A method for dynamically allocating tenant-isolated resources, an electronic device, and a storage medium
By verifying tenant identities and scoring behavioral profiles, resource quotas are dynamically adjusted, solving the problem of resource abuse in multi-tenant environments and achieving fair and stable resource allocation among tenants.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HANGZHOU YSCREDIT CO LTD
- Filing Date
- 2026-05-21
- Publication Date
- 2026-07-10
AI Technical Summary
In a multi-tenant environment with shared database resources, existing technologies lack continuous monitoring and risk assessment of tenant behavior, making it impossible to effectively identify malicious tenants and resource abuse. Furthermore, static allocation strategies are difficult to dynamically adjust resource quotas based on actual tenant usage, affecting the fairness and stability of the system.
By verifying tenant identity information and allocating temporary isolation space, continuously collecting behavioral data to build dynamic behavioral profiles, calculating behavioral compliance scores, and dynamically adjusting resource quotas according to security levels, resource isolation and fair scheduling among tenants are achieved.
It achieves service quality assurance for high-trust tenants and limits resource abuse by low-trust tenants, thereby improving the rationality of resource allocation strategies and the stability of the system.
Smart Images
Figure CN122365544A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of database resource management technology, and in particular to a dynamic allocation method, electronic device and storage medium for tenant isolation resources. Background Technology
[0002] In a multi-tenant environment with shared database resources, different tenants share the underlying computing and storage resources. In existing technologies, tenants are typically allocated a fixed quota or temporary space after registration and can then directly use the database service. However, this mechanism lacks continuous monitoring and risk assessment of tenant behavior, and cannot effectively identify malicious tenants, resource abuse, or abnormal operating patterns, such as idle resource usage, excessively large queries, or frequent data exports. Furthermore, existing resource isolation schemes mostly employ static allocation strategies, making it difficult to dynamically adjust resource quotas based on actual tenant usage. This results in high-credit tenants not receiving guaranteed service quality, while low-credit tenants may excessively consume public resources, affecting the fairness and stability of the entire system.
[0003] Therefore, there is an urgent need for a method that can dynamically assess tenant security levels and flexibly allocate isolation resources. Summary of the Invention
[0004] To address the aforementioned technical problems, this invention provides a dynamic allocation method, electronic device, and storage medium for tenant isolation resources, which can ensure the service quality of high-trust tenants and limit the resource abuse of low-trust tenants, thereby achieving effective resource isolation and fair scheduling among tenants.
[0005] According to a first aspect of the present invention, a method for dynamically allocating tenant isolation resources is provided, comprising the following steps: S1, in response to each tenant's database registration request, verifies the tenant's identity information and allocates temporary isolation space and corresponding tenant identifier to tenants that pass the verification.
[0006] S2. For any tenant identifier, during the period when the corresponding tenant uses database resources, the tenant's behavioral data is continuously collected, and the influence weight of each behavioral dimension in the behavioral data on the consumption of database resources is calculated to construct a dynamic behavioral profile of the tenant.
[0007] S3 calculates the tenant's behavior compliance score by weighted summation based on the normalized values and corresponding impact weights of each behavior dimension in the tenant's dynamic behavior profile, and dynamically determines the tenant's current security level based on the behavior compliance score.
[0008] S4, when the isolation resource allocation trigger condition is detected, the isolation resource quota allocated to the tenant is dynamically adjusted based on the tenant's temporary isolation space and the tenant's current security level to achieve resource isolation between different tenants.
[0009] According to a second aspect of the present invention, a non-transitory computer-readable storage medium is provided, wherein at least one instruction or at least one program is stored therein, the at least one instruction or the at least one program being loaded and executed by a processor to implement the above-described method for dynamic allocation of tenant-isolated resources.
[0010] According to a third aspect of the present invention, an electronic device is provided, including a processor and the aforementioned non-transitory computer-readable storage medium.
