple database nodes to further improve access efficiency. It also supports related queries based on the abnormal propagation path, facilitating abnormal analysis and optimized calculations. The database also uses incremental storage technology to reduce redundant storage of historical records, further improving system performance.

[0299] Through the above detailed disclosure, the system database of the present invention realizes efficient storage, real-time query and historical analysis of backup tasks, and provides important data support for the normal operation of other modules.

[0300] The present invention further provides a computer backup data monitoring device. The specific implementation of the device is described below in conjunction with the working principle of the device of the present invention.

[0301] A processor, for executing program logic related to a task status monitoring module, an abnormality prediction and propagation module, and a global optimization scheduling module;

[0302] Memory, used to store task status information, exception records, optimization model parameters and scheduling plans;

[0303] A communication module is used to communicate with the task nodes of the distributed backup system to collect task status information and issue resource allocation instructions in real time;

[0304] The display module is used to display the task status, exception propagation path and scheduling optimization results.

[0305] The device of this embodiment can be used to execute the above system embodiments, and its principles and technical effects are similar, which will not be repeated here.

[0306] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A computer backup data monitoring system, characterized in that: include: A task status monitoring module is used to collect status information of backup tasks and generate a task status vector, wherein the status information includes task completion rate, resource allocation ratio and task priority; The anomaly prediction and propagation module is used to simulate the propagation path and intensity of anomalies in backup tasks and generate anomaly intensity vectors based on the anomaly propagation results; The global optimization scheduling module is used to calculate the optimal resource allocation scheme of the task based on the task state vector and the abnormal intensity vector, and adjust the scheduling priority of the task; User interaction module, used to display backup task status, exception propagation information and optimization scheduling scheme in real time, and receive task weights and resource limits input by users; The system database is used to store historical status information of tasks, exception records, scheduling plans and related parameters of optimization models.

2. A computer backup data monitoring system according to claim 1, characterized in that: The task status monitoring module includes: A data collection unit, used to collect task completion rate, resource allocation ratio and task priority from distributed backup tasks; A state modeling unit is used to construct a task state vector based on the collected task state information, and the task state vector represents the dynamic change of the task completion rate; The task trend analysis unit is used to predict the task completion time and possible delay based on the task state vector.

3. A computer backup data monitoring system according to claim 1, characterized in that: The anomaly prediction and propagation module includes: an abnormality detection unit, used to detect abnormal initial conditions of the backup task according to the task status information; Anomaly propagation simulation unit, used to calculate the propagation path of anomalies based on anomaly diffusion model, which describes the spatiotemporal variation of anomaly intensity; The exception feedback unit is used to pass the exception propagation results to the global optimization scheduling module to adjust the scheduling plan.

4. A computer backup data monitoring system according to claim 3, characterized in that: The anomaly diffusion model is used to describe the diffusion behavior of anomaly intensity in a task network, and the diffusion behavior includes the change of anomaly intensity over time and the propagation range between task nodes.

5. A computer backup data monitoring system according to claim 1, characterized in that: The global optimization scheduling module includes: An optimization calculation unit is used to construct an objective function including task completion rate, resource allocation cost and abnormal intensity; A scheduling generation unit is used to solve the optimal resource allocation solution based on the objective function generated by the optimization calculation unit and dynamically adjust the scheduling priority of the task; The task adjustment unit is used to send resource allocation instructions to the task status monitoring module according to the optimal resource allocation plan.

6. A computer backup data monitoring system according to claim 5, characterized in that: The objective function is used to balance the task completion rate, resource allocation cost and abnormal intensity, which is defined as the overall optimization goal in time and task space. The goals include maximizing the task completion rate, minimizing resource consumption and minimizing the global impact of abnormal intensity.

7. A computer backup data monitoring system according to claim 1, characterized in that: The user interaction module comprises: Status display unit, used to display task completion status, exception propagation path and resource allocation plan in real time; A parameter adjustment unit, used to receive the task priority and resource limit parameters input by the user, and pass the adjusted parameters to the global optimization scheduling module; Feedback unit, used to provide optimization results and abnormal response suggestions to users.

8. A computer backup data monitoring system according to claim 1, characterized in that: The system database includes: A historical status storage unit, used to store historical status information of backup tasks, including task completion rate and resource allocation records; An exception record storage unit is used to store exception information that occurs during the backup task, including the initial conditions and propagation paths of the exception; The optimization parameter storage unit is used to store the weight parameters of the global optimization objective function and the model calculation results.

9. A computer backup data monitoring system according to claim 1, characterized in that: When the anomaly propagation result exceeds a preset threshold, the anomaly prediction and propagation module sends an anomaly feedback signal to the global optimization scheduling module. The global optimization scheduling module dynamically adjusts resource allocation to prioritize tasks in the anomaly area and simultaneously optimizes the scheduling order of the remaining tasks.

10. A computer backup data monitoring device, applied to the computer backup data monitoring system according to any one of claims 1 to 9, characterized in that: include: A processor, for executing program logic related to a task status monitoring module, an abnormality prediction and propagation module, and a global optimization scheduling module; Memory, used to store task status information, exception records, optimization model parameters and scheduling plans; A communication module is used to communicate with the task nodes of the distributed backup system to collect task status information and issue resource allocation instructions in real time; The display module is used to display the task status, exception propagation path and scheduling optimization results.

Images