A virtualization and container-based resource pool security management method

Abstract

The application relates to the technical field of resource pool management, in particular to a resource pool security management method based on virtualization and containerization. The method comprises the following steps: constructing a virtual resource pool based on resource equipment parameters; setting a plurality of running nodes in the virtual resource pool according to historical task parameters, and generating a first resource allocation strategy according to all the running nodes; setting a task sub-container according to a to-be-executed task, and selecting a to-be-executed node according to the characteristic parameters of the task sub-container; obtaining feedback data packets of each task sub-container according to a preset feedback time node, and judging whether to generate a correction instruction according to all the feedback data packets. The virtual resource pool is constructed by virtualizing all the hardware equipment, a plurality of running nodes are established, different types of to-be-executed tasks are adapted, various running resources in the virtual resource pool are dynamically allocated according to the characteristic parameters of each running node, and the scheduling and allocation efficiency of the running resources is improved.

Description

Technical Field

[0001] The present application relates to the technical field of resource pool management, and in particular to a resource pool security management method based on virtualization and containerization. Background Art

[0002] A resource pool refers to the abstraction, integration, and centralized management of physical or virtual computing resources (such as CPU, memory, storage), network resources (such as bandwidth, IP addresses, ports), or other specific functional resources (such as industrial equipment, sensor access points, licenses), forming a collection of shared resources that can be dynamically allocated and recycled on demand.

[0003] With the rapid development of new services such as cloud computing, big data, and the Internet of Things, data centers are increasingly demanding operational resources. However, traditional resource pool management methods are inefficient and unable to meet these elastic demands. Summary of the Invention

[0004] The purpose of this application is: to solve the above technical problems, this application provides a resource pool security management method based on virtualization and containerization, aiming to improve the operation scheduling efficiency of the resource pool and improve the utilization rate of operating resources.

[0005] In some embodiments of the present application, a resource pool security management method based on virtualization and containerization is provided, including:

[0006] Build a virtual resource pool based on resource device parameters;

[0007] Set up multiple running nodes in the virtual resource pool based on historical task parameters, and generate a first-level resource allocation strategy based on all running nodes;

[0008] Set the task subcontainer according to the task to be executed, and select the node to be executed according to the characteristic parameters of the task subcontainer;

[0009] The feedback data packets of each task sub-container are obtained according to the preset feedback time node, and it is determined whether to generate a correction instruction based on all the feedback data packets.

[0010] In some embodiments of the present application, when multiple running nodes are set, the following steps are included:

[0011] Set multiple task characteristic indicators based on historical task parameters;

[0012] Generate multiple task categories based on all task characteristic indicators and establish a task category sequence P, P=(p1, p2…p i …p m ), where p i is the i-th task category; m is the number of task categories;

[0013] According to the task category sequence P, set p in sequence i is the target task category;

[0014] Generate the demand evaluation value s of the target task category;

[0015] s= β i *v i ;

[0016] Among them, θ1 is the number of demand evaluation indicators; β i is the influencing factor of the i-th demand evaluation indicator; vi is the reference value of the i-th demand evaluation indicator in the target task category;

[0017] Set the number of running nodes of the target task category according to the demand evaluation value v;

[0018] Set the number of running nodes for each task category in turn;

[0019] Establish a running node sequence A, A=(a1, a2…a i …a n ), where a i is the i-th running node; n is the number of running nodes, and n>m;

[0020] A task category-running node mapping table is set according to the running node number sequence A and the task category number sequence P.

[0021] In some embodiments of the present application, when generating a primary resource allocation strategy based on all running nodes, the process includes:

[0022] Set a in sequence according to the number of running nodes A i is the target node;

[0023] Obtain the feature data packet of the task category corresponding to the target node according to the task category-running node mapping table;

[0024] Set the target node's operating architecture based on the characteristic data packet;

[0025] Generate resource sub-strategy for the target node;

[0026] Generate the operation architecture and resource sub-strategy of each operation node in turn;

[0027] Generate a first-level resource allocation strategy based on all resource sub-strategies.

