Resource scheduling method, device, equipment, medium and product

Abstract

The invention discloses a resource scheduling method, device and equipment, a medium and a product. The method comprises the steps of determining at least one to-be-assessed scheduling mode corresponding to a resource scheduling request, and determining at least one to-be-assessed scheduling mode according to current state data of a plurality of resource objects in the to-be-assessed scheduling mode, preset tolerable state values corresponding to the resource objects under associated assessment dimensions and preset importance coefficients corresponding to the assessment dimensions, determining a first assessment attribute of the to-be-assessed scheduling mode, and determining a second assessment attribute of the to-be-assessed scheduling mode based on the current state data of the at least one resource object under the same assessment dimension; determining a third evaluation attribute of the to-be-evaluated scheduling mode based on the first evaluation attribute and the second evaluation attribute; and determining a target scheduling mode based on each third evaluation attribute, and processing the target task based on the target scheduling mode. The reliability and efficiency of task processing are improved while it is ensured that the allocated target scheduling mode meets the real-time processing requirement of the task.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer processing, and in particular to a resource scheduling method, device, equipment, medium and product. BACKGROUND

[0002] With the rapid development of information technology, especially the wide application of emerging technologies such as cloud computing, big data and artificial intelligence, the number and complexity of various computing tasks are increasing, and the demand for computing resources is also increasing. In order to ensure that the computing task can be accurately and efficiently executed, it is usually necessary to allocate corresponding resources to it.

[0003] At present, the existing resource allocation method is mainly based on the explicit demand of the computing task for the resource, and the remaining resources are allocated to the corresponding computing task by filtering out the resources that do not meet these demands.

[0004] However, in the face of high concurrency complex scenarios, the allocated resources are difficult to meet the task processing demand, and there are problems of low efficiency and poor performance in processing tasks. SUMMARY

[0005] The present application provides a resource scheduling method, device, equipment, medium and product to ensure that the target scheduling mode of the allocated resources meets the real-time processing demand of the task while improving the reliability and efficiency of processing the task.

[0006] According to an aspect of the present application, a resource scheduling method is provided, which comprises:

[0007] Upon receiving a resource scheduling request, at least one to-be-evaluated scheduling mode corresponding to the resource scheduling request is determined, and current state data of a plurality of resource objects under the at least one to-be-evaluated scheduling mode is obtained; wherein each resource object is associated with at least one evaluation dimension;

[0008] For each to-be-evaluated scheduling mode, based on the current state data of the plurality of resource objects in the to-be-evaluated scheduling mode, the preset tolerable state value of the resource object in each evaluation dimension associated therewith, and the preset importance coefficient corresponding to each evaluation dimension, a first evaluation attribute of the to-be-evaluated scheduling mode in each evaluation dimension is determined, and based on the current state data of at least one resource object in the same evaluation dimension, a second evaluation attribute of the to-be-evaluated scheduling mode in the evaluation dimension is determined;

[0009] Based on at least one first evaluation attribute and second evaluation attribute corresponding to the to-be-evaluated scheduling mode, a third evaluation attribute corresponding to the to-be-evaluated scheduling mode is determined;

[0010] determine a target scheduling mode from the at least one to-be-evaluated scheduling mode based on each of the third evaluation attributes, so as to process a target task corresponding to the resource scheduling request based on the target scheduling mode.

[0011] According to another aspect of the present application, there is provided a resource scheduling device, which comprises:

[0012] a to-be-evaluated scheduling mode determination module, configured to determine at least one to-be-evaluated scheduling mode corresponding to a resource scheduling request and acquire current state data of a plurality of resource objects in the at least one to-be-evaluated scheduling mode when the resource scheduling request is received, wherein each of the resource objects is associated with at least one evaluation dimension;

[0013] a second evaluation attribute determination module, configured to determine, for each of the to-be-evaluated scheduling modes, a first evaluation attribute of the to-be-evaluated scheduling mode in each of the evaluation dimensions based on the current state data of the plurality of resource objects in the to-be-evaluated scheduling mode, preset tolerable state values corresponding to the resource objects in each of the evaluation dimensions associated with the resource objects, and preset importance coefficients corresponding to each of the evaluation dimensions, and determine a second evaluation attribute of the to-be-evaluated scheduling mode in the evaluation dimension based on the current state data of at least one of the resource objects in the evaluation dimension;

[0014] a third evaluation attribute determination module, configured to determine a third evaluation attribute corresponding to the to-be-evaluated scheduling mode based on at least one of the first evaluation attribute and the second evaluation attribute corresponding to the to-be-evaluated scheduling mode;

[0015] a target scheduling mode determination module, configured to determine a target scheduling mode from the at least two to-be-evaluated scheduling modes based on each of the third evaluation attributes, so as to process a target task corresponding to the resource scheduling request based on the target scheduling mode.

[0016] According to another aspect of the present application, there is provided an electronic device, which comprises:

[0017] at least one processor; and a memory connected with the at least one processor in communication; wherein

[0018] the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the resource scheduling method according to any one of the embodiments of the present application.

[0019] According to another aspect of the present application, there is provided a computer readable storage medium storing computer instructions for causing a processor to implement the resource scheduling method according to any of the embodiments of the present application when executed by the processor.

[0020] According to another aspect of the present application, there is provided a computer program product comprising a computer program for implementing the resource scheduling method according to any of the embodiments of the present application when executed by a processor.

[0021] The technical scheme of the embodiment of the present application comprises the following steps: when a resource scheduling request is received, at least one to-be-evaluated scheduling mode corresponding to the resource scheduling request is determined, and current state data of a plurality of resource objects under the at least one to-be-evaluated scheduling mode is acquired; based on the current state data of the plurality of resource objects in the to-be-evaluated scheduling mode, preset tolerable state values corresponding to the resource objects under each evaluation dimension associated with the resource objects, and preset importance coefficients corresponding to each evaluation dimension, a first evaluation attribute of the to-be-evaluated scheduling mode under each evaluation dimension is determined, and based on the current state data of at least one resource object under the same evaluation dimension, a second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension is determined; based on at least one first evaluation attribute and second evaluation attribute corresponding to the to-be-evaluated scheduling mode, a third evaluation attribute corresponding to the to-be-evaluated scheduling mode is determined; and based on each third evaluation attribute, a target scheduling mode is determined from the at least one to-be-evaluated scheduling mode, so as to process a target task corresponding to the resource scheduling request based on the target scheduling mode, thereby solving the problem that in the prior art, a resource is filtered based on a demand, and the allocated resource cannot meet the real-time processing demand of the task, resulting in low efficiency and poor performance in processing the task, and achieving the following: when a resource scheduling request is received, at least one to-be-evaluated scheduling mode corresponding to the resource scheduling request is preliminarily determined, and current state data of a plurality of resource objects under the at least one to-be-evaluated scheduling mode is acquired. Further, based on the current state data of the plurality of resource objects in the to-be-evaluated scheduling mode, preset tolerable state values corresponding to the resource objects under each evaluation dimension associated with the resource objects, and preset importance coefficients corresponding to each evaluation dimension, the first evaluation attribute of the to-be-evaluated scheduling mode under each evaluation dimension is comprehensively evaluated, and based on the current state data of at least one resource object under the same evaluation dimension, the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension is determined, so as to ensure that the evaluation attribute meets the real-time task processing demand, and to realize multi-dimensional and multi-level evaluation of the performance of the to-be-evaluated scheduling mode under each evaluation dimension, avoid one-sidedness of single-index evaluation, and improve evaluation accuracy. Further, by comprehensively evaluating at least one first evaluation attribute and second evaluation attribute corresponding to the to-be-evaluated scheduling mode, a third evaluation attribute corresponding to the to-be-evaluated scheduling mode is determined, and finally, based on each third evaluation attribute, a target scheduling mode is selected, so that the allocated target scheduling mode meets the real-time processing demand of the task, and the reliability and efficiency of processing the task are improved.

[0022] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to make the technical solution in the embodiments of the present application clearer, the accompanying drawings needed in the embodiments description will be briefly introduced as follows. Obviously, the accompanying drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of the accompanying drawings.

[0024] Figure 1 is a flow chart of a resource scheduling method according to the first embodiment of the present application;

[0025] Figure 2 is a flow chart of a resource scheduling method according to the second embodiment of the present application;

[0026] Figure 3 is a flow chart of a resource scheduling method according to the second embodiment of the present application;

[0027] Figure 4 is a structural schematic diagram of a resource scheduling device according to the third embodiment of the present application;

[0028] Figure 5 is a structural schematic diagram of an electronic device implementing the resource scheduling method according to the embodiments of the present application. DETAILED DESCRIPTION

[0029] In order to make the technical solution in the embodiments of the present application clearer, the accompanying drawings needed in the embodiments description will be briefly introduced as follows. Obviously, the accompanying drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of the accompanying drawings.

[0030] It should be noted that the terms "first", "second", and the like in the description and claims of the present application and the above-mentioned accompanying drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not necessarily have to include only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to the process, method, product or device.

[0031] Embodiment One

[0032] Figure 1is a flowchart of a resource scheduling method according to an embodiment of the present application. The embodiment can be applied to the case of allocating resource objects for processing related tasks. The method can be executed by a resource scheduling device, which can be implemented in the form of hardware and / or software, and can be configured in a computing device. As shown in Figure 1 the method comprises the following steps.

