A cloud computing task scheduling method based on privacy protection

Abstract

The present invention discloses a cloud computing task scheduling method based on privacy protection, which relates to the field of task processing technology and includes the following steps: S1, obtaining tasks to be scheduled and performing privacy processing on the tasks to be scheduled; S2, after the privacy processing, generating an influence matrix between the tasks to be scheduled and each container based on a number of containers of the cloud platform, and determining the scheduling unit of the tasks to be scheduled according to the influence matrix; S3, completing the scheduling processing according to the scheduling unit of the tasks to be scheduled. The present invention converts the security attributes of the container nodes into vector form through a hash vectorizer, determines the influence matrix in combination with the task priority; then captures the synergistic effect of risk and priority through vector operations, avoiding the one-sidedness of single indicator decision-making; and finally enables the tasks to be scheduled to run in units of containers, reducing interference between tasks. At the same time, the present invention enhances the robustness against unknown attacks and overcomes the defense shortcomings of traditional scheduling methods.

Description

Technical Field

[0001] The present invention relates to the technical field of task processing, and in particular to a cloud computing task scheduling method based on privacy protection. Background Art

[0002] With the rapid development of cloud computing technology, cloud platforms have become a vital infrastructure for businesses and individuals to store data and run applications. However, task scheduling in cloud computing environments faces numerous challenges, with privacy protection and efficient resource utilization being two core issues. To address these challenges, researchers have proposed a variety of privacy-preserving cloud computing task scheduling methods.

[0003] In traditional cloud computing task scheduling, scheduled tasks are typically transmitted and processed in plaintext within the cloud platform, potentially leading to the leakage of user privacy data. For example, tasks may contain sensitive user information, such as identity information and transaction records. Once this information is obtained by malicious attackers, it will cause serious losses to the user. Therefore, privacy protection of scheduled tasks is a key step in ensuring user data security. However, existing scheduling methods often focus on resource utilization or cost as a single objective and lack integrated optimization for privacy protection. For example, tasks may be scheduled into containers that share the same physical machine, exposing the risk of side-channel attacks. Summary of the Invention

[0004] In order to solve the above problems, the present invention proposes a cloud computing task scheduling method based on privacy protection.

[0005] The technical solution of the present invention is: a cloud computing task scheduling method based on privacy protection includes the following steps:

[0006] S1. Obtain the tasks to be scheduled and perform privacy processing on them.

[0007] S2. After privacy processing, based on several containers on the cloud platform, an influence matrix is generated between the task to be scheduled and each container, and the scheduling unit of the task to be scheduled is determined according to the influence matrix;

[0008] S3. Complete the scheduling process according to the scheduling unit of the task to be scheduled.

[0009] Furthermore, in S1, the firewall of the cloud platform is used to set the IP address of the client corresponding to the task to be scheduled in the access whitelist of the cloud platform to complete the privacy processing.

[0010] Furthermore, S2 includes the following sub-steps:

[0011] S21. After privacy processing, obtain several containers from the cloud platform;

[0012] S22. Calculate the impact matrix between the scheduled task and each container based on the node risk factor of each container in the cloud platform;

[0013] S23. Determine the scheduling unit of the task to be scheduled according to the influence matrix between the task to be scheduled and each container.

[0014] The beneficial effect of this further solution is that, in this invention, containers encapsulate the code, runtime libraries, configuration files, and dependencies required by a task, becoming the smallest schedulable unit. The row and column values of the impact matrix are used as the influence weights between tasks and containers to quantify scheduling priorities, preventing high-priority tasks from being assigned to high-risk containers (such as nodes with unpatched vulnerabilities), thereby reducing the risk of data leakage or service interruption. Through risk quantification models and priority-driven scheduling, task execution efficiency is guaranteed while mitigating the risk of data leakage and system intrusion (for example, scheduling financial transaction tasks to compliant containers).

[0015] Furthermore, S22 includes the following sub-steps:

[0016] S221. Convert the node risk factor of the container into a node risk vector using a hash vectorizer.