[0028] In some embodiments of the present application, selecting a node to be executed based on characteristic parameters of a task subcontainer includes:

[0029] According to the task category sequence P, set pi as the task category to be compared;

[0030] Generate a similarity evaluation value c between the task subcontainer and the task category to be compared;

[0031] Generate similar evaluation values ​​of task sub-containers and each task category in sequence, and establish similar evaluation value sequence C, C = (c1, c2…c i …c m ), where c i is the similarity evaluation value between the task subcontainer and the i-th task category;

[0032] Set the maximum value c in the similarity evaluation value sequence C max The corresponding task category is the first-level task category;

[0033] According to the task category-running node mapping table, set the running node corresponding to the first-level task category as the first-level node;

[0034] Generate load evaluation values ​​for each first-level node;

[0035] Establish a load evaluation value series D, D = (d1, d2…d i …d n2 ), where d i is the load evaluation value of the i-th first-level node; n2 is the number of first-level nodes;

[0036] Set the minimum value d in the load evaluation value sequence D min The corresponding first-level node is the node to be executed.

[0037] In some embodiments of the present application, generating a similarity evaluation value c between a task subcontainer and a task category to be compared includes:

[0038] c=e1*Q1*[ η 1i *w i ]+e2*Q2*[ η 2i *t i ];

[0039] Wherein, e1 is the preset first weight coefficient; e2 is the preset second weight coefficient; Q1 is the preset first fixed coefficient; Q2 is the preset second fixed coefficient; θ2 is the number of resource characteristic indicators; η 1i is the influencing factor of the i-th resource characteristic indicator; w i is the similarity between the resource feature index of the task subcontainer and the task category to be compared; θ3 is the number of historical evaluation indicators; η 2i is the impact factor of the i-th historical evaluation index; t i is the reference value of the i-th historical evaluation indicator.

[0040] In some embodiments of the present application, establishing the load evaluation value sequence D includes:

[0041] Select the target first-level node according to all first-level nodes in sequence;

[0042] Generate the load evaluation value d of the target first-level node;

[0043] d= α i *j i ;

[0044] Among them, θ4 is the number of load evaluation indicators; α i is the influencing factor of the i-th load evaluation index; j i is the reference value of the i-th load evaluation index in the target first-level node.

[0045] In some embodiments of the present application, determining whether to generate a correction instruction based on all feedback data packets includes:

[0046] Set a in sequence according to the number of running nodes A i is the node to be monitored;

[0047] Set all task sub-containers in the sub-node to be monitored as running sub-containers;

[0048] Set the running sub-container sequence Y of the node to be monitored, Y=(y1, y2…y i …y r ), where y i is the i-th running sub-container of the node to be monitored; r is the number of running sub-containers of the node to be monitored;

[0049] Set y in sequence according to the number of running sub-containers Y i Run a subcontainer for the target;

[0050] Get the feedback data packet of the target running sub-container at the current feedback time node;

[0051] Generate the running evaluation value f of the target running sub-container at the current feedback time node;

[0052] Preset operation evaluation value threshold F1;

[0053] If f <F1,当前反馈时间节点生成目标运行子容器的一级修正指令;

[0054] It is determined in turn whether each running sub-container generates a first-level correction instruction at the current feedback time node.

[0055] In some embodiments of the present application, generating the operation evaluation value f includes:

[0056] f= g i *u i ;

[0057] Among them, θ5 is the number of preset operation evaluation indicators; g i is the influencing factor of the i-th operation evaluation index; u i The reference value of the i-th running evaluation indicator of the target running sub-container at the current feedback time node.

[0058] In some embodiments of the present application, when determining whether to generate a correction instruction based on all feedback data packets, the following steps are further included:

[0059] Generate the operation evaluation value sequence F of the node to be monitored at the current feedback time node, F=(f1, f2…f i …f r ), where f i is the running evaluation value of the i-th running sub-container of the node to be monitored; r is the number of running sub-containers of the node to be monitored;

[0060] Generate the performance deviation value h of the node to be monitored based on the running evaluation value series F;

[0061] Set the performance deviation threshold H1;

[0062] If h>H1, generate a secondary correction instruction.