[0011] The present invention has at least the following beneficial effects: This invention provides a dynamic allocation method for tenant isolation resources. Upon responding to each tenant's database registration request, it allocates temporary isolation space and a tenant identifier to verified tenants, establishing a benchmark for subsequent dynamic resource adjustments. During tenant usage of database resources, it continuously collects tenant behavior data and calculates the impact weight of each behavior dimension on database resource consumption. This objectively quantifies the degree of database resource consumption by different behaviors, avoiding biases from subjective weighting, and constructs a dynamic behavior profile reflecting the actual usage characteristics of tenants. This provides a reliable data foundation for subsequent accurate behavior compliance scoring and dynamic resource allocation. Then, based on the tenant's... The system calculates tenant behavior compliance scores by normalizing the values of each behavioral dimension of a user and their corresponding impact weights, thereby dynamically determining the current security level of the tenant. This enables the system to automatically identify high-risk tenants and trusted tenants, achieving differentiated access control and improving the rationality of resource allocation strategies. When the conditions for triggering isolation resource allocation are detected, the system dynamically adjusts the tenant's isolation resource quota based on the tenant's temporary isolation space and the tenant's current security level. By dynamically adjusting the tenant's isolation resource quota through the tenant's security level, the system can ensure the service quality of high-trust tenants and limit the resource abuse of low-trust tenants, thereby achieving effective resource isolation and fair scheduling among tenants. Attached Figure Description
[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0013] Figure 1 A flowchart illustrating a dynamic allocation method for tenant isolation resources provided in an embodiment of the present invention. Detailed Implementation
[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0015] This invention provides a method for dynamically allocating tenant-isolated resources, such as... Figure 1 As shown, the method includes the following steps: S1, responding to each tenant's database registration request, verifies the tenant's identity information and allocates temporary isolation space and a corresponding tenant identifier to tenants who pass the verification. This can be understood as: temporary isolation space refers to a separate isolation container allocated to each tenant within the database, allowing users to perform data transmission, storage, and usage operations within that container. In practical implementations, it can be used for relational databases, real-time analytics databases, and other types of databases.
[0016] Specifically, verifying tenant identity information includes, but is not limited to, screening for the authenticity of identity, the validity of contact information, the reasonableness of the usage scenario, and basic risks. For example, the reasonableness of the usage scenario includes whether the claimed business type and purpose are reasonable; screening for basic risks refers to whether the tenant is on a blacklist and whether there are any records of historical malicious behavior.
[0017] The above-mentioned verification of basic identity information can prevent malicious or invalid tenants from occupying database resources, ensure that only legitimate tenants can enter the subsequent dynamic management process, allocate temporary isolation space, provide tenants with the initial minimum available resources, avoid tenants being unable to use the resources immediately after registration, and establish a benchmark for subsequent dynamic adjustment of resources.
[0018] S2. For any tenant identifier, during the period when the corresponding tenant uses database resources, the tenant's behavioral data is continuously collected, and the influence weight of each behavioral dimension in the behavioral data on the consumption of database resources is calculated to construct a dynamic behavioral profile of the tenant.
[0019] Specifically, each behavioral dimension includes at least usage cycle, usage frequency, operation type, data volume per operation, resource consumption, and behavioral pattern; data collection for each behavioral dimension is performed in units of sliding time windows. For example, for the above behavioral dimensions, the corresponding abnormal operation behaviors include: high-frequency operation during abnormal periods, large-scale export of sensitive data, continuous large-scale data query, long-term high resource consumption but low output, idle state, or abnormal concurrent operation, etc.
[0020] Furthermore, the weighting of the impact of each behavioral dimension on database resource consumption in the calculated behavioral data includes the following steps: S201: At each preset time point, collect the historical behavior data sequence of each behavioral dimension and the consumption data sequence of each resource type for the tenant, and perform dimensionless processing on the historical behavior data sequence and the consumption data sequence of each resource type; this can be understood as: setting several preset time points in the sliding time window.
[0021] Specifically, the database resources include several resource types, such as CPU, memory, and GPU.
[0022] Furthermore, when performing dimensionless processing on historical behavior data sequences and consumption data sequences, the min-max normalization method can be used to normalize each data value to between 0 and 1.
[0023] Preferably, step S201 further includes the following steps: S2011, for each behavioral dimension and each resource type, obtain the original observation sequence at each preset time point; this can be understood as: obtaining the original behavioral data sequence corresponding to each behavioral dimension and the original consumption data sequence corresponding to each resource type.
[0024] S2012 uses a preset triple exponential smoothing method to smooth each original observation value sequence, and uses the smoothed sequence value as the corresponding historical behavior data sequence or the corresponding resource type consumption data sequence.
[0025] Specifically, the process of smoothing each original observation sequence using a preset triple exponential smoothing method includes the following steps: S2012a, for any original observation sequence, obtain the initial horizontal component S. L Initial trend component b L and the initial periodic component c i .
[0026] Specifically, initial level components Where L is the preset period length, x i This refers to the original observation value at the i-th preset time point. That is, the mean of the original observation values within the first collection period is used as the initial level component; for example, the collection period for behavioral data can be set to n days. When collecting behavioral data by hour, the period length L is 24, that is, the first period is from the 1st hour to the 24th hour.