[0017] S222. Convert the task priority of the task to be scheduled into a priority vector using a hash vectorized container;

[0018] S223, obtaining the difference between the node risk vector and the priority vector;

[0019] S224. Perform a dot product of the node risk vector and the priority vector;

[0020] S225. Calculate the influence matrix between the task to be scheduled and the container using the activation function based on the difference result and the dot product result.

[0021] The beneficial effect of this further solution is that, in this invention, a hash vectorizer converts container node risk factors (such as vulnerability scores, historical intrusion events, and compliance status) into vectors to quantify node security. Task priorities (high / medium / low) are mapped to numerical vectors (e.g., high = 1.0, medium = 0.5, low = 0.1) to reflect task importance, thus converting unstructured security metrics into computable vectors.

[0022] Furthermore, in S225, the influence matrix between the task to be scheduled and the container The expression is:

[0023] ;

[0024] Where, represents the node risk vector, represents the priority vector, Represents a splicing operation, represents the dot product operation, represents the first trainable parameter matrix, represents the second trainable parameter matrix, represents the third trainable parameter matrix, represents the activation function, Represents a transpose operation.

[0025] The beneficial effect of the above further solution is that, in the present invention, the vectors can be concatenated to form a matrix with one row and four columns. Therefore, the first trainable parameter matrix can be a matrix with four rows and one column, the second trainable parameter matrix can be a matrix with one row and one column, and the rows and columns of the third trainable parameter matrix can be determined based on actual conditions and should be a square matrix. The introduction of nonlinearity through the tanh activation function enables the present invention to learn complex relationships and, combined with random perturbations, defend against attacks based on historical behavior speculation. Through trainable parameters, the present invention can automatically adapt to dynamic changes in the cloud platform (such as new vulnerabilities and task modes).

[0026] Furthermore, S23 includes the following sub-steps:

[0027] S231. Calculate the row and column values of the influence matrix between the task to be scheduled and the container, and use the row and column values as the influence weights between the task to be scheduled and the container;

[0028] S232, obtain the buffer overflow vulnerability of the scheduled task on the cloud platform;

[0029] S233. Determine a final scheduling coefficient between the task to be scheduled and the container based on a buffer overflow vulnerability of the task to be scheduled on the cloud platform and an impact weight between the task to be scheduled and the container.

[0030] S234: Use the container with the maximum final scheduling coefficient as the scheduling unit of the task to be scheduled.

[0031] The beneficial effect of the above further scheme is: in the present invention, a buffer overflow vulnerability refers to a phenomenon in which a buffer overflow occurs when a program attempts to put data into a certain location in its memory because there is not enough space. The high-dimensional impact matrix is compressed into a scalar weight to facilitate fast sorting and comparison of container priorities. The row and column values reflect the most significant association patterns between tasks and containers (such as the strong synergistic effect of risk and priority), ignoring minor feature noise. The present invention also identifies buffer overflow vulnerabilities in task codes through static analysis or runtime detection, records the size of the vulnerability memory, and selects the container with the largest d value in combination with the impact weight, thereby achieving a comprehensive optimization of risk, priority and security, and avoiding local optimization (such as scheduling only by resource utilization).

[0032] Furthermore, in S233, the final scheduling coefficient between the task to be scheduled and the container The calculation formula is:

[0033] ;

[0034] Where, Indicates the influence weight between the task to be scheduled and the container. represents a constant, It means taking a random number between 0 and 1. Indicates the attack frequency of the scheduled task on the cloud platform. Indicates the memory size of the buffer overflow vulnerability of the scheduled task on the cloud platform.

[0035] The beneficial effects of this invention are as follows: It uses a hash vectorizer to convert the security attributes of container nodes into vector form, combining them with task priorities to determine an impact matrix; vector operations are then used to capture the synergistic effects of risk and priority, avoiding the one-sidedness of single-metric decision-making; and ultimately, scheduled tasks are run as containers, reducing inter-task interference. Furthermore, this invention enhances robustness against unknown attacks, overcoming the defense shortcomings of traditional scheduling methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Flowchart of the privacy-preserving cloud computing task scheduling method. DETAILED DESCRIPTION

[0037] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0038] like Figure 1 As shown, the present invention provides a cloud computing task scheduling method based on privacy protection, comprising the following steps:

[0039] S1. Obtain the tasks to be scheduled and perform privacy processing on them.