[0063] In some embodiments of the present application, generating the performance deviation value u includes:

[0064] h=e3*Q3*[ λ i *k i ]+e4*Q4*(d'-d'1);

[0065] Among them, e3 is the preset third weight coefficient; e4 is the preset fourth weight coefficient; Q3 is the preset third fixed coefficient; Q4 is the preset fourth fixed coefficient; θ6 is the number of auxiliary evaluation indicators; λ i is the influencing factor of the i-th auxiliary evaluation index; k i is the reference value of the i-th auxiliary evaluation index generated based on the running evaluation value series; d' is the load evaluation value of the node to be monitored at the current feedback time node; d'1 is the expected load value of the node to be monitored at the current feedback time node.

[0066] Compared with the prior art, the resource pool security management method based on virtualization and containerization in the embodiment of the present application has the following advantages:

[0067] By virtualizing all hardware devices to build a virtual resource pool, and by establishing multiple operating nodes to adapt to different types of tasks to be executed, various types of operating resources in the virtual resource pool are dynamically allocated according to the characteristic parameters of each operating node, thereby improving the scheduling and allocation efficiency of operating resources.

[0068] By containerizing the tasks to be executed and selecting the corresponding running nodes, the running safety of the tasks to be executed is guaranteed and the running efficiency of each task to be executed is improved. By periodically collecting the resource consumption data of each task sub-container, the resource allocation strategy is dynamically corrected in time to ensure the running performance of the task sub-container. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a flow chart of a resource pool security management method based on virtualization and containerization in the preferred embodiment of the present application. DETAILED DESCRIPTION

[0070] The following embodiments are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0071] In the description of this application, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application.

[0072] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature specified as "first" or "second" may explicitly or implicitly include one or more of such features. Throughout this application, unless otherwise specified, "plurality" means two or more.

[0073] In the description of this application, it should be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in this application based on the specific circumstances.

[0074] like Figure 1As shown, a resource pool security management method based on virtualization and containerization in a preferred embodiment of the present application includes:

[0075] S101: Constructing a virtual resource pool based on resource device parameters;

[0076] S102: Setting multiple running nodes in the virtual resource pool according to historical task parameters, and generating a primary resource allocation strategy based on all running nodes;

[0077] S103: Setting a task subcontainer according to the task to be executed, and selecting a node to be executed according to characteristic parameters of the task subcontainer;

[0078] S104: Acquire feedback data packets of each task sub-container according to a preset feedback time node, and determine whether to generate a correction instruction based on all feedback data packets.

[0079] Specifically, resource devices include physical resource devices such as CPU, memory, storage, etc. and virtual resources such as network, etc. A virtual resource pool is constructed by abstracting and virtualizing all resource devices.

[0080] Specifically, when setting up multiple running nodes, include:

[0081] Set multiple task characteristic indicators based on historical task parameters;

[0082] Generate multiple task categories based on all task characteristic indicators and establish a task category sequence P, P=(p1, p2…p i …p m ), where p i is the i-th task category; m is the number of task categories;

[0083] According to the task category sequence P, set p in sequence i is the target task category;

[0084] Generate the demand evaluation value s of the target task category;

[0085] s= β i *v i ;

[0086] Among them, θ1 is the number of demand evaluation indicators; β i is the influencing factor of the i-th demand evaluation indicator; vi is the reference value of the i-th demand evaluation indicator in the target task category;

[0087] Set the number of running nodes of the target task category according to the demand evaluation value v;

[0088] Set the number of running nodes for each task category in turn;

[0089] Establish a running node sequence A, A=(a1, a2…a i …a n ), where a i is the i-th running node; n is the number of running nodes, and n>m;

[0090] A task category-running node mapping table is set according to the running node number sequence A and the task category number sequence P.

[0091] Specifically, the larger the demand evaluation value, the larger the corresponding number of running nodes. The mapping relationship between the demand evaluation value and the number of running nodes can be set according to historical parameters.

[0092] Specifically, multiple sub-resource pools are set in the virtual resource pool based on all resource parameters of the virtual resource pool, and a single sub-resource pool is set as a single running node.

[0093] Specifically, historical task parameters refer to all user-required tasks that the system has executed. By analyzing and processing historical task parameters, multiple task characteristic indicators are generated.

[0094] Specifically, task characteristic indicators include, but are not limited to, business type, operational load, and the maximum demand for various operational resources. By quantifying each task characteristic indicator, multiple value ranges are generated for each task characteristic indicator. By randomly combining these value ranges, various task categories are constructed.