[0027] Furthermore, the initial trend component b L =0.
[0028] Furthermore, the initial periodic component c i =x i -S L .
[0029] S2012b, for time points t=L+1, L+2, ..., calculate the current horizontal component S at the current time point according to the following formula. t Current trend component b t and the current periodic component c t .
[0030] Specifically, S t =α(x t -c t-L )+(1-α)(S t-1 +b t-1 ), where S t x is the smoothed value corresponding to the t-th preset time point. t S represents the original observation value at the t-th preset time point. t-1 b is the smoothed value corresponding to the (t-1)th preset time point. t-1 c represents the trend component at the (t-1)th preset time point. t-L This refers to the periodic component at the same time point as the previous acquisition cycle.
[0031] b t =β(S t -S t-L )+(1-β)b t-1 , where b t S represents the trend component at the t-th preset time point. t-L This refers to the horizontal component at the same time point in the previous acquisition cycle. That is, b t It is used to reflect the direction and rate of change of the data sequence near that point in time.
[0032] c t =γ(x t -S t-L -b t-1 )+(1-γ)c t-L .
[0033] Wherein, α, β and γ are preset smoothing coefficients, with values ranging from 0 to 1.
[0034] The aforementioned triple exponential smoothing method can eliminate noise by filtering out random fluctuations through recursive smoothing of the horizontal component. Simultaneously, it uses trend components to track the direction of behavioral changes and periodic components to characterize the regular fluctuations of day and night or weekdays, preserving the trend and periodic characteristics of the data sequence. This provides high-quality input data for subsequent correlation coefficient calculations, enabling the calculation of the influence weights of the behavioral dimension to distinguish between normal periodic peaks and abnormal sudden behaviors. This avoids misjudging regular fluctuations as risks, reduces the interference of instantaneous abnormal behaviors on behavioral compliance scoring, and thus improves the accuracy of subsequent behavioral compliance scoring and the robustness of safety level determination. Ultimately, this achieves a more reasonable and fair dynamic allocation of tenant isolation resources.
[0035] S202, for any resource type's consumption data sequence, calculate the absolute difference between the historical behavior data sequence and the consumption data sequence for each behavior dimension at each preset time point. Then, calculate the correlation coefficient between each historical behavior data sequence and the current resource type at each preset time point based on the global maximum and minimum differences. This can be understood as follows: the global maximum difference refers to the maximum value among all absolute differences, and the global minimum difference refers to the minimum value among all absolute differences. A larger correlation coefficient indicates a more synchronized trend between the historical behavior data sequence and the consumption data sequence, thus increasing the weight of the influence on the current resource type.
[0036] Specifically, the correlation coefficient is expressed as follows: ξ jk =(D min +ρ×D max ) / (D jk +ρ×D max ), where ξ jk D represents the correlation coefficient between the j-th historical behavior data sequence at the k-th preset time point and the current resource type. min D is the global minimum difference. max D is the global maximum difference. jk ρ is the absolute difference between the j-th historical behavior data sequence and the current resource type consumption data sequence at the k-th preset time point, and ρ is the resolution coefficient, which ranges from 0 to 1.
[0037] S203, for each historical behavior data sequence, take the average of the correlation coefficients between the historical behavior data sequence and the current resource type at all preset time points to obtain the correlation degree between the behavior dimension corresponding to the historical behavior data sequence and the current resource type.
[0038] S204 normalizes the correlation between each behavioral dimension and the current resource type to obtain the weight of each behavioral dimension's influence on the consumption of the current resource type. For example, the normalized value corresponding to each correlation degree is the sum of each correlation degree.
[0039] S205, based on the weighted impact of each behavioral dimension on the consumption of each resource type, and the preset weight of each resource type, the weighted sum is used to calculate the weighted impact of each behavioral dimension on the database resource consumption. Those skilled in the art can set the preset weight for each resource type according to actual needs.
[0040] As described above, by collecting tenant behavior data and calculating the impact weight of each behavior dimension on resource consumption, the degree of consumption of different behaviors on different resource types can be objectively quantified, avoiding the bias of subjective weighting. The weighted calculation of the contribution weight of each behavior dimension to the overall database resource consumption is obtained, and a dynamic behavior profile reflecting the actual usage characteristics of tenants is constructed, providing a reliable data foundation for subsequent accurate behavior compliance scoring calculation and dynamic resource allocation.