[0033] S110, upon receiving a resource scheduling request, determining at least one to-be-evaluated scheduling mode corresponding to the resource scheduling request, and obtaining current state data of a plurality of resource objects under the at least one to-be-evaluated scheduling mode.

[0034] The resource scheduling request can be a program or code for requesting allocation of resource objects for processing one or more tasks. The to-be-evaluated scheduling mode refers to a resource to-be-evaluated scheduling mode that needs to be evaluated, and different types of resource objects are included in the to-be-evaluated scheduling mode. The resource object refers to a specific resource that can be used for scheduling, for example, the resource object can include but is not limited to computing resources (such as CPU, GPU, NPU), storage resources (such as memory, disk), network resources (such as bandwidth or network access method, etc.), etc. The evaluation dimension refers to the index dimension used to evaluate the to-be-evaluated scheduling mode, and each resource object is associated with at least one evaluation dimension, i.e., one resource object needs to be evaluated under one or more evaluation dimensions. For example, the evaluation dimension includes but is not limited to performance dimension (such as computing power, storage speed, read-write speed, network bandwidth, resource utilization, job queuing waiting time, number of queued jobs, etc.), cost dimension (such as resource usage cost, network access cost, etc.), reliability dimension (such as node failure rate, network stability, etc.), etc. The current state data can refer to the real-time state value of the resource object, for example, the current state data of the computing resource object includes but is not limited to CPU usage, GPU usage, number of queued jobs, job queuing waiting time, energy consumption index, etc. The current state data of the storage resource object includes but is not limited to available resource amount, memory usage, storage capacity, read-write speed, storage type, etc. The current state data of the network access method includes but is not limited to bandwidth utilization, network remaining available bandwidth, network transmission delay, network packet loss rate, network jitter, etc.

[0035] In the embodiment, when receiving a resource scheduling request corresponding to a to-be-processed task, various requirement information such as resource requirement (for example, requirement including type, quantity, priority, use time, performance, and the like of a resource object), task requirement (for example, task type: computation-intensive task, data-intensive task, or interactive task, and the like; task priority; task dependency relationship; task execution time; task output requirement, and the like), and network requirement (for example, bandwidth requirement, delay requirement, security requirement, reliability requirement, and the like) can be extracted from the resource scheduling request. Further, a network access mode meeting the requirement can be screened according to the network requirement. According to the resource requirement and the task requirement, a computation resource object and a storage resource object meeting the condition can be screened from the resource object set. The network access mode, the computation resource object, and the storage resource object, and the like screened can be used to determine a plurality of possible to-be-evaluated scheduling modes. Further, current state data of all resource objects corresponding to each to-be-evaluated scheduling mode can be obtained through an interface or a function. At this time, in order to avoid repeated acquisition of data, a union set of all resource objects in each to-be-evaluated scheduling mode can be processed to obtain current state data of each resource object in the union set, so as to evaluate real-time performance of the to-be-evaluated scheduling mode in processing the task according to the current state data.

[0036] For example, it is assumed that there are three to-be-evaluated scheduling modes, which are scheduling mode 1, scheduling mode 2, and scheduling mode 3. The scheduling mode 1 includes resource object A, resource object B, and resource object C; the scheduling mode 2 includes resource object B, resource object C, and resource object D; and the scheduling mode 3 includes resource object C, resource object D, and resource object E. Then, the union set of resource objects in all to-be-evaluated scheduling modes is resource object A, resource object B, resource object C, resource object D, and resource object E. At this time, current state data of resource object A, resource object B, resource object C, resource object D, and resource object E needs to be obtained.

[0037] In S120, for each to-be-evaluated scheduling mode, a first evaluation attribute of the to-be-evaluated scheduling mode in each evaluation dimension is determined based on current state data of a plurality of resource objects in the to-be-evaluated scheduling mode, a preset tolerable state value corresponding to each resource object in each evaluation dimension to which the resource object is associated, and a preset importance coefficient corresponding to each evaluation dimension.

[0038] The preset tolerable state value can refer to a maximum or minimum state value of the state data of the resource object in the evaluation dimension that can be accepted. For example, the CPU usage of the resource object 1 is not higher than the preset tolerable state value 80%; the network bandwidth of the resource object 1 is at least the preset tolerable state value 100 Mbps; the resource usage cost of the resource object 2 is not higher than the preset tolerable state value 80% of the budget; the node failure rate of the resource object 2 is not lower than the preset tolerable state value 0.1%; and the network stability of the resource object 3 reaches the preset tolerable state value 89.9%. It should be noted that the preset tolerable state values of different resource objects in the same evaluation dimension can be the same or different. The preset importance coefficient can refer to the influence degree of the evaluation dimension on the overall evaluation result. For example, the importance coefficient of the computing power in the performance evaluation dimension is 0.5, the importance coefficient of the storage speed is 0.3, and the importance coefficient of the network bandwidth is 0.2; the importance coefficient of the resource usage cost in the cost evaluation dimension is 0.6, and the importance coefficient of the network access cost is 0.4; the importance coefficient of the node failure rate in the reliability evaluation dimension is 0.7, and the importance coefficient of the network stability is 0.3; or the importance coefficient of the performance evaluation dimension is higher than that of the cost evaluation dimension. It can be understood that the greater the preset importance coefficient is, the higher the influence degree of the evaluation dimension or the state data of the resource object in the evaluation dimension on the overall evaluation result is. The first evaluation attribute is used to represent the performance evaluation result of the to-be-evaluated scheduling mode in processing the task in different evaluation dimensions. The second evaluation attribute is the performance evaluation result determined based on the current state data of the resource object. In this embodiment, each evaluation attribute can be represented by a numerical value or a percentage, for example, the higher the numerical value of the evaluation attribute is, the more reliable and better the effect of using the to-be-evaluated scheduling mode to process the task is.

[0039] It should be noted that the evaluation of each to-be-evaluated scheduling mode is the same, and any to-be-evaluated scheduling mode can be taken as an example for introduction.

[0040] In the embodiment, for a to-be-evaluated scheduling mode, the current state data of the resource object in the to-be-evaluated scheduling mode and the preset tolerable state value corresponding to the resource object in different evaluation dimensions are compared to determine whether the current state data of the resource object meets the requirement. Further, the evaluation attribute corresponding to the evaluation dimension can be determined as the first evaluation attribute according to the comparison result corresponding to each resource object in the same evaluation dimension and the preset importance coefficient corresponding to the evaluation dimension. For example, whether the network bandwidth in the current state data is lower than the preset tolerable state value 100 Mbps is judged: if yes, it is considered that the comparison result of the network bandwidth is not met the requirement, and if no, it is considered that the comparison result of the network bandwidth meets the requirement. If the number of the resource objects meeting the requirement in the same evaluation dimension is more, the preset importance coefficient is higher, and the value of the first evaluation attribute of the to-be-evaluated scheduling mode in the evaluation dimension is higher. Alternatively, the current state data of the resource object in the to-be-evaluated scheduling mode and the preset tolerable state value corresponding to the resource object in the evaluation dimension associated with the resource object are processed to obtain the quotient value corresponding to the resource object. The closer the quotient value is to 1, the more the resource object can meet the task processing requirement, and the farther the quotient value deviates from 1, the less the resource object can meet the task processing requirement. If the number of the resource objects closer to 1 in the same evaluation dimension is more, the preset importance coefficient is higher, and the value of the first evaluation attribute of the to-be-evaluated scheduling mode in the evaluation dimension is higher.

[0041] In the embodiment, based on the current state data of the plurality of resource objects in the to-be-evaluated scheduling mode, the preset tolerable state value corresponding to the resource object in each evaluation dimension associated with the resource object, and the preset importance coefficient corresponding to each evaluation dimension, the first evaluation attribute of the to-be-evaluated scheduling mode in each evaluation dimension is determined, including: based on the current state data of at least one resource object in the same evaluation dimension, the preset tolerable state value corresponding to the resource object in the evaluation dimension, and the preset importance coefficient of the evaluation dimension, the fourth evaluation attribute of the evaluation dimension is determined; wherein the evaluation dimension includes at least two; for each evaluation dimension, based on the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attribute of the other evaluation dimension different from the current evaluation dimension, the first evaluation attribute of the to-be-evaluated scheduling mode in the current evaluation dimension is determined.

[0042] It can be understood that the current state data of each resource object under the same evaluation dimension can be compared with the preset tolerable state value corresponding to the evaluation dimension to determine the comparison result of each resource object under the same evaluation dimension. If the comparison result under the same evaluation dimension is that the more resource objects meet the requirements, the higher the preset importance coefficient of the evaluation dimension is, it can be considered that the value of the fourth evaluation attribute of the evaluation dimension under the to-be-evaluated scheduling mode is higher. For example, the proportion of the resource objects meeting the requirements under the same evaluation dimension to all resource objects under the evaluation dimension can be determined, and the proportion is multiplied by the preset importance coefficient of the evaluation dimension as the fourth evaluation attribute corresponding to the evaluation dimension. Alternatively, the preset tolerable state value of the resource object under the evaluation dimension can also be used to evaluate the current state data of the resource object to obtain an evaluation result, and the fourth evaluation attribute of the current evaluation dimension can be calculated according to the evaluation result of each resource object under the current evaluation dimension and the preset importance coefficient. Correspondingly, the fourth evaluation attribute corresponding to each evaluation dimension can be obtained. Further, the fourth evaluation attributes of the other evaluation dimensions except the current evaluation dimension can be summed to obtain a sum value, and the quotient value between the fourth evaluation attribute of the current evaluation dimension and the sum value is taken as the first evaluation attribute of the current evaluation dimension under the to-be-evaluated scheduling mode, so as to comprehensively evaluate the to-be-evaluated scheduling mode according to the plurality of first evaluation attributes.

