阅读说明：本技术 服务等级协议映射方法、电子设备、及存储介质 (Service level agreement mapping method, electronic device, and storage medium ) 是由 李静 李福昌 董秋丽 曹亘 于 2021-07-19 设计创作，主要内容包括：本申请提供一种服务等级协议映射方法、电子设备、及存储介质,涉及通信技术领域。该方法包括：获取服务等级协议SLA参数对应的服务质量QoS参数集和SLA评价等级集；QoS参数集合包括一个或多个QoS参数；SLA评价等级集包括至少两个评价等级；根据QoS参数集和SLA评价等级集,建立评价矩阵；根据评价矩阵和QoS参数集中的每个QoS参数的权重,确定QoS参数集对SLA评价等级集中的每个评价等级的隶属度；确定QoS参数集对SLA评价等级集中的每个评价等级的隶属度中,最大隶属度对应的评价等级为QoS参数集对应的评价等级。该方法中提供的SLA参数与QoS参数之间的映射方式,可以适用于不同的业务服务、不同的承载网络类型,具有普适性。(The application provides a service level protocol mapping method, electronic equipment and a storage medium, and relates to the technical field of communication. The method comprises the following steps: acquiring a service quality QoS parameter set and an SLA evaluation level set corresponding to service level agreement SLA parameters; the set of QoS parameters includes one or more QoS parameters; the SLA rating level set comprises at least two rating levels; establishing an evaluation matrix according to the QoS parameter set and the SLA evaluation grade set; determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the evaluation matrix and the weight of each QoS parameter in the QoS parameter set; and determining the membership degree of the QoS parameter set to each evaluation grade in the SLA evaluation grade set, wherein the evaluation grade corresponding to the maximum membership degree is the evaluation grade corresponding to the QoS parameter set. The mapping mode between the SLA parameters and the QoS parameters provided by the method can be suitable for different service services and different bearer network types, and has universality.)

1. A service level agreement mapping method, the method comprising:

acquiring a service quality QoS parameter set and an SLA evaluation level set corresponding to service level agreement SLA parameters; the set of QoS parameters comprises one or more QoS parameters; the SLA rating level set comprises at least two rating levels;

establishing an evaluation matrix according to the QoS parameter set and the SLA evaluation level set;

determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the evaluation matrix and the weight of each QoS parameter in the QoS parameter set;

and determining the membership degree of the QoS parameter set to each evaluation grade in the SLA evaluation grade set, wherein the evaluation grade corresponding to the maximum membership degree is the evaluation grade corresponding to the QoS parameter set.

2. The method of claim 1, wherein determining the degree of membership of the set of QoS parameters to each rating level in the set of SLA rating levels according to the rating matrix and the weight of each QoS parameter in the set of QoS parameters comprises:

acquiring the weight of each QoS parameter in the QoS parameter set;

generating a weight matrix according to the weight of each QoS parameter in the QoS parameter set;

and determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the weight matrix and the evaluation matrix.

3. The method of claim 2, wherein determining the degree of membership of the set of QoS parameters to each rating level in the set of SLA rating levels according to the weight matrix and the rating matrix comprises:

acquiring a fuzzy synthesis relation calculation result of the weight matrix and the evaluation matrix;

and determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the weight matrix and the fuzzy synthetic relation calculation result of the evaluation matrix.

4. The method of claim 1, wherein establishing an evaluation matrix based on the set of QoS parameters and the set of SLA evaluation levels comprises:

acquiring the membership degree of each QoS parameter in the QoS parameter set to each evaluation level in the SLA evaluation level set;

determining the normalization result of the membership degree of each QoS parameter to each evaluation grade according to the membership degree of each QoS parameter to each evaluation grade in the SLA evaluation grade set;

and establishing an evaluation matrix according to the normalization result of the membership of each QoS parameter to each evaluation grade.

5. The method of claim 4, wherein obtaining the degree of membership of each QoS parameter in the set of QoS parameters to each rating level in the set of SLA rating levels comprises:

acquiring the number of each QoS parameter which meets each evaluation grade;

and obtaining the membership degree of each QoS parameter to each evaluation grade in the SLA evaluation grade set according to the number of the QoS parameters meeting each evaluation grade.

6. The method according to any of claims 1-5, wherein the QoS parameters include latency, throughput, and packet loss rate.

7. The method of any of claims 1-5, wherein said SLA evaluation levels comprise level 1, level 2, level 3, level 4, and level 5.

8. The method of claim 2, wherein the obtaining the weight of each QoS parameter in the set of QoS parameters comprises:

determining the information gain of each QoS parameter in the QoS parameter set according to a limit gradient lifting model;

determining the value of the importance degree of each QoS parameter in the QoS parameter set according to the information gain of each QoS parameter in the QoS parameter set;

and calculating the importance ratio of each QoS parameter according to the value of the importance degree of each QoS parameter in the QoS parameter set to obtain the weight of each QoS parameter.

9. An electronic device, comprising: a processor and a memory;

the memory stores instructions executable by the processor;

the processor is configured to, when executing the instructions, cause the electronic device to implement the method of any of claims 1-8.

10. A computer readable storage medium having stored thereon computer program instructions; computer program instructions which, when executed by an electronic device, cause the electronic device to carry out the method of any one of claims 1 to 8.

Technical Field

The present application relates to the field of communications technologies, and in particular, to a service level agreement mapping method, an electronic device, and a storage medium.

Background

An SLA (service level agreement) is an agreement between an enterprise or an individual (which may be referred to as a service provider) providing a service and a user (which may be referred to as a service user) using the service, in terms of quality of service, priority, obligation for responsibility, and the like. Different SLA parameters are determined by service provider and service consumer negotiations in SLAs for different services.