[0041] S3 calculates the tenant's behavior compliance score by weighted summation based on the normalized values and corresponding impact weights of each behavior dimension in the tenant's dynamic behavior profile, and dynamically determines the tenant's current security level based on the behavior compliance score.
[0042] It should be noted that the impact weight reflects the actual impact of each behavioral dimension on the consumption of database resources. That is, the greater the impact of a behavioral dimension, the more critical it is. The normalized value of a behavioral dimension reflects the quality of its performance, such as whether the frequency is too high or the data volume is appropriate. By weighting and calculating, the higher the behavioral compliance score, the more standardized the tenant's performance is on the key behavioral dimensions, and the higher the current security level is determined.
[0043] Specifically, the process of dynamically determining a tenant's current security level based on behavioral compliance scoring includes the following steps: S301, Obtain the preset security level classification rules; the security level classification rules include several consecutive scoring intervals, each scoring interval corresponding to a security level; those skilled in the art can set the specific numerical range of each scoring interval according to actual needs, which will not be elaborated here.
[0044] Specifically, the security levels include at least high, medium and low levels.
[0045] S302, match the behavior compliance score with each score interval to determine the score interval corresponding to the behavior compliance score and the corresponding current security level, and associate and store the determined current security level with the corresponding tenant identifier.
[0046] The above-mentioned approach, by calculating the behavior compliance score through weighted summation and determining the security level, abstracts the multidimensional behavior of tenants into a single quantifiable behavior compliance score indicator. This enables the system to automatically identify high-risk tenants and trusted tenants, achieve differentiated access control, avoid a one-size-fits-all resource allocation strategy, and improve the rationality of resource allocation strategies and database stability.
[0047] S4, when the isolation resource allocation trigger condition is detected, the isolation resource quota allocated to the tenant is dynamically adjusted based on the tenant's temporary isolation space and the tenant's current security level to achieve resource isolation between different tenants; it can be understood as: on the basis of sharing the underlying public data resources, the database resources are allocated in isolation.
[0048] Specifically, the triggering condition refers to being triggered when a user issues a query or at a fixed time period.
[0049] Specifically, step S4 includes the following steps: S401, determine the resource adjustment coefficient λ corresponding to the tenant based on the tenant's current security level; where high level corresponds to λ > 1, medium level corresponds to λ = 1, and low level corresponds to 0 < λ < 1. That is, the resource usage permission corresponding to high level is to allow the resource quota to be expanded upward, the resource usage permission corresponding to medium level is to maintain the current resource quota unchanged, and the resource usage permission corresponding to low level is to limit the resource quota or reduce the scheduling priority according to demand.
[0050] S402, the product of the resource limit corresponding to the tenant's temporary isolation space and the resource adjustment coefficient λ corresponding to the tenant is used as the tenant's target isolation resource quota.
[0051] S403 employs a preset resource isolation mechanism to adjust isolation resource quotas, thereby adjusting the resource limit corresponding to the tenant's temporary isolation space to the target isolation resource quota. In specific implementations, the Control Groups mechanism or namespace mechanism provided by the Linux kernel can be used to achieve resource isolation between tenants and prevent cross-tenant interference.
[0052] Based on the temporary isolation space, the isolation resource quota is dynamically adjusted according to the tenant's security level. This enables the dynamic allocation of isolation resources to tenants with different compliance levels while sharing underlying resources. This not only ensures the performance experience of high-trust tenants but also restricts the abuse of resources by low-trust tenants, effectively achieving resource isolation and fair scheduling among tenants.
[0053] Furthermore, the method also includes a dynamic resource adjustment method based on behavioral feedback, specifically comprising the following steps: S10: Each time the current security level of a tenant is determined, the current security level is compared with the security level of the tenant determined in the previous step.
[0054] S20: If the current security level is higher than the previous security level, the resource adjustment coefficient λ corresponding to the tenant is increased according to the preset adjustment step size; if the current security level is lower than the previous security level, the resource adjustment coefficient λ corresponding to the tenant is decreased to the target value corresponding to the current security level, and an audit log recording is triggered; if the current security level is the same as the previous security level, the resource adjustment coefficient λ corresponding to the tenant remains unchanged. Those skilled in the art can set the preset adjustment step size and the target value corresponding to each security level according to actual needs, which will not be elaborated here.
[0055] It should be noted that if the current resource adjustment coefficient λ is no greater than the target value corresponding to the current security level, then there is no need to lower it further; this can be understood as: the target value corresponding to the lower level is the lower limit of the resource adjustment coefficient λ.