[0043] The advantage of such a setting is that the fourth evaluation attribute of the evaluation dimension is preliminarily evaluated by combining the current state data of at least one resource object under the same evaluation dimension, the preset tolerable state value of each resource object under the evaluation dimension, and the preset importance coefficient of the evaluation dimension. Further, the first evaluation attribute of the current evaluation dimension under the to-be-evaluated scheduling mode can be determined by considering the evaluation result of the current evaluation dimension and the influence of other evaluation dimensions on the current evaluation dimension, which can improve the accuracy and comprehensiveness of the evaluation of the to-be-evaluated scheduling mode.

[0044] It should be noted that the resource objects under different evaluation dimensions may be state values greater than or less than the evaluation attribute. For example, the greater the bandwidth of the resource object, the higher the ability to process data (such as faster data transmission, download or upload speed); the smaller the consumption cost of the resource object, the less the resource consumption of processing tasks. That is, the evaluation standards of different evaluation dimensions are different. In this embodiment, in order to improve the accuracy of the evaluation of the to-be-evaluated scheduling mode, different evaluation dimensions can be distinguished and evaluated and analyzed by distinguishing the positive indicators and the negative indicators, so as to ensure the accuracy of the evaluation.

[0045] In this embodiment, there are at least two implementation manners for determining the fourth evaluation attribute of the evaluation dimension, which will be introduced below.

[0046] An implementation can be: in the case of evaluating a dimension as a positive indicator dimension, for each resource object under the evaluation dimension, determining a first state performance attribute corresponding to the resource object based on a quotient value between current state data of the resource object and a preset tolerable state value corresponding to the resource object under the evaluation dimension; and determining a fourth evaluation attribute of the evaluation dimension based on the first state performance attribute of each resource object and a preset importance coefficient corresponding to the evaluation dimension.

[0047] The positive indicator dimension can refer to an evaluation dimension in which the greater the state value of a resource object, the better the performance. For example, computing power, storage speed, read-write speed, network bandwidth, resource utilization, and the like are all positive indicator dimensions. The first state performance attribute can be used to represent the closeness of the state data of the resource object to the preset tolerable state value. For a positive indicator dimension, the greater the first state performance attribute, the better the state data of the resource object.

[0048] Specifically, when the evaluation dimension is a positive indicator dimension, for each resource object under the evaluation dimension, the current state data of the resource object and the preset tolerable state value corresponding to the resource object under the evaluation dimension can be quotient processed to obtain a quotient value. The quotient value is taken as the first state performance attribute corresponding to the resource object. Accordingly, the first state performance attribute corresponding to each resource object can be obtained. The first state performance attributes corresponding to all resource objects under the evaluation dimension are summed to obtain a sum value, or the first state performance attributes corresponding to all resource objects under the evaluation dimension are summed and averaged to obtain a mean value. The sum value or the mean value is taken as a third intermediate value. The third intermediate value and the preset importance coefficient corresponding to the evaluation dimension are multiplied to obtain the fourth evaluation attribute of the evaluation dimension. Accordingly, the fourth evaluation attribute of each evaluation dimension that is a positive indicator dimension can be obtained.

[0049] For example, the fourth evaluation attribute of the evaluation dimension can be determined based on formula (1), which can be expressed as: wherein A ia represents the fourth evaluation attribute of the a-th positive indicator dimension under the i-th to-be-evaluated scheduling mode; h represents the h-th resource object under the a-th positive indicator dimension in the i-th to-be-evaluated scheduling mode; S a (t) represents the current state data of the t-th resource object under the a-th positive indicator dimension, represents the preset tolerable state value corresponding to the t-th resource object under the a-th positive indicator dimension; a a is the preset importance coefficient of the a-th positive indicator dimension, and ∑ represents a summation function.

[0050] Another implementation manner can be: in the case that the evaluation dimension is a negative indicator dimension, for each resource object under the evaluation dimension, a second state performance attribute of the resource object is determined based on a quotient value between current state data of the resource object and a preset tolerable state value corresponding to the resource object under the evaluation dimension and a preset parameter value; and a fourth evaluation attribute of the evaluation dimension is determined based on the second state performance attribute of each resource object and a preset importance coefficient corresponding to the evaluation dimension.

[0051] The negative indicator dimension can refer to an evaluation dimension in which the smaller the state value of a resource object is, the better the performance is. For example, the cost operation queuing waiting time, the number of queued waiting jobs, the resource use cost, the network access cost, and the like are all negative indicator dimensions. For the negative indicator dimension, the larger the second state performance attribute is, the better the state data of the resource object is.

[0052] Specifically, when the evaluation dimension is a negative indicator dimension, for each resource object under the evaluation dimension, the current state data of the resource object and the preset tolerable state value corresponding to the resource object under the evaluation dimension can be quotient processed to obtain a quotient value. The difference between the preset parameter value and the quotient value is taken as the second state performance attribute of the resource object. Correspondingly, the second state performance attribute corresponding to each resource object can be obtained. The second state performance attributes corresponding to all resource objects under the evaluation dimension are summed to obtain a sum value, or the second state performance attributes corresponding to all resource objects under the evaluation dimension are summed and averaged to obtain a mean value. The sum value or the mean value is taken as a fourth intermediate value. The fourth intermediate value and the preset importance coefficient corresponding to the evaluation dimension are multiplied to obtain the fourth evaluation attribute of the evaluation dimension. Correspondingly, the fourth evaluation attribute of each evaluation dimension that is a negative indicator dimension can be obtained.

[0053] For example, the fourth evaluation attribute of the evaluation dimension can be determined based on formula (2), which can be represented as: wherein A ij represents the fourth evaluation attribute of the jth negative indicator dimension under the ith to-be-evaluated scheduling mode; h represents h resource objects under the jth negative indicator dimension in the ith to-be-evaluated scheduling mode; S j (t) represents the current state data of the tth resource object under the jth negative indicator dimension, represents the preset tolerable state value corresponding to the tth resource object under the jth negative indicator dimension; a j is the preset importance coefficient of the jth negative indicator dimension, and ∑ represents a summation function.

[0054] In the case that the positive indicator dimension and the negative indicator dimension are included in the at least two evaluation dimensions, the manner of determining the first evaluation attribute of the to-be-evaluated scheduling manner under the current evaluation dimension based on the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attribute of the other evaluation dimension different from the current evaluation dimension includes at least two implementation manners. The following will introduce each implementation manner of determining the first evaluation attribute of the to-be-evaluated scheduling manner under the current evaluation dimension.

[0055] One implementation manner can be that in the case that the current evaluation dimension is the positive indicator dimension, a first intermediate value is determined based on the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attribute of the other positive indicator dimension different from the current evaluation dimension, and the first evaluation attribute of the to-be-evaluated scheduling manner under the current evaluation dimension is determined based on the fourth evaluation attribute of the current evaluation dimension and the first intermediate value.

[0056] In the embodiment, in the case that the current evaluation dimension is the positive indicator dimension, the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attribute of all the other positive indicator dimensions different from the current evaluation dimension can be summed to obtain a first sum value, or the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attribute of all the other positive indicator dimensions different from the current evaluation dimension can be summed and averaged to obtain a first average value; the first sum value or the first average value is taken as the first intermediate value. Further, the fourth evaluation attribute of the current evaluation dimension and the first intermediate value can be processed by quotient value to obtain the first evaluation attribute of the to-be-evaluated scheduling manner under the current evaluation dimension.

[0057] For example, the first evaluation attribute of the to-be-evaluated scheduling manner under the current evaluation dimension can be determined based on formula (3), which can be expressed as: wherein ω ia is the first evaluation attribute of the ith to-be-evaluated scheduling manner under the ath positive indicator dimension, and the larger the first evaluation attribute is, the better the performance of the to-be-evaluated scheduling manner under the positive indicator dimension is. b represents the total number of the positive indicator dimensions. S k (t) represents the current state data of the tth resource object of the kth positive indicator dimension in the ith to-be-evaluated scheduling manner. represents the current state data of the tth resource object under the kth positive indicator dimension, α k is a preset importance coefficient of the kth positive indicator dimension.

[0058] The fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attribute of other positive index dimensions related to the fourth evaluation attribute of the current evaluation dimension are comprehensively considered to determine the first intermediate value, and the first evaluation attribute of the to-be-evaluated scheduling manner in the current evaluation dimension is derived, so that the comprehensive performance of the to-be-evaluated scheduling manner in the current evaluation dimension can be more comprehensively reflected, and the one-sidedness caused by single-dimension evaluation can be avoided. By introducing the attribute information of other related dimensions, the influence of different factors on the evaluation result can be better balanced, so that the accuracy and reliability of the evaluation are improved.