However, in the SLA parameters, many values of the SLA parameters cannot be directly obtained, and the SLA parameters need to be mapped to other quality of service (QoS) parameters for comprehensive analysis and calculation.

Currently, the mapping manner for mapping the SLA parameter to the QoS parameter is generally defined by human. For different service services and different bearer network types, the mapping manner of mapping the SLA parameters to the QoS parameters may be different.

Disclosure of Invention

The embodiment of the application provides a service level agreement mapping method, an electronic device and a storage medium, which can realize mapping between SLA parameters and QoS parameters, and the mapping mode can be suitable for different service services and different bearer network types, and has universality.

In a first aspect, an embodiment of the present application provides a service level agreement mapping method, where the method includes:

acquiring a service quality QoS parameter set and an SLA evaluation level set corresponding to service level agreement SLA parameters; the set of QoS parameters includes one or more QoS parameters; the SLA rating level set comprises at least two rating levels; establishing an evaluation matrix according to the QoS parameter set and the SLA evaluation grade set; determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the evaluation matrix and the weight of each QoS parameter in the QoS parameter set; and determining the membership degree of the QoS parameter set to each evaluation grade in the SLA evaluation grade set, wherein the evaluation grade corresponding to the maximum membership degree is the evaluation grade corresponding to the QoS parameter set.

In the method, the SLA parameter can be divided into a plurality of (e.g., at least two) service levels (or referred to as evaluation levels), and a mapping manner between the SLA parameter and the QoS parameter corresponding to the SLA parameter is provided by means of machine learning and mathematical computation. The mapping mode between the SLA parameters and the QoS parameters provided in the service level agreement mapping method can be suitable for different service services and different bearer network types, and has universality. In addition, the mapping between the SLA parameters and the QoS parameters in the service level agreement mapping method is more objective.

In one possible design, determining a degree of membership of the set of QoS parameters to each rating level in the set of SLA rating levels based on the rating matrix and a weight of each QoS parameter in the set of QoS parameters comprises: acquiring the weight of each QoS parameter in a QoS parameter set; and generating a weight matrix according to the weight of each QoS parameter in the QoS parameter set. And determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the weight matrix and the evaluation matrix.

In another possible design, determining the membership of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the weight matrix and the evaluation matrix includes: acquiring a fuzzy synthetic relation calculation result of the weight matrix and the evaluation matrix; and determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the fuzzy synthetic relation calculation result of the weight matrix and the evaluation matrix.

In yet another possible design, an evaluation matrix is established based on the set of QoS parameters and the set of SLA evaluation levels, including: acquiring the membership degree of each QoS parameter in the QoS parameter set to each evaluation level in the SLA evaluation level set; determining the normalization result of the membership degree of each QoS parameter to each evaluation grade according to the membership degree of each QoS parameter to each evaluation grade in the SLA evaluation grade set; and establishing an evaluation matrix according to the normalization result of the membership of each QoS parameter to each evaluation grade.

In yet another possible design, the degree of membership of each QoS parameter to each rating level in the set of SLA rating levels includes: acquiring the number of QoS parameters meeting each evaluation grade; and obtaining the membership degree of each QoS parameter to each evaluation grade in the SLA evaluation grade set according to the number of the QoS parameters meeting each evaluation grade.

In yet another possible design, the QoS parameters may include latency, throughput, and packet loss rate.

In yet another possible design, the SLA evaluation levels include level 1, level 2, level 3, level 4, and level 5.

Optionally, the obtaining the weight of each QoS parameter in the QoS parameter set includes: determining the information gain of each QoS parameter in the QoS parameter set according to a limit gradient lifting model; determining the value of the importance degree of each QoS parameter in the QoS parameter set according to the information gain of each QoS parameter in the QoS parameter set; and calculating the importance ratio of each QoS parameter according to the value of the importance degree of each QoS parameter in the QoS parameter set to obtain the weight of each QoS parameter.

In a second aspect, an embodiment of the present application provides a service level agreement mapping apparatus, including: the device comprises an acquisition module and a processing module.

The acquisition module is used for acquiring a service quality QoS parameter set and an SLA evaluation level set corresponding to the SLA parameters of the service level agreement; the processing module is used for establishing an evaluation matrix according to the QoS parameter set and the SLA evaluation level set, determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the evaluation matrix and the weight of each QoS parameter in the QoS parameter set, determining the maximum membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set, and determining the evaluation level corresponding to the maximum membership degree as the evaluation level corresponding to the QoS parameter set.

In one possible design, the obtaining module is further configured to obtain a weight of each QoS parameter in the QoS parameter set; and the processing module is also used for generating a weight matrix according to the weight of each QoS parameter in the QoS parameter set, and determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the weight matrix and the evaluation matrix.

In another possible design, the obtaining module is further configured to obtain a fuzzy synthetic relationship calculation result of the weight matrix and the evaluation matrix; and the processing module is also used for determining the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the fuzzy synthetic relation calculation result of the weight matrix and the evaluation matrix.

In yet another possible design, the obtaining module is further configured to obtain a membership degree of each QoS parameter in the QoS parameter set to each evaluation level in the SLA evaluation level set; the processing module is further used for determining the normalization result of the membership degree of each QoS parameter to each evaluation level according to the membership degree of each QoS parameter to each evaluation level in the SLA evaluation level set, and establishing an evaluation matrix according to the normalization result of the membership degree of each QoS parameter to each evaluation level.

In another possible design, the obtaining module is further configured to obtain the number of QoS parameters that satisfy each evaluation level; and the processing module is also used for obtaining the membership degree of each QoS parameter to each evaluation level in the SLA evaluation level set according to the number of the QoS parameters meeting each evaluation level.

Optionally, the obtaining module is