[0056] S30 If a tenant is found to have pre-defined high-risk behavior, the current security level assessment will be skipped, the security level will be directly downgraded to a low level, and the resource adjustment coefficient λ≤1 will be maintained for at least one pre-defined penalty period. This can be understood as: within a pre-defined penalty period, even if the tenant's security level is improved, the resource adjustment coefficient for the tenant will not be increased.
[0057] In another scenario, if a tenant's security level remains high for multiple consecutive times, the basic resource quota for the temporary isolation space will be increased to the current resource limit as the new benchmark.
[0058] The above-mentioned mechanism, through dynamic comparison of security levels and two-way adjustment of resource adjustment coefficients, enables refined management of tenant permissions. Resource usage permissions are gradually expanded when the security level is increased and rapidly reduced when the security level is decreased. Simultaneously, a circuit breaker mechanism for high-risk behaviors is established. Once malicious operations are detected, the normal scoring process is immediately skipped, resulting in a direct downgrade and halting of permission expansion, effectively ensuring the overall security and fair allocation of database resources.
[0059] In a preferred embodiment, the method further includes an iterative update step for the profile and weights: S01, with a preset period as the time interval, extract all tenant behavior data and consumption data of each resource type accumulated within the current preset period to obtain new historical behavior data sequences for each behavior dimension and new consumption data sequences for each resource type.
[0060] S02, re-execute steps S201-S204 to update the weight matrix of the impact of each behavior dimension on the consumption of each resource type.
[0061] S03, compare the updated consumption impact weight matrix with the original consumption impact weight matrix. If the rate of change of weight for any behavioral dimension exceeds a preset rate of change threshold, a weight change alarm is triggered. Those skilled in the art can set the preset rate of change threshold according to actual needs, which will not be elaborated here.
[0062] As described above, by periodically performing impact weight calculations, the long-term evolution trend of tenant group behavior patterns can be captured, ensuring that the weight matrix always remains consistent with the latest data distribution characteristics. Furthermore, a weight change rate monitoring and alarm mechanism has been introduced, which makes it easier for operations and maintenance personnel to identify abnormal weight drifts or sudden changes in data distribution, ensuring the reliability and interpretability of compliance scoring.
[0063] Embodiments of the present invention also provide a non-transitory computer-readable storage medium, which can be disposed in an electronic device to store at least one instruction or at least one program related to implementing a method in the method embodiments, wherein the at least one instruction or the at least one program is loaded and executed by the processor to implement the dynamic allocation method for tenant isolation resources provided in the above embodiments.
[0064] Embodiments of the present invention also provide an electronic device, including a processor and the aforementioned non-transitory computer-readable storage medium.
[0065] While specific embodiments of the invention have been described in detail by way of example, those skilled in the art should understand that the examples are for illustrative purposes only and not intended to limit the scope of the invention. It should also be understood that various modifications can be made to the embodiments without departing from the scope and spirit of the invention. The scope of the invention is defined by the appended claims.
Claims
1. A method for dynamically allocating tenant-isolated resources, characterized in that, The method includes the following steps: S1, in response to each tenant's database registration request, verifies the tenant's identity information and allocates temporary isolation space and corresponding tenant identifier to tenants that pass the verification; S2, for any tenant identifier, during the period when the corresponding tenant uses database resources, continuously collect the tenant's behavior data and calculate the impact weight of each behavior dimension in the behavior data on the consumption of database resources, so as to build a dynamic behavior profile of the tenant. S3 calculates the tenant’s behavior compliance score by weighted summation based on the normalized values and corresponding impact weights of each behavior dimension in the tenant’s dynamic behavior profile, and dynamically determines the tenant’s current security level based on the behavior compliance score. S4, when the isolation resource allocation trigger condition is detected, the isolation resource quota allocated to the tenant is dynamically adjusted based on the tenant's temporary isolation space and the tenant's current security level to achieve resource isolation between different tenants.
2. The dynamic allocation method for tenant isolation resources according to claim 1, characterized in that, Each behavioral dimension includes at least the usage cycle, usage frequency, operation type, data volume per operation, resource consumption, and behavior pattern; data collection for each behavioral dimension is performed in units of sliding time windows.