[0059] Another implementation manner can be that, in a case where the current evaluation dimension is a negative index dimension, a second intermediate value is determined based on the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attribute of other negative index dimensions different from the current evaluation dimension, and the first evaluation attribute of the to-be-evaluated scheduling manner in the current evaluation dimension is determined based on the fourth evaluation attribute of the current evaluation dimension and the second intermediate value.

[0060] In the embodiment, in a case where the current evaluation dimension is a negative index dimension, the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attribute of all other negative index dimensions different from the current evaluation dimension can be summed to obtain a second sum value, or the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attribute of all other negative index dimensions different from the current evaluation dimension can be summed and averaged to obtain a second average value; the second sum value or the second average value is taken as the second intermediate value. Further, the fourth evaluation attribute of the current evaluation dimension and the second intermediate value can be processed by quotient value to obtain the first evaluation attribute of the to-be-evaluated scheduling manner in the current evaluation dimension.

[0061] For example, the first evaluation attribute of the to-be-evaluated scheduling manner in the current evaluation dimension can be determined based on formula (4), which can be represented as: wherein ω ij is the first evaluation attribute of the i th to-be-evaluated scheduling manner in the j th negative index dimension, and the larger the first evaluation attribute is, the better the performance of the to-be-evaluated scheduling manner in the negative index dimension is; c represents the total number of negative index dimensions.

[0062] The advantage of this mode is that by comprehensively considering the fourth evaluation attribute of the current negative indicator dimension and the fourth evaluation attribute of other negative indicator dimensions, determining the second intermediate value, and combining the fourth evaluation attribute of the current dimension to determine the first evaluation attribute, this mode can effectively avoid the deviation caused by single dimension evaluation, fully consider the correlation and mutual influence between different negative indicator dimensions, thereby more comprehensively reflecting the comprehensive table of the to-be-evaluated scheduling mode under the current negative indicator dimension, improving the accuracy and objectivity of the evaluation result, and at the same time ensuring that the first evaluation attribute of the to-be-evaluated scheduling mode under different evaluation dimensions is positively correlated with the performance evaluation of the to-be-evaluated scheduling mode under different evaluation dimensions, thereby guaranteeing the accuracy of subsequent evaluation.

[0063] In S130, the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension is determined based on the current state data of at least one resource object under the same evaluation dimension.

[0064] In this embodiment, the current state data of all resource objects under the same evaluation dimension can be processed, such as calculating the cost consumption, comprehensive performance, etc., to obtain the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension. Accordingly, the second evaluation attribute of the to-be-evaluated scheduling mode under each evaluation dimension can be obtained.

[0065] In this embodiment, the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension is determined based on the current state data of at least one resource object under the same evaluation dimension, including: based on a preset condition, evaluating the current state data of at least one resource object under the same evaluation dimension to obtain the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension. The preset condition can be pre-configured and used to evaluate whether the performance of the resource object under different evaluation dimensions meets the specific requirements. Optionally, the preset condition at least includes a computing power performance evaluation condition, a storage performance evaluation condition, a network performance evaluation condition, a security performance evaluation condition, an energy consumption evaluation condition, and a cost evaluation condition. The second evaluation attribute can be used to reflect whether the performance of the resource object under the corresponding evaluation dimension meets the requirements.

[0066] Specifically, in the case where the preset condition is the computing power performance evaluation condition, the current state data of all resource objects under the same evaluation dimension can be evaluated based on the computing power performance evaluation condition, such as the CPU usage rate being not more than 80%, the GPU computing power being at least 1000 GFLOPS, etc., to obtain the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension.

[0067] In a case where the preset condition is the storage performance evaluation condition, storage performance evaluation can be performed on the current state data of all resource objects under the same evaluation dimension based on the storage performance evaluation condition, such as that the storage capacity is at least 1 TB and the read-write speed is at least 100 MB / s, to obtain the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension.

[0068] In a case where the preset condition is the network performance evaluation condition, network performance evaluation can be performed on the current state data of all resource objects under the same evaluation dimension based on the network performance evaluation condition, such as that the network bandwidth is at least 100 Mbps, the delay is no more than 10 ms, and the packet loss rate is no more than 0.1%, to obtain the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension.

[0069] In a case where the preset condition is the security performance evaluation condition, network performance evaluation can be performed on the current state data of all resource objects under the same evaluation dimension based on the security performance evaluation condition, such as that the data encryption level is a preset level and the access control policy meets a preset standard, to obtain the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension.

[0070] In a case where the preset condition is the energy consumption evaluation condition, energy consumption evaluation can be performed on the current state data of all resource objects under the same evaluation dimension based on the energy consumption evaluation condition, such as that the device power consumption is no more than 1000 W and the energy efficiency ratio is at least 2 GFLOPS / W, to obtain the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension.

[0071] In a case where the preset condition is the cost evaluation condition, cost evaluation can be performed on the current state data of all resource objects under the same evaluation dimension based on the cost evaluation condition, such as that the source use cost is no more than 80% of the budget and the network access cost is no more than 20% of the budget, to obtain the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension.

[0072] S140, based on at least one first evaluation attribute and a second evaluation attribute corresponding to the to-be-evaluated scheduling mode, determining a third evaluation attribute corresponding to the to-be-evaluated scheduling mode.

[0073] In this embodiment, the to-be-evaluated scheduling mode can be comprehensively evaluated based on all first evaluation attributes and all second evaluation attributes corresponding to the to-be-evaluated scheduling mode, to obtain the third evaluation attribute. For example, the first evaluation attribute and the second evaluation attribute under the same evaluation dimension in the to-be-evaluated scheduling mode can be weighted to obtain the third evaluation attribute corresponding to the to-be-evaluated scheduling mode; or all first evaluation attributes and all second evaluation attributes are summed to obtain the third evaluation attribute corresponding to the to-be-evaluated scheduling mode.

[0074] The third evaluation attribute corresponding to the to-be-evaluated scheduling manner is determined based on at least one first evaluation attribute and a second evaluation attribute corresponding to the to-be-evaluated scheduling manner, including: determining a fifth evaluation attribute corresponding to an evaluation dimension based on the first evaluation attribute and the second evaluation attribute corresponding to the same evaluation dimension under the to-be-evaluated scheduling manner; and determining the third evaluation attribute corresponding to the to-be-evaluated scheduling manner based on the fifth evaluation attribute corresponding to each evaluation dimension.

[0075] Specifically, the first evaluation attribute and the second evaluation attribute corresponding to the same evaluation dimension under the to-be-evaluated scheduling manner can be multiplied to obtain the fifth evaluation attribute corresponding to the evaluation dimension, and accordingly, the fifth evaluation attribute corresponding to each evaluation dimension can be obtained. Furthermore, the fifth evaluation attribute corresponding to each evaluation dimension is summed to obtain the third evaluation attribute corresponding to the to-be-evaluated scheduling manner.

[0076] Exemplarily, the third evaluation attribute corresponding to the to-be-evaluated scheduling manner can be determined based on formula (5), which can be expressed as: wherein, Score i is the third evaluation attribute corresponding to the i th to-be-evaluated scheduling manner, ω id is the first evaluation attribute of the i th to-be-evaluated scheduling manner under the d th evaluation dimension (including a positive index dimension and a negative index dimension), z id is the second evaluation attribute of the i th to-be-evaluated scheduling manner under the d th evaluation dimension. Score i is higher, indicating that the performance of the to-be-evaluated scheduling manner in processing the task is better.

[0077] The technical scheme provided in the embodiment integrates the first evaluation attribute and the second evaluation attribute of the to-be-evaluated scheduling manner under each evaluation dimension, further determines the fifth evaluation attribute corresponding to each evaluation dimension, and finally determines the third evaluation attribute of the to-be-evaluated scheduling manner based on the fifth evaluation attribute of all evaluation dimensions, thereby realizing multi-dimensional and multi-level comprehensive evaluation of the to-be-evaluated scheduling manner and avoiding the limitation of single attribute evaluation. By comprehensively analyzing multiple evaluation attributes under the same evaluation dimension, the finally determined third evaluation attribute can more comprehensively and accurately evaluate the overall performance of the scheduling manner, thereby improving the comprehensiveness and accuracy of the evaluation.

[0078] S150, determining a target scheduling manner from the to-be-evaluated scheduling manners based on the third evaluation attribute corresponding to each to-be-evaluated scheduling manner, to process a target task corresponding to the resource scheduling request based on the target scheduling manner.

[0079] The target task refers to a specific task that needs to be executed in the resource scheduling request. For example, the target task includes a computing task, a data processing task, a storage task, etc.

[0080] In the embodiment, the third evaluation attribute corresponding to each to-be-evaluated scheduling manner can be sorted from high to low, and the to-be-evaluated scheduling manner corresponding to the highest third evaluation attribute is determined as the target scheduling manner. Then, the target task corresponding to the resource scheduling request can be processed based on each resource object in the target scheduling manner, or the target scheduling manner can be assigned to the requester corresponding to the resource scheduling request, so that the requester processes the target task corresponding to the resource scheduling request based on each resource object in the target scheduling manner.