[0040] S2. After privacy processing, based on several containers on the cloud platform, an influence matrix is generated between the task to be scheduled and each container, and the scheduling unit of the task to be scheduled is determined according to the influence matrix;

[0041] S3. Complete the scheduling process according to the scheduling unit of the task to be scheduled.

[0042] In an embodiment of the present invention, in S1, the firewall of the cloud platform is used to set the IP address of the client corresponding to the task to be scheduled in the access whitelist of the cloud platform to complete the privacy processing.

[0043] In this embodiment of the present invention, S2 includes the following sub-steps:

[0044] S21. After privacy processing, obtain several containers from the cloud platform;

[0045] S22. Calculate the impact matrix between the scheduled task and each container based on the node risk factor of each container in the cloud platform;

[0046] S23. Determine the scheduling unit of the task to be scheduled according to the influence matrix between the task to be scheduled and each container.

[0047] In this invention, containers encapsulate the code, runtime libraries, configuration files, and dependencies required for a task, becoming the smallest schedulable unit. The row and column values of the impact matrix are used as the influence weights between tasks and containers to quantify scheduling priorities, avoiding assigning high-priority tasks to high-risk containers (such as nodes with unpatched vulnerabilities), thereby reducing the risk of data leakage or service interruption. Through risk quantification models and priority-driven scheduling, we can ensure task execution efficiency while reducing the risk of data leakage and system intrusion (for example, by scheduling financial transaction tasks to compliant containers).

[0048] In this embodiment of the present invention, S22 includes the following sub-steps:

[0049] S221. Convert the node risk factor of the container into a node risk vector using a hash vectorizer.

[0050] S222. Convert the task priority of the task to be scheduled into a priority vector using a hash vectorized container;

[0051] S223, obtaining the difference between the node risk vector and the priority vector;

[0052] S224. Perform a dot product of the node risk vector and the priority vector;

[0053] S225. Calculate the influence matrix between the task to be scheduled and the container using the activation function based on the difference result and the dot product result.

[0054] In this paper, a hash vectorizer is used to convert container node risk factors (such as vulnerability scores, historical intrusion events, and compliance status) into vectors to quantify node security. Task priorities (high / medium / low) are mapped to numerical vectors (e.g., high = 1.0, medium = 0.5, low = 0.1) to reflect task importance, completing the transformation of unstructured security metrics into computable vectors.

[0055] In the embodiment of the present invention, in S225, the influence matrix between the task to be scheduled and the container The expression is:

[0056] ;

[0057] Where, represents the node risk vector, represents the priority vector, Represents a splicing operation, represents the dot product operation, represents the first trainable parameter matrix, represents the second trainable parameter matrix, represents the third trainable parameter matrix, represents the activation function, Represents a transpose operation.

[0058] In this invention, the concatenated vectors can form a matrix with one row and four columns. Therefore, the first trainable parameter matrix can be a matrix with four rows and one column, the second trainable parameter matrix can be a matrix with one row and one column, and the third trainable parameter matrix can be a square matrix based on actual conditions. The introduction of nonlinearity through the tanh activation function enables the invention to learn complex relationships and, combined with random perturbations, defend against attacks based on historical behavior speculation. Through trainable parameters, the invention can automatically adapt to dynamic changes in the cloud platform (such as new vulnerabilities and task modes).

[0059] In this embodiment of the present invention, S23 includes the following sub-steps:

[0060] S231. Calculate the row and column values of the influence matrix between the task to be scheduled and the container, and use the row and column values as the influence weights between the task to be scheduled and the container;

[0061] S232, obtain the buffer overflow vulnerability of the scheduled task on the cloud platform;

[0062] S233. Determine a final scheduling coefficient between the task to be scheduled and the container based on a buffer overflow vulnerability of the task to be scheduled on the cloud platform and an impact weight between the task to be scheduled and the container.

[0063] S234: Use the container with the maximum final scheduling coefficient as the scheduling unit of the task to be scheduled.