[0095] Specifically, a historical data package for the target task category is generated based on historical task parameters. This data package includes all previously executed user-required tasks belonging to the target task category. Reference values ​​for each demand evaluation indicator are generated based on this historical data package. A higher reference value for each demand evaluation indicator indicates a greater amount of resources required by the target task category.

[0096] Specifically, demand evaluation indicators include but are not limited to the average task execution time, total task volume, task execution frequency and other parameters related to resource demand. The value of the influencing factor of each demand evaluation indicator is set according to its correlation with resource demand. The greater the correlation, the larger the corresponding impact factor value. The specific value range can be set according to historical parameters.

[0097] It is understandable that in the above embodiment, multiple task categories are set based on the analysis of historical task parameters, and the demand evaluation value of each task category dynamically adjusts the corresponding number of running nodes to improve the running efficiency of all tasks to be executed.

[0098] In a preferred embodiment of the present application, when generating a first-level resource allocation strategy based on all running nodes, it includes:

[0099] Set a in sequence according to the number of running nodes A i is the target node;

[0100] Obtain the feature data packet of the task category corresponding to the target node according to the task category-running node mapping table;

[0101] Set the target node's operating architecture based on the characteristic data packet;

[0102] Generate resource sub-strategy for the target node;

[0103] Generate the operation architecture and resource sub-strategy of each operation node in turn;

[0104] Generate a first-level resource allocation strategy based on all resource sub-strategies.

[0105] Specifically, the feature data package includes the expected demand for various types of operating resources and the required operating architecture of the task category. The operating architecture corresponding to the task node is set according to the feature data package, thereby improving the task execution efficiency of the target node.

[0106] Specifically, different task types may not necessarily require the same amount of resources. For example, for tasks with low load but requiring timely data exchange, their CPU usage can be appropriately reduced while network resource usage can be increased.

[0107] Specifically, the expected demand for each resource type for a task category is set based on the actual consumption of each resource in its historical data package. The higher the consumption, the greater the expected demand. A target node resource sub-policy is generated based on the total expected demand, and all resources in the virtual resource pool are scheduled according to the total resource sub-policy.

[0108] Specifically, when selecting a node to be executed based on the characteristic parameters of the task subcontainer, it includes:

[0109] According to the task category sequence P, set pi as the task category to be compared;

[0110] Generate a similarity evaluation value c between the task subcontainer and the task category to be compared;

[0111] Generate similar evaluation values ​​of task sub-containers and each task category in sequence, and establish similar evaluation value sequence C, C = (c1, c2…c i …c m ), where c i is the similarity evaluation value between the task subcontainer and the i-th task category;

[0112] Set the maximum value c in the similarity evaluation value sequence C max The corresponding task category is the first-level task category;

[0113] According to the task category-running node mapping table, set the running node corresponding to the first-level task category as the first-level node;

[0114] Generate load evaluation values ​​for each first-level node;

[0115] Establish a load evaluation value series D, D = (d1, d2…d i …d n2 ), where d i is the load evaluation value of the i-th first-level node; n2 is the number of first-level nodes;

[0116] Set the minimum value d in the load evaluation value sequence D min The corresponding first-level node is the node to be executed.

[0117] Specifically, the larger the similarity evaluation value is, the more suitable the running node corresponding to the task category to be compared is for running the current task subcontainer. By analyzing all runs corresponding to the task category to be compared, the task subcontainer is placed in the most idle running node for running.

[0118] Specifically, task sub-containers can share all running resources of the running node.

[0119] Specifically, when generating the similarity evaluation value c between the task subcontainer and the task category to be compared, it includes:

[0120] c=e1*Q1*[ η 1i *w i ]+e2*Q2*[ η 2i *t i ];

[0121] Wherein, e1 is the preset first weight coefficient; e2 is the preset second weight coefficient; Q1 is the preset first fixed coefficient; Q2 is the preset second fixed coefficient; θ2 is the number of resource characteristic indicators; η 1i is the influencing factor of the i-th resource characteristic indicator; w i is the similarity between the resource feature index of the task subcontainer and the task category to be compared; θ3 is the number of historical evaluation indicators; η 2i is the impact factor of the i-th historical evaluation index; t i is the reference value of the i-th historical evaluation indicator.