[0081] For example, the third evaluation attribute of the to-be-evaluated scheduling manner 1 is 0.5, the third evaluation attribute of the to-be-evaluated scheduling manner 2 is 0.7, the third evaluation attribute of the to-be-evaluated scheduling manner 3 is 0.88, and the third evaluation attribute of the to-be-evaluated scheduling manner 4 is 0.82. At this time, the to-be-evaluated scheduling manner 3 can be selected as the target scheduling manner.

[0082] The technical solution provided in the embodiment can initially determine at least one to-be-evaluated scheduling manner corresponding to the resource scheduling request, and then acquire the current state data of the plurality of resource objects under the at least one to-be-evaluated scheduling manner. Then, based on the current state data of the plurality of resource objects in the to-be-evaluated scheduling manner, the preset tolerable state value corresponding to the resource object in each evaluation dimension, and the preset importance coefficient corresponding to each evaluation dimension, the first evaluation attribute of the to-be-evaluated scheduling manner in each evaluation dimension is comprehensively evaluated, and the second evaluation attribute of the to-be-evaluated scheduling manner in the evaluation dimension is determined based on the current state data of at least one resource object in the same evaluation dimension, so as to ensure that the evaluation attribute meets the real-time task processing requirement, and to realize the multi-dimensional and multi-level evaluation of the performance of the to-be-evaluated scheduling manner in each evaluation dimension, avoid the one-sidedness of single-index evaluation, and improve the evaluation accuracy. Further, the third evaluation attribute corresponding to the to-be-evaluated scheduling manner is determined by comprehensively evaluating at least one first evaluation attribute and second evaluation attribute of the to-be-evaluated scheduling manner, and finally, the target scheduling manner is selected based on each third evaluation attribute, so that the assigned target scheduling manner meets the real-time processing requirement of the task, improves the reliability and efficiency of processing the task, and solves the problem in the prior art that the assigned resource cannot meet the real-time processing requirement of the task, resulting in low efficiency and poor performance in processing the task.

[0083] Embodiment Two

[0084] Figure 2 It is a flowchart of a resource scheduling method according to the embodiment two of the present application, which is further refined on the basis of the foregoing embodiment. The specific implementation can be referred to the technical solution of the embodiment. Among them, the same or corresponding technical terms as the above embodiments are not described here.

[0085] As shown in Figure 2 , the method specifically comprises the following steps:

[0086] S210, when receiving the resource scheduling request, determining the resource scheduling demand and / or scheduling task content corresponding to the resource scheduling request.

[0087] The resource scheduling demand refers to the specific demand information of the resource object embodied in the resource scheduling request. The scheduling task content refers to the specific task description information involved in the resource scheduling request.

[0088] Referring to Figure 3 , when receiving the resource scheduling request, the resource scheduling request can be parsed, and the demand fields are extracted from the request, and the content in these demand fields is taken as the resource scheduling demand. For example, the resource scheduling demand includes but is not limited to the type of resource object (such as computing resource, storage resource, network bandwidth, etc.), the number of resources (such as how many computing resource objects are needed, the size of storage capacity, etc.), performance indicators, time range of use (such as the start time and end time of resource use), priority (such as the degree of urgency of the task), resource performance requirements (such as the number of CPU cores, GPU computing power, storage read / write speed, etc.), network requirements (such as whether low delay and high bandwidth network access is needed, whether a specific network topology structure is needed, etc.), etc. If the resource scheduling request contains specific task description (such as operation steps, operation content, etc.), these information is extracted as the scheduling task content. For example, the scheduling task content includes but is not limited to the type of task (such as computing-intensive, data-intensive or I / O-intensive task), the task dependency relationship (such as whether there is a sequence between tasks, whether parallel execution is needed), the geographical location requirement of the task (such as whether the resource object needs to be executed close to the data source or the user end), the network access requirement of the task (such as whether the task needs a specific network access method to obtain data or communicate with the outside), the steps of the task to be executed, the specific content of the operation, etc.

[0089] The resource scheduling request can also be checked, such as checking whether it conforms to the predetermined format specification, checking its integrity, etc., so as to assign a target scheduling mode to the resource scheduling request that passes the verification.

[0090] Next, the technical solutions provided by the present embodiment are explained in terms of flow, taking the scheduling platform executing the technical solutions provided by the present embodiment as an example. When the client and the scheduling platform interact, the specific implementation manner can be: when the client receives the resource scheduling request of the user, the client sends the resource scheduling demand in the resource scheduling request to the scheduling platform through the API mode, the scheduling platform performs calculation according to the resource scheduling method provided by the present embodiment, and determines the target scheduling mode. Alternatively, the client sends the entire content of the computing task (i.e., the target task) to the scheduling platform as the resource scheduling request, including the resource scheduling demand of the computing task and the related input data of the computing task (i.e., the scheduling task content); and then the scheduling platform performs scheduling and subsequent operations of the computing task. After receiving the resource scheduling request of a computing task, the scheduling platform first performs demand analysis, determines the resource object that needs to be scheduled by the scheduling platform through task decomposition, and obtains the target scheduling mode. The client obtains the scheduling result (including the target scheduling mode) from the message returned by the scheduling platform, and then performs subsequent operations on the computing task according to the scheduling result.

[0091] S220, determining at least one network access mode based on the resource scheduling demand and / or the scheduling task content, and determining at least one candidate resource node from the resource object set.

[0092] The network access mode can be an Internet access mode or a dedicated line access mode, and the dedicated line access mode can be divided into different types (such as OTN dedicated line, IP-RAN dedicated line, PON dedicated line, etc.). The resource object set includes resource nodes such as computing resource objects and storage resource objects.

[0093] Continuing to refer to Figure 3According to the network access requirement in the task content, at least one network access mode can be selected. For example, if the task requires fast response (such as real-time calculation, online interaction scene, etc.), a low-latency network access mode is preferred, such as a fiber network or a 5G network; if the task is a data-intensive task (such as large-scale data transmission, video stream processing, etc.), a high-bandwidth network access mode is selected, such as a gigabit Ethernet or a high-speed fiber network; if the task needs to be executed on a mobile device or communicate with a mobile device, a wireless network access mode is preferred, such as Wi-Fi or 5G. If the task is a task involving sensitive data, a high-security network access mode is selected, such as a dedicated virtual private network (VPN) or an encrypted fiber network. Further, according to the network requirements in the resource scheduling requirements, the selected network access mode can be further refined to obtain at least one network access mode. Alternatively, if the resource scheduling requirements specify the geographical location or network topology of the resources, the network access mode that matches it is selected. For example, if the resources are distributed in multiple data centers, a network access mode that can support multi-point interconnection is selected, such as an MPLS (Multiprotocol Label Switching) network. Alternatively, according to the task priority in the resource scheduling requirements, the network bandwidth and access mode are allocated, for example, high-priority tasks can be allocated higher network bandwidth or better network access mode. For example, the network connection mode can be the connection network between the client, the computing resource object, and the storage resource object.

[0094] At the same time, according to the task type, task dependency relationship, geographical location requirement of the task, step required to be executed by the task, specific content of the operation, etc. in the task content, the computing resource object that meets the requirements can be selected from the resource object set as the candidate resource node. Alternatively, according to the computing resource requirement (such as CPU core number, memory capacity, calculation time, etc.) and storage resource requirement (such as storage space size, read-write speed, data persistence, etc.) in the resource scheduling requirements, the candidate resource node that meets the condition can be selected from the resource object set. Alternatively, according to the resource scheduling requirements and the task content, the candidate resource node that meets the condition is comprehensively determined.

[0095] It should be noted that according to the resource scheduling requirements and / or scheduling task content, the types of various computing, storage, network resource objects that need to be scheduled for the current to-be-processed task can be determined to filter the candidate resource nodes that meet the conditions in the resource object subset of the resource object type that needs to be scheduled in the resource object set, and the search efficiency is improved. For example, for a to-be-processed task, the user only has one network access method, i.e., the Internet, to select different computing clusters to submit the to-be-processed task. In this scenario, the network resource object does not belong to the schedulable resource object type, and the computing resource object represented by the computing cluster belongs to the schedulable resource object type. At this time, the Internet network access method can be regarded as a resource object, and the remaining network access methods do not need to be searched in the resource object set. For another example, for a to-be-processed task, data streams need to be sent to the same computing and storage integrated computing cluster through different networks or network paths for processing. In this scenario, the computing resource object and the storage resource object do not belong to the schedulable resource object type, and the network resource object (i.e., the network access method) belongs to the schedulable resource object type. For another example, for a to-be-processed task, one of multiple copy data sources needs to be selected as a data source center, and the data set file stored in the center needs to be transmitted to the data set cache layer of the target data center for processing. In this scenario, the storage resource object that stores the data set file belongs to the schedulable resource object type.

[0096] For example, the types of resource objects that need to be scheduled for each to-be-processed task can be seen from Table 1 below.