[0064] In the present invention, a buffer overflow vulnerability refers to a phenomenon in which a buffer overflow occurs when a program attempts to put data into a certain location in its memory because there is not enough space. The high-dimensional impact matrix is compressed into a scalar weight to facilitate fast sorting and comparison of container priorities. The row and column values reflect the most significant association patterns between tasks and containers (such as the strong synergistic effect of risk and priority), ignoring minor feature noise. The present invention also identifies buffer overflow vulnerabilities in task codes through static analysis or runtime detection, records the size of the vulnerability memory, and selects the container with the largest d value in combination with the impact weight, thereby achieving a comprehensive optimization of risk, priority and security, and avoiding local optimization (such as scheduling only by resource utilization).

[0065] In the embodiment of the present invention, in S233, the final scheduling coefficient between the task to be scheduled and the container The calculation formula is:

[0066] ;

[0067] Where, Indicates the influence weight between the task to be scheduled and the container. represents a constant, Indicates taking a random number between 0 and 1. Indicates the attack frequency of the scheduled task on the cloud platform. Indicates the memory size of the buffer overflow vulnerability of the scheduled task on the cloud platform.

[0068] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and it should be understood that the scope of protection of the present invention is not limited to such specific descriptions and embodiments. Those skilled in the art can make various other specific variations and combinations based on the technical teachings disclosed in the present invention without departing from the essence of the present invention, and such variations and combinations are still within the scope of protection of the present invention.

Claims

1. A cloud computing task scheduling method based on privacy protection, characterized in that: The following steps are involved: S1. Obtain the tasks to be scheduled and perform privacy processing on them. S2. After privacy processing, based on several containers on the cloud platform, an influence matrix is generated between the task to be scheduled and each container, and the scheduling unit of the task to be scheduled is determined according to the influence matrix; S3, completing the scheduling process according to the scheduling unit of the task to be scheduled; The S2 includes the following sub-steps: S21. After privacy processing, obtain several containers from the cloud platform; S22. Calculate the impact matrix between the scheduled task and each container based on the node risk factor of each container in the cloud platform; S23. Determine the scheduling unit of the task to be scheduled based on the influence matrix between the task to be scheduled and each container; The S22 includes the following sub-steps: S221. Convert the node risk factor of the container into a node risk vector using a hash vectorizer. S222. Convert the task priority of the task to be scheduled into a priority vector using a hash vectorized container; S223, obtaining the difference between the node risk vector and the priority vector; S224. Perform a dot product of the node risk vector and the priority vector; S225. Calculate the influence matrix between the task to be scheduled and the container using the activation function based on the difference result and the dot product result.

2. The privacy protection-based cloud computing task scheduling method according to claim 1, characterized in that: In S1, the cloud platform's firewall is used to set the IP address of the client corresponding to the task to be scheduled in the cloud platform's access whitelist to complete privacy processing.

3. The privacy protection-based cloud computing task scheduling method according to claim 1, characterized in that: In S225, the influence matrix between the task to be scheduled and the container The expression is: ； Where, represents the node risk vector, represents the priority vector, Represents a splicing operation, represents the dot product operation, represents the first trainable parameter matrix, represents the second trainable parameter matrix, represents the third trainable parameter matrix, represents the activation function, Represents a transpose operation.

4. The privacy protection-based cloud computing task scheduling method according to claim 1, characterized in that: The S23 includes the following sub-steps: S231. Calculate the row and column values of the influence matrix between the task to be scheduled and the container, and use the row and column values as the influence weights between the task to be scheduled and the container; S232, obtain the buffer overflow vulnerability of the scheduled task on the cloud platform; S233. Determine a final scheduling coefficient between the task to be scheduled and the container based on a buffer overflow vulnerability of the task to be scheduled on the cloud platform and an impact weight between the task to be scheduled and the container. S234: Use the container with the maximum final scheduling coefficient as the scheduling unit of the task to be scheduled.

5. The privacy protection-based cloud computing task scheduling method according to claim 4, characterized in that: In S233, the final scheduling coefficient between the task to be scheduled and the container The calculation formula is: ； Where, Indicates the influence weight between the task to be scheduled and the container. represents a constant, Indicates taking a random number between 0 and 1. Indicates the attack frequency of the scheduled task on the cloud platform. Indicates the memory size of the buffer overflow vulnerability of the scheduled task on the cloud platform.

Images