3. The dynamic allocation method for tenant isolation resources according to claim 1, characterized in that, In step S2, the weighting of the impact of each behavioral dimension on database resource consumption in the calculated behavioral data includes the following steps: S201, at each preset time point, collect historical behavior data sequences for each behavioral dimension and consumption data sequences for each resource type of the tenant, and perform dimensionless processing on the historical behavior data sequences and consumption data sequences for each resource type; the database resources include several resource types; S202, for any resource type consumption data sequence, calculate the absolute difference between the historical behavior data sequence and the consumption data sequence of each behavior dimension at each preset time point, and calculate the correlation coefficient between each historical behavior data sequence and the current resource type at each preset time point based on the global maximum difference and the global minimum difference; S203, For each historical behavior data sequence, take the average of the correlation coefficients between the historical behavior data sequence and the current resource type at all preset time points to obtain the correlation degree between the behavior dimension corresponding to the historical behavior data sequence and the current resource type; S204, normalize the correlation between each behavior dimension and the current resource type to obtain the weight of the influence of each behavior dimension on the consumption of the current resource type; S205. Based on the weight of each behavioral dimension's impact on the consumption of each resource type, and the preset weight of each resource type, the weight of each behavioral dimension's impact on the database resource consumption is calculated by weighted summation.
4. The dynamic allocation method for tenant isolation resources according to claim 3, characterized in that, Step S201 also includes the following steps: S2011, for each behavioral dimension and each resource type, obtain the original observation sequence at each preset time point; S2012 uses a preset triple exponential smoothing method to smooth each original observation value sequence, and uses the smoothed sequence value as the corresponding historical behavior data sequence or the corresponding resource type consumption data sequence.
5. The dynamic allocation method for tenant isolation resources according to claim 1, characterized in that, In step S3, the dynamic determination of the tenant's current security level based on behavioral compliance scoring includes the following steps: S301, Obtain a preset security level classification rule; the security level classification rule includes several consecutive scoring intervals, each scoring interval corresponding to a security level; the security level includes at least high level, medium level and low level; S302, match the behavior compliance score with each score interval to determine the score interval corresponding to the behavior compliance score and the corresponding current security level, and associate and store the determined current security level with the corresponding tenant identifier.
6. The dynamic allocation method for tenant isolation resources according to claim 5, characterized in that, Step S4 includes the following steps: S401, determine the resource adjustment coefficient λ corresponding to the tenant based on the tenant's current security level; where λ > 1 corresponds to high level, λ = 1 corresponds to medium level, and 0 < λ < 1 corresponds to low level. S402, the product of the resource limit corresponding to the tenant's temporary isolation space and the resource adjustment coefficient λ corresponding to the tenant is used as the tenant's target isolation resource quota; S403 uses a preset resource isolation mechanism to adjust the isolation resource quota, so as to adjust the resource limit corresponding to the tenant's temporary isolation space to the target isolation resource quota.
7. The dynamic allocation method for tenant isolation resources according to claim 6, characterized in that, The method further includes the following steps: S10, each time the current security level of a tenant is determined, the current security level is compared with the security level of the tenant determined in the previous step; S20, if the current security level is higher than the previous security level, the resource adjustment coefficient λ corresponding to the tenant is increased according to the preset upward adjustment step size; if the current security level is lower than the previous security level, the resource adjustment coefficient λ corresponding to the tenant is reduced to the target value corresponding to the current security level, and the audit log is triggered; if the current security level is the same as the previous security level, the resource adjustment coefficient λ corresponding to the tenant remains unchanged. S30. If a tenant is found to have pre-defined high-risk behavior, the current security level assessment will be skipped, the security level will be directly downgraded to a low level, and the resource adjustment coefficient λ≤1 will be maintained for at least one pre-defined penalty period.
8. The dynamic allocation method for tenant isolation resources according to claim 3, characterized in that, The method also includes a profile and weight iterative update step: S01, with a preset period as the time interval, extract all tenant behavior data and consumption data of each resource type accumulated within the current preset period to obtain new historical behavior data sequences of each behavior dimension and new consumption data sequences of each resource type; S02, re-execute steps S201-S204 to update the weight matrix of the impact of each behavior dimension on the consumption of each resource type; S03, compare the updated consumption impact weight matrix with the previous consumption impact weight matrix. If the weight change rate of any behavior dimension exceeds the preset change rate threshold, a weight change alarm will be triggered.
9. A non-transitory computer-readable storage medium, wherein the storage medium stores at least one instruction or at least one program segment, characterized in that, The at least one instruction or the at least one program segment is loaded and executed by the processor to implement the dynamic allocation method for tenant isolation resources as described in any one of claims 1-8.
10. An electronic device, characterized in that, Includes a processor and the non-transitory computer-readable storage medium as described in claim 9.