[0097] Table 1

[0098] To-be-processed task ID Resource object type that needs to be scheduled To-be-processed task 1 Computing resource object To-be-processed task 2 Storage resource object To-be-processed task 3 Network resource object To-be-processed task 4 Computing resource object, network resource object To-be-processed task 5 Computing resource object, storage resource object To-be-processed task 6 Computing resource object, storage resource object, network resource object

[0099] In this embodiment, determining at least one candidate resource node from the resource object set comprises: determining resource constraint information based on the resource scheduling requirements and / or the scheduling task content; and determining at least one candidate resource node from the resource object set based on the resource constraint information.

[0100] The resource scheduling constraint information includes hard scheduling constraint information and / or soft scheduling constraint information. The hard scheduling constraint information is a hard requirement for the resource object. For example, the hard scheduling constraint information refers to a resource requirement that must be met. If these constraints are not met, the task cannot be executed. The soft scheduling constraint information is a preferred requirement for the resource object. The soft scheduling constraint information refers to constraint information for further optimizing resource allocation on the premise of meeting the hard constraint. The soft scheduling constraint information can not be necessarily met, but meeting these constraint information can improve the accuracy of resource allocation and the performance of task execution.

[0101] Specifically, the resource scheduling requirements and / or scheduling task content can be analyzed to determine the resource constraint information. For example, the hard scheduling constraint information includes resource types (such as the task must be assigned to a specific type of resource object), resource performance indicators (such as the minimum number of CPU cores, the minimum computing power of GPU, the minimum read-write speed of storage, etc.), network requirements (such as the specific network access method required by the task), geographic location requirements (such as the task must be executed on a resource in a specific geographic location), resource availability (such as the resource object required by the task must be available within the requested time period and cannot conflict with other tasks). The soft scheduling constraint information includes resource performance optimization indicators (such as, on the basis of meeting the minimum performance requirements, preferentially selecting resources with higher performance), cost optimization indicators (such as, on the premise of meeting the task requirements, preferentially selecting resources with lower cost), load balancing indicators (such as, try to assign tasks to nodes with lower load to avoid overloading some nodes and improve the overall performance of the system), reliability optimization indicators (such as: preferentially selecting nodes with low failure rate and good maintenance). Further, the multiple candidate resource nodes that meet the hard scheduling constraint information can be filtered from the resource object set in combination with the resource constraint information. For example, filter out the nodes that support the required resource type, the performance indicators meet the minimum requirements, support the required network access method, are located in the required geographic location, and the resource objects are available within the requested time period, as candidate resource nodes. Further, if the soft scheduling constraint information is included in the resource constraint information, the candidate resource nodes can be filtered and soft constraint evaluation can be performed on the one or more candidate resource nodes that meet the soft scheduling constraint information to select the candidate resource nodes as the alternative resource nodes. For example, according to the performance indicators (such as CPU core number, GPU computing power, etc.) in the soft scheduling constraint information, the candidate resource nodes are sorted, and the candidate resource nodes with higher performance are preferentially selected as the alternative resource nodes. In the case of equivalent performance, the candidate resource nodes with lower cost can be preferentially selected as the alternative resource nodes.

[0102] Exemplarily, the resource scheduling requirement and / or the scheduling task content can be analyzed to determine the hard scheduling constraint information. The hard scheduling constraint information is matched with the object attributes of the computing resource objects and the storage resource objects in the resource object set, the resource objects that do not meet the hard scheduling constraint information are removed, and the resource objects that meet the hard scheduling constraint information are left as candidate resource nodes. For example, the object attributes of the computing resource objects include: the application name (or application ID) deployed on the computing resource object, the application version number (or version ID) deployed on the computing resource object, the specification model of the computing resource of the computing resource object (such as the CPU, GPU, NPU specification model), the number of remaining available computing resources of the computing resource object, the resource utilization of the computing resource object, the geographical related attributes of the computing resource object (such as the core computing resource object, the center computing resource object, and the edge computing resource object), the number of jobs queued on the computing resource object, the job queuing duration on the computing resource object, the energy consumption index (such as PUE and power consumption unit cost) of the data center where the computing resource object is located, and the cost (such as the unit cost of computing power usage and core hour / card hour unit cost). The object attributes of the storage resource objects include: the data content name (or content ID) stored, the data content version number (or version ID) stored, the storage type (object storage, block storage, and file storage), the storage location attribute (centralized and distributed), the number of available storage resources, and the storage resource utilization. The object attributes of the network access mode include: the network type, the network bandwidth utilization, the network remaining available bandwidth, the network transmission delay, the network packet loss rate, the network jitter, and the line cost.

[0103] It should be noted that the candidate resource nodes can be integrated or separated, and in the separated scenario, the computing resource objects and the storage resource objects that meet the hard scheduling constraint information are determined respectively. If the soft scheduling constraint information is not included in the resource constraint information, the candidate resource nodes can be used as the alternative resource nodes. If the soft scheduling constraint information is included in the resource constraint information, the candidate resource nodes are evaluated in different dimensions according to the soft scheduling constraint information, the evaluation results of each candidate resource node in each dimension are obtained, and the alternative resource nodes are selected from the candidate resource nodes according to the evaluation results. Optionally, the dimensions include but are not limited to the dimensions of computing power, storage, network, cost, energy consumption, security, region, and experience.

[0104] S230, each network access mode and each alternative resource node are combined to obtain a plurality of to-be-evaluated scheduling modes.

[0105] Continuing to refer to Figure 3The plurality of candidate resource nodes and the plurality of network access modes are combined to obtain a plurality of to-be-evaluated scheduling modes. For example, the to-be-evaluated scheduling mode includes various combinations of the determined computing resource object, the storage resource object, and the network access mode. The network access mode and the candidate resource node correspond to the resource object. That is, the candidate resource node and the network access mode in the to-be-evaluated scheduling mode are resource objects in the to-be-evaluated scheduling mode. After determining the plurality of to-be-evaluated scheduling modes, a third evaluation attribute corresponding to the to-be-evaluated scheduling mode is determined, and then the target scheduling mode is determined based on the third evaluation attribute.

[0106] For example, it is assumed that the storage resource object in the to-be-evaluated scheduling mode does not need to be scheduled, and only the computing resource object and the network access mode need to be scheduled. The resource objects in each to-be-evaluated scheduling mode can be referred to Table 2.

[0107] Table 2

[0108] Sequence number To-be-evaluated scheduling manner Computing resource object Network access manner 1 To-be-evaluated scheduling manner 1 Computing resource object 1 Network access manner 1 2 To-be-evaluated scheduling manner 2 Computing resource object 1 Network access manner 2 3 To-be-evaluated scheduling manner 3 Computing resource object 2 Network access manner 1 4 To-be-evaluated scheduling manner 4 Computing resource object 2 Network access manner 2 5 To-be-evaluated scheduling manner 5 Computing resource object 2 Network access manner 3

[0109] S240, current state data of each resource object in the to-be-evaluated scheduling mode is determined.

[0110] S250, for each to-be-evaluated scheduling mode, based on the current state data of each resource object in the to-be-evaluated scheduling mode, the preset tolerable state value corresponding to each evaluation dimension, and the preset importance coefficient, a first evaluation attribute corresponding to each evaluation dimension is determined, and based on the current state data of the resource object under the same evaluation dimension, a second evaluation attribute of the evaluation dimension is determined.

[0111] S260, based on the first evaluation attribute and the second evaluation attribute corresponding to each evaluation dimension in the to-be-evaluated scheduling mode, a third evaluation attribute corresponding to the to-be-evaluated scheduling mode is determined.

[0112] S270, based on the third evaluation attribute corresponding to each to-be-evaluated scheduling mode, a target scheduling mode is determined from the to-be-evaluated scheduling mode, so as to process a target task corresponding to the resource scheduling request based on the target scheduling mode.

[0113] The technical scheme of the embodiment improves the reliability of resource scheduling by determining the resource scheduling requirement and / or the scheduling task content when the resource scheduling request is received, and then determining at least one network access mode in combination with the resource scheduling requirement and / or the scheduling task content, and selecting at least one candidate resource node from the resource object set. Finally, the network access mode and the candidate resource node are combined to form a plurality of preliminary to-be-evaluated scheduling modes, which effectively improves the efficiency of resource scheduling and further ensures the accuracy of subsequent target scheduling mode determination.

[0114] Embodiment three

[0115] Figure 4 is a structural schematic diagram of a resource scheduling device according to Embodiment Three of the present application. As shown in the figure, the device comprises a to-be-evaluated scheduling mode determining module 310, a second evaluation attribute determining module 320, a third evaluation attribute determining module 330 and a target scheduling mode determining module 340. Figure 4

[0116] The to-be-evaluated scheduling mode determining module 310 is configured to, upon receiving a resource scheduling request, determine at least one to-be-evaluated scheduling mode corresponding to the resource scheduling request and acquire current state data of a plurality of resource objects under the at least one to-be-evaluated scheduling mode, wherein each resource object is associated with at least one evaluation dimension.

[0117] The second evaluation attribute determining module 320 is configured to, for each to-be-evaluated scheduling mode, determine a first evaluation attribute of the to-be-evaluated scheduling mode under each evaluation dimension based on current state data of a plurality of resource objects in the to-be-evaluated scheduling mode, preset tolerable state values of the resource objects under each evaluation dimension associated therewith and preset importance coefficients corresponding to each evaluation dimension, and determine a second evaluation attribute of the to-be-evaluated scheduling mode under the same evaluation dimension based on current state data of at least one resource object under the evaluation dimension.