[0122] Specifically, resource characteristic indicators include but are not limited to the demand ratio of task sub-containers for various types of operating resources and the demand ratio of the task categories to be compared for various types of operating resources. The closer the demand ratios are, the greater the corresponding similarity.

[0123] Specifically, the greater the degree of dependence of a task category on the resource represented by the resource characteristic indicator, the greater the impact factor of the resource characteristic indicator.

[0124] Specifically, historical evaluation indicators include but are not limited to whether the current task subcontainer has been executed in the historical data packet of the task category to be compared, the number of executions, the historical average execution time, and other parameters for determining whether the task to be executed corresponding to the current task subcontainer is a historical execution task.

[0125] Specifically, all parameters in the model are normalized by presetting a first fixed coefficient and a second fixed coefficient, so that each parameter in the model is in the same value range.

[0126] Specifically, when establishing the load evaluation value series D, it includes:

[0127] Select the target first-level node according to all first-level nodes in sequence;

[0128] Generate the load evaluation value d of the target first-level node;

[0129] d= α i *j i ;

[0130] Among them, θ4 is the number of load evaluation indicators; α i is the influencing factor of the i-th load evaluation index; j i is the reference value of the i-th load evaluation index in the target first-level node.

[0131] Specifically, the load evaluation indicators include, but are not limited to, load value, the number of tasks to be executed in the running node, the average waiting time of each task to be executed, and other parameter values ​​associated with the running status.

[0132] Specifically, the greater the correlation between the load evaluation index and the operating status, the greater the corresponding impact factor value.

[0133] It can be understood that in the above embodiment, by containerizing the tasks to be executed and then selecting the corresponding running nodes, the running security of the tasks to be executed is guaranteed and the running efficiency of each task to be executed is improved.

[0134] In a preferred embodiment of the present application, determining whether to generate a correction instruction based on all feedback data packets includes:

[0135] Set a in sequence according to the number of running nodes A i is the node to be monitored;

[0136] Set all task sub-containers in the sub-node to be monitored as running sub-containers;

[0137] Set the running sub-container sequence Y of the node to be monitored, Y=(y1, y2…y i …y r ), where y i is the i-th running sub-container of the node to be monitored; r is the number of running sub-containers of the node to be monitored;

[0138] Set y in sequence according to the number of running sub-containers Y i Run a subcontainer for the target;

[0139] Get the feedback data packet of the target running sub-container at the current feedback time node;

[0140] Generate the running evaluation value f of the target running sub-container at the current feedback time node;

[0141] Preset operation evaluation value threshold F1;

[0142] If f <F1,当前反馈时间节点生成目标运行子容器的一级修正指令;

[0143] It is determined in turn whether each running sub-container generates a first-level correction instruction at the current feedback time node.

[0144] Specifically, the operational evaluation value threshold can be set based on historical parameters.

[0145] Specifically, the first-level correction instruction refers to dynamically correcting the resource usage of the target running sub-container to ensure the operating efficiency of the target running sub-container. The correction process is to reallocate the resources within the running node.

[0146] Specifically, when generating the running evaluation value f, it includes:

[0147] f= g i *u i ;

[0148] Among them, θ5 is the number of preset operation evaluation indicators; g i is the influencing factor of the i-th operation evaluation index; u i The reference value of the i-th running evaluation indicator of the target running sub-container at the current feedback time node.

[0149] Specifically, a larger running evaluation value indicates a higher overall running performance of the running sub-container.

[0150] It is understandable that in the above embodiment, by periodically collecting resource consumption data of each task sub-container, the resource allocation strategy is dynamically modified in a timely manner to ensure the operating performance of the task sub-container.

[0151] In a preferred embodiment of the present application, when determining whether to generate a correction instruction based on all feedback data packets, the method further includes:

[0152] Generate the operation evaluation value sequence F of the node to be monitored at the current feedback time node, F=(f1, f2…f i …f r ), where f i is the running evaluation value of the i-th running sub-container of the node to be monitored; r is the number of running sub-containers of the node to be monitored;

[0153] Generate the performance deviation value h of the node to be monitored based on the running evaluation value series F;

[0154] Set the performance deviation threshold H1;

[0155] If h>H1, generate a secondary correction instruction.

[0156] Specifically, the performance deviation value threshold can be set according to historical parameters. The larger the performance deviation value, the lower the overall operating efficiency of the node to be monitored.