[0118] The third evaluation attribute determining module 330 is configured to determine a third evaluation attribute corresponding to the to-be-evaluated scheduling mode based on at least one first evaluation attribute and second evaluation attribute corresponding to the to-be-evaluated scheduling mode.

[0119] The target scheduling mode determining module 340 is configured to determine a target scheduling mode from at least two to-be-evaluated scheduling modes based on each third evaluation attribute, so as to process a target task corresponding to the resource scheduling request based on the target scheduling mode.

[0120] ​The technical scheme of the embodiment comprises the following steps: when a resource scheduling request is received, at least one to-be-evaluated scheduling mode corresponding to the resource scheduling request is preliminarily determined, and current state data of a plurality of resource objects under the at least one to-be-evaluated scheduling mode is acquired. Then, based on the current state data of the plurality of resource objects under the to-be-evaluated scheduling mode, preset tolerable state values corresponding to the resource objects under each evaluation dimension associated with the resource objects, and preset importance coefficients corresponding to each evaluation dimension, a first evaluation attribute of the to-be-evaluated scheduling mode under each evaluation dimension is comprehensively evaluated, and based on the current state data of at least one resource object under the same evaluation dimension, a second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension is determined, so as to ensure that the evaluation attribute meets the real-time task processing requirement, to realize multi-dimensional and multi-level evaluation of the performance of the to-be-evaluated scheduling mode under each evaluation dimension, avoid one-sidedness of single-index evaluation, and improve evaluation accuracy. Further, by comprehensively determining at least one first evaluation attribute and second evaluation attribute corresponding to the to-be-evaluated scheduling mode, a third evaluation attribute corresponding to the to-be-evaluated scheduling mode is determined, and finally, based on each third evaluation attribute, a target scheduling mode is selected, so that the allocated target scheduling mode meets the real-time processing requirement of the task, and improves the reliability and efficiency of processing the task.

[0121] On the basis of the above device, optionally, the to-be-evaluated scheduling mode determination module 310 comprises:

[0122] The analysis unit is configured to determine resource scheduling requirements and / or scheduling task content corresponding to the resource scheduling request when the resource scheduling request is received.

[0123] The alternative resource node determination unit is configured to determine at least one network access mode based on the resource scheduling requirements and / or scheduling task content, and determine at least one alternative resource node from the resource object set.

[0124] The to-be-evaluated scheduling mode determination unit is configured to combine each network access mode and each alternative resource node to obtain a plurality of to-be-evaluated scheduling modes; wherein the network access mode and the alternative resource node correspond to the resource object.

[0125] On the basis of the above device, optionally, the alternative resource node determination unit comprises:

[0126] The resource constraint information determination subunit is configured to determine resource constraint information based on the resource scheduling requirements and / or scheduling task content; wherein the resource scheduling constraint information comprises hard scheduling constraint information and / or soft scheduling constraint information; the hard scheduling constraint information is a hard requirement for the resource object; and the soft scheduling constraint information is a preferential requirement for the resource object.

[0127] The alternative resource node determination subunit is configured to determine at least one alternative resource node from the resource object set based on the resource constraint information.

[0128] On the basis of the above device, the second evaluation attribute determination module 320 comprises:

[0129] The fourth evaluation attribute determination unit is configured to determine a fourth evaluation attribute of the evaluation dimension based on the current state data of at least one resource object under the same evaluation dimension, the preset tolerable state value corresponding to the resource object under the evaluation dimension, and the preset importance coefficient of the evaluation dimension; wherein the evaluation dimension comprises at least two.

[0130] The first evaluation attribute determination unit is configured to determine a first evaluation attribute of the to-be-evaluated scheduling mode under a current evaluation dimension based on the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attributes of other evaluation dimensions different from the current evaluation dimension.

[0131] On the basis of the above device, the fourth evaluation attribute determination unit comprises:

[0132] The first state performance attribute determination unit is configured to determine a first state performance attribute corresponding to each resource object under the evaluation dimension based on the quotient value between the current state data of the resource object and the preset tolerable state value corresponding to the resource object under the evaluation dimension when the evaluation dimension is a positive index dimension.

[0133] The fourth evaluation attribute determination first unit is configured to determine the fourth evaluation attribute of the evaluation dimension based on the first state performance attribute of each resource object and the preset importance coefficient corresponding to the evaluation dimension.

[0134] On the basis of the above device, the fourth evaluation attribute determination unit comprises:

[0135] The second state performance attribute determination unit is configured to determine a second state performance attribute corresponding to each resource object under the evaluation dimension based on the quotient value between the current state data of the resource object and the preset tolerable state value corresponding to the resource object under the evaluation dimension and a preset parameter value when the evaluation dimension is a negative index dimension.

[0136] The fourth evaluation attribute determination second unit is configured to determine the fourth evaluation attribute of the evaluation dimension based on the second state performance attribute of each resource object and the preset importance coefficient corresponding to the evaluation dimension.

[0137] On the basis of the above-mentioned device, optionally, the first evaluation attribute determination unit is specifically configured to, in the case that the current evaluation dimension is a positive index dimension, determine a first intermediate value based on a fourth evaluation attribute of the current evaluation dimension and a fourth evaluation attribute of other positive index dimensions that are different from the current evaluation dimension, and determine the first evaluation attribute of the to-be-evaluated scheduling mode under the current evaluation dimension based on the fourth evaluation attribute of the current evaluation dimension and the first intermediate value; in the case that the current evaluation dimension is a negative index dimension, determine a second intermediate value based on a fourth evaluation attribute of the current evaluation dimension and a fourth evaluation attribute of other negative index dimensions that are different from the current evaluation dimension, and determine the first evaluation attribute of the to-be-evaluated scheduling mode under the current evaluation dimension based on the fourth evaluation attribute of the current evaluation dimension and the second intermediate value.

[0138] On the basis of the above-mentioned device, optionally, the second evaluation attribute determination module 320 comprises:

[0139] The second evaluation attribute determination unit is configured to evaluate the current state data of at least one resource object under the same evaluation dimension based on preset conditions to obtain the second evaluation attribute of the to-be-evaluated scheduling mode under the evaluation dimension; wherein the preset conditions at least include an algorithm performance evaluation condition, a storage performance evaluation condition, a network performance evaluation condition, a security performance evaluation condition, an energy consumption evaluation condition and a cost evaluation condition.

[0140] On the basis of the above-mentioned device, optionally, the third evaluation attribute determination module 330 comprises:

[0141] The fifth evaluation attribute determination unit is configured to determine the fifth evaluation attribute corresponding to the evaluation dimension based on the first evaluation attribute and the second evaluation attribute corresponding to the same evaluation dimension under the to-be-evaluated scheduling mode.

[0142] The third evaluation attribute determination unit is configured to determine the third evaluation attribute corresponding to the to-be-evaluated scheduling mode based on the fifth evaluation attribute corresponding to each evaluation dimension.

[0143] The resource scheduling device provided in the embodiments of the present application can execute the resource scheduling method provided in any embodiment of the present application, and has the corresponding function modules and beneficial effects of the execution method.

[0144] Embodiment four

[0145] Figure 5Schematic diagram of the structure of an electronic device that implements the resource scheduling method of an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0146] like Figure 5 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory 12 and a random access memory 13, that is communicatively connected to the at least one processor 11. The memory stores a computer program that can be executed by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory 12 or the computer program loaded from the storage unit 18 into the random access memory 13. The random access memory 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, the read-only memory 12, and the random access memory 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0147] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0148] The processor 11 may be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the resource scheduling method.

[0149] In some embodiments, the resource scheduling method can be implemented as a computer program tangibly embodied in a computer readable storage medium, e.g., storage unit 18. In some embodiments, parts or all of the computer program can be loaded and / or installed onto electronic device 10 via read-only memory 12 and / or communication unit 19. When the computer program is loaded onto random access memory 13 and executed by processor 11, one or more steps of the above-described resource scheduling method can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the resource scheduling method by way of other means, e.g., by way of firmware.

[0150] Various implementations of the systems and techniques described above can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a programmable logic device (PLD), a computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0151] Computer programs used to implement the methods of the application can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the computer program, when executed by the processor of the machine, implements the functions / acts specified in the flowcharts and / or block diagrams. The computer program can be executed entirely on a machine, partially on a machine, partially on a machine as part of a standalone software package, or entirely on a remote machine or server.

[0152] In the context of the present invention, computer-readable storage medium can be a tangible medium that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage medium can include but is not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage medium can be a machine-readable signal medium. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device or any suitable combination of the foregoing.

[0153] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0154] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0155] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. Servers can be cloud servers, also known as cloud computing servers or cloud hosts, which are a host product in the cloud computing service system to solve the defects of large management difficulty and weak business scalability in traditional physical hosts and VPS services.

[0156] In particular, the processes described above with reference to the flowcharts can be implemented as a computer software program in accordance with embodiments of the present application. For example, embodiments of the present application include a computer program product comprising a computer program carried on a non-transitory computer readable medium, the computer program comprising program code for executing the methods illustrated by the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via the communication unit 19, or installed from the storage unit 18, or installed from the read only memory 12. When the computer program is executed by the processor 11, the above-mentioned functions defined in the methods of the embodiments of the present application are performed.