[0157] Specifically, the secondary correction instruction refers to reallocating resources within the virtual resource pool to ensure the operating efficiency within each operating node.

[0158] Specifically, when generating the performance deviation value u, it includes:

[0159] h=e3*Q3*[ λ i *k i ]+e4*Q4*(d'-d'1);

[0160] Among them, e3 is the preset third weight coefficient; e4 is the preset fourth weight coefficient; Q3 is the preset third fixed coefficient; Q4 is the preset fourth fixed coefficient; θ6 is the number of auxiliary evaluation indicators; λ i is the influencing factor of the i-th auxiliary evaluation index; k i is the reference value of the i-th auxiliary evaluation index generated based on the running evaluation value series; d' is the load evaluation value of the node to be monitored at the current feedback time node; d'1 is the expected load value of the node to be monitored at the current feedback time node.

[0161] Specifically,

[0162] Specifically, all parameters in the model are normalized by presetting the third fixed coefficient and the fourth fixed coefficient, so that each parameter in the model is in the same value range.

[0163] According to the first concept of this application, a virtual resource pool is constructed by virtualizing all hardware devices, and by establishing multiple operating nodes to adapt to different types of tasks to be executed, various types of operating resources in the virtual resource pool are dynamically allocated according to the characteristic parameters of each operating node, thereby improving the scheduling and allocation efficiency of operating resources.

[0164] According to the second concept of this application, by containerizing the tasks to be executed and then selecting the corresponding running nodes, the running safety of the tasks to be executed is guaranteed, the running efficiency of each task to be executed is improved, and by periodically collecting the resource consumption data of each task sub-container, the resource allocation strategy is dynamically corrected in time to ensure the running performance of the task sub-container.

[0165] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and replacements can be made without departing from the technical principles of the present application. These improvements and replacements should also be regarded as the scope of protection of the present application.

Claims

1. A resource pool security management method based on virtualization and containerization, characterized in that: including: Construct a virtual resource pool based on resource device parameters; Set multiple running nodes in the virtual resource pool according to historical task parameters, and generate a first-level resource allocation policy based on all running nodes; Set task sub-containers according to the tasks to be executed, and select the nodes to be executed according to the characteristic parameters of the task sub-containers; Obtain the feedback data packets of each task sub-container according to the preset feedback time nodes, and determine whether to generate a correction instruction according to all feedback data packets; When setting multiple running nodes, including: Set multiple task characteristic indicators according to historical task parameters; Generate multiple task categories based on all task characteristic indicators and establish a task category sequence P, P=(p1, p2…p i …p m ), where p i is the i-th task category; m is the number of task categories; According to the task category sequence P, set p in sequence i is the target task category; Generate the demand evaluation value s of the target task category; s= b i *v i ? Among them, θ1 is the number of demand evaluation indicators; β i is the influencing factor of the i-th demand evaluation indicator; vi is the reference value of the i-th demand evaluation indicator in the target task category; Set the number of running nodes of the target task category according to the demand evaluation value s; Set the number of running nodes of each task category in turn; Establish a running node sequence A, A=(a1, a2…a i …a n ), where a i is the i-th running node; n is the number of running nodes, and n>m; Set the task category-running node mapping table according to the running node sequence A and the task category sequence P; When generating a first-level resource allocation policy based on all running nodes, including: Set a in sequence according to the number of running nodes A i is the target node; Obtain the characteristic data packets of the task category corresponding to the target nodes according to the task category-running node mapping table; Set the running architecture of the target nodes according to the characteristic data packets; Generate the resource sub-policy of the target nodes; Generate the running architectures and resource sub-policies of each running node in turn; Generate a first-level resource allocation policy according to all resource sub-policies; When selecting the nodes to be executed according to the characteristic parameters of the task sub-containers, including: Set pi as the task category to be compared in turn according to the task category sequence P; Generate the similarity evaluation value c between the task sub-container and the task category to be compared; Generate similar evaluation values ​​of task sub-containers and each task category in turn, and establish similar evaluation value sequence C, C = (c1, c2…c i …c m ), where c i is the similarity evaluation value between the task subcontainer and the i-th task category; Set the maximum value c in the similarity evaluation value sequence C max The corresponding task category is the first-level task category; Set the running nodes corresponding to the first-level task categories as the first-level nodes according to the task category-running node mapping table; Generate the load evaluation values of each first-level node; Establish a load evaluation value series D, D = (d1, d2…d i …d n2 ), where d i is the load evaluation value of the i-th first-level node; n2 is the number of first-level nodes; Set the minimum value d in the load evaluation value sequence D min The corresponding first-level node is the node to be executed.

2. The resource pool security management method based on virtualization and containerization according to claim 1, characterized in that: When generating the similarity evaluation value c between the task sub-container and the task category to be compared, including: c=e1*Q1*[ or 1i *w i ]+e2*Q2*[ or 2i *t i ]; Wherein, e1 is the preset first weight coefficient; e2 is the preset second weight coefficient; Q1 is the preset first fixed coefficient; Q2 is the preset second fixed coefficient; θ2 is the number of resource characteristic indicators; η 1i is the influencing factor of the i-th resource characteristic indicator; w i is the similarity between the resource feature index of the task subcontainer and the task category to be compared; θ3 is the number of historical evaluation indicators; η 2i is the impact factor of the i-th historical evaluation index; t i is the reference value of the i-th historical evaluation indicator.

3. The resource pool security management method based on virtualization and containerization according to claim 2, characterized in that: When establishing the load evaluation value sequence D, including: Select the target first-level nodes in turn according to all first-level nodes; Generate the load evaluation value d of the target first-level node; d= a i *j i ? Among them, θ4 is the number of load evaluation indicators; α i is the influencing factor of the i-th load evaluation index; j i is the reference value of the i-th load evaluation index in the target first-level node.

4. The resource pool security management method based on virtualization and containerization according to claim 3, characterized in that: When determining whether to generate a correction instruction according to all feedback data packets, including: Set a in sequence according to the number of running nodes A i is the node to be monitored; Set all task sub-containers in the sub-nodes to be monitored as running sub-containers; Set the running sub-container sequence Y of the node to be monitored, Y=(y1, y2…y i …y r ), where y i is the i-th running sub-container of the node to be monitored; r is the number of running sub-containers of the node to be monitored; Set y in sequence according to the number of running sub-containers Y i Run a subcontainer for the target; Obtain the feedback data packet of the target running sub-container at the current feedback time node; Generate the running evaluation value f of the target running sub-container at the current feedback time node; Preset the running evaluation value threshold F1; If f < F1, generate a first-level correction instruction for the target running sub-container at the current feedback time node; Judge whether to generate a first-level correction instruction for each running sub-container at the current feedback time node in turn.

5. The resource pool security management method based on virtualization and containerization according to claim 4, characterized in that: When generating the running evaluation value f, including: f= g i *u i ; Among them, θ5 is the number of preset operation evaluation indicators; g i is the influencing factor of the i-th operation evaluation index; u i The reference value of the i-th running evaluation indicator of the target running sub-container at the current feedback time node.

6. The resource pool security management method based on virtualization and containerization according to claim 5, characterized in that: When determining whether to generate a correction instruction according to all feedback data packets, it also includes: Generate the operation evaluation value sequence F of the node to be monitored at the current feedback time node, F=(f1, f2…f i …f r ), where f i is the running evaluation value of the i-th running sub-container of the node to be monitored; r is the number of running sub-containers of the node to be monitored; Generate the performance deviation value h of the sub-nodes to be monitored according to the running evaluation value sequence F; Set the performance deviation value threshold H1; If h > H1, generate a second-level correction instruction.

7. The resource pool security management method based on virtualization and containerization according to claim 6, characterized in that: When generating the performance deviation value u, including: h=e3*Q3*[ λ i *k i ]+e4*Q4*(d'-d'1); Among them, e3 is the preset third weight coefficient; e4 is the preset fourth weight coefficient; Q3 is the preset third fixed coefficient; Q4 is the preset fourth fixed coefficient; θ6 is the number of auxiliary evaluation indicators; λ i is the influencing factor of the i-th auxiliary evaluation index; k i is the reference value of the i-th auxiliary evaluation index generated based on the running evaluation value series; d' is the load evaluation value of the node to be monitored at the current feedback time node; d'1 is the expected load value of the node to be monitored at the current feedback time node.

Images