[0157] Embodiments of the present application also provide a computer program product comprising a computer program which, when executed by a processor, implements the resource scheduling method provided by any of the embodiments of the present application.

[0158] In the implementation of the computer program product, the computer program code for performing the operations of the present application can be written in one or more programming languages or combinations of languages including object-oriented programming languages such as Java, Smalltalk, C++, conventional procedural programming languages such as the "C" programming language, or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, through the Internet using an Internet service provider).

[0159] It should be understood that the various forms of flow shown above can be reordered, added to, or deleted from, for example. The steps described in the present application can be executed in parallel, in sequence, or in a different order, as long as the desired results of the technical solutions of the present application can be achieved, and the present application is not limited herein.

[0160] The above detailed description does not limit the scope of the application. Various modifications, combinations, sub-combinations and alternatives can be made to the detailed embodiment within the scope of the application. Any modification, equivalent replacement and improvement made without departing from the spirit and principle of the application shall fall within the scope of the application.

Claims

1. A resource scheduling method, characterized in that: include: Upon receiving a resource scheduling request, determining at least one scheduling mode to be evaluated corresponding to the resource scheduling request, and obtaining current state data of a plurality of resource objects under the at least one scheduling mode to be evaluated; wherein each resource object is associated with at least one evaluation dimension; For each of the scheduling modes to be evaluated, based on the current state data of multiple resource objects in the scheduling mode to be evaluated, the preset tolerable state values ​​corresponding to the resource objects under each of the evaluation dimensions associated therewith, and the preset importance coefficients corresponding to each of the evaluation dimensions, determine the first evaluation attribute of the scheduling mode to be evaluated under each of the evaluation dimensions, and based on the current state data of at least one of the resource objects under the same evaluation dimension, determine the second evaluation attribute of the scheduling mode to be evaluated under the evaluation dimension; determining a third evaluation attribute corresponding to the scheduling mode to be evaluated based on at least one of the first evaluation attribute and the second evaluation attribute corresponding to the scheduling mode to be evaluated; Based on each of the third evaluation attributes, a target scheduling mode is determined from at least one of the scheduling modes to be evaluated, so as to process a target task corresponding to the resource scheduling request based on the target scheduling mode.

2. The method according to claim 1, characterized in that The step of determining, upon receiving a resource scheduling request, at least one scheduling mode to be evaluated corresponding to the resource scheduling request includes: Upon receiving a resource scheduling request, determining resource scheduling requirements and / or scheduling task content corresponding to the resource scheduling request; Based on the resource scheduling requirements and / or scheduling task content, determining at least one network access mode, and determining at least one candidate resource node from the resource object set; Each of the network access modes and each of the candidate resource nodes are combined to obtain a plurality of scheduling modes to be evaluated; wherein both the network access modes and the candidate resource nodes correspond to the resource objects.

3. The method according to claim 2, characterized in that The determining of at least one candidate resource node from the resource object set includes: Determining resource constraint information based on the resource scheduling requirements and / or scheduling task content; wherein the resource scheduling constraint information includes hard scheduling constraint information and / or soft scheduling constraint information; the hard scheduling constraint information is a hard requirement for the resource object; the soft scheduling constraint information is a preferential requirement for the resource object; At least one candidate resource node is determined from the resource object set based on the resource constraint information.

4. The method according to claim 1, wherein The determining, based on current status data of multiple resource objects in the scheduling mode to be evaluated, preset tolerable status values ​​corresponding to the resource objects under each of the evaluation dimensions associated therewith, and preset importance coefficients corresponding to each of the evaluation dimensions, of the first evaluation attribute of the scheduling mode to be evaluated under each of the evaluation dimensions includes: Determining a fourth evaluation attribute of the evaluation dimension based on current status data of at least one resource object under the same evaluation dimension, a preset tolerable status value corresponding to the resource object under the evaluation dimension, and a preset importance coefficient of the evaluation dimension; wherein the evaluation dimension includes at least two; For each of the evaluation dimensions, based on the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attributes of other evaluation dimensions different from the current evaluation dimension, the first evaluation attribute of the scheduling mode to be evaluated under the current evaluation dimension is determined.

5. The method according to claim 4, characterized in that The determining of the fourth evaluation attribute of the evaluation dimension based on current status data of at least one resource object under the same evaluation dimension, a preset tolerable status value corresponding to the resource object under the evaluation dimension, and a preset importance coefficient of the evaluation dimension includes: In a case where the evaluation dimension is a positive indicator dimension, for each resource object under the evaluation dimension, determining a first state performance attribute corresponding to the resource object based on a quotient between current state data of the resource object and a preset tolerable state value corresponding to the resource object under the evaluation dimension; Based on the first state performance attribute of each resource object and the preset importance coefficient corresponding to the evaluation dimension, a fourth evaluation attribute of the evaluation dimension is determined.

6. The method according to claim 4, characterized in that The determining of the fourth evaluation attribute of the evaluation dimension based on current status data of at least one resource object under the same evaluation dimension, a preset tolerable status value corresponding to the resource object under the evaluation dimension, and a preset importance coefficient of the evaluation dimension includes: In the case where the evaluation dimension is a negative indicator dimension, for each resource object under the evaluation dimension, determining a second state performance attribute corresponding to the resource object based on a quotient between the current state data of the resource object and a preset tolerable state value corresponding to the resource object under the evaluation dimension and a preset parameter value; Based on the second state performance attribute of each resource object and the preset importance coefficient corresponding to the evaluation dimension, a fourth evaluation attribute of the evaluation dimension is determined.

7. The method according to claim 4, characterized in that The determining, based on the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attributes of other evaluation dimensions different from the current evaluation dimension, the first evaluation attribute of the scheduling mode to be evaluated under the current evaluation dimension includes: In a case where the current evaluation dimension is a positive indicator dimension, determining a first intermediate value based on a fourth evaluation attribute of the current evaluation dimension and fourth evaluation attributes of other positive indicator dimensions different from the current evaluation dimension, and determining a first evaluation attribute of the scheduling mode to be evaluated under the current evaluation dimension based on the fourth evaluation attribute of the current evaluation dimension and the first intermediate value; In the case where the current evaluation dimension is a negative indicator dimension, a second intermediate value is determined based on the fourth evaluation attribute of the current evaluation dimension and the fourth evaluation attributes of other negative indicator dimensions different from the current evaluation dimension, and based on the fourth evaluation attribute of the current evaluation dimension and the second intermediate value, the first evaluation attribute of the scheduling method to be evaluated under the current evaluation dimension is determined.

8. The method according to claim 1, characterized in that The determining, based on the current state data of at least one resource object under the same evaluation dimension, a second evaluation attribute of the scheduling mode to be evaluated under the evaluation dimension includes: Based on a preset condition, current state data of at least one resource object under the same evaluation dimension is evaluated to obtain a second evaluation attribute of the scheduling mode to be evaluated under the evaluation dimension; Among them, the preset conditions include at least computing power performance evaluation conditions, storage performance evaluation conditions, network performance evaluation conditions, security performance evaluation conditions, energy consumption evaluation conditions and cost evaluation conditions.

9. The method according to claim 1, characterized in that The determining, based on at least one of the first evaluation attribute and the second evaluation attribute corresponding to the scheduling mode to be evaluated, a third evaluation attribute corresponding to the scheduling mode to be evaluated includes: Determine a fifth evaluation attribute corresponding to the evaluation dimension based on the first evaluation attribute and the second evaluation attribute corresponding to the same evaluation dimension under the scheduling mode to be evaluated; Based on the fifth evaluation attribute corresponding to each of the evaluation dimensions, a third evaluation attribute corresponding to the scheduling mode to be evaluated is determined.

10. A resource scheduling device, characterized in that: include: a scheduling mode determination module to be evaluated, configured to, upon receiving a resource scheduling request, determine at least one scheduling mode to be evaluated corresponding to the resource scheduling request, and obtain current state data of a plurality of resource objects under the at least one scheduling mode to be evaluated; wherein each resource object is associated with at least one evaluation dimension; a second evaluation attribute determination module for determining, for each of the scheduling modes to be evaluated, a first evaluation attribute of the scheduling mode to be evaluated under each of the evaluation dimensions based on current status data of multiple resource objects in the scheduling mode to be evaluated, preset tolerable status values ​​corresponding to the resource objects under each of the evaluation dimensions associated therewith, and a preset importance coefficient corresponding to each of the evaluation dimensions, and determining a second evaluation attribute of the scheduling mode to be evaluated under the evaluation dimension based on the current status data of at least one of the resource objects under the same evaluation dimension; a third evaluation attribute determining module, configured to determine a third evaluation attribute corresponding to the scheduling mode to be evaluated based on at least one of the first evaluation attribute and the second evaluation attribute corresponding to the scheduling mode to be evaluated; The target scheduling mode determination module is configured to determine a target scheduling mode from at least two scheduling modes to be evaluated based on each of the third evaluation attributes, so as to process a target task corresponding to the resource scheduling request based on the target scheduling mode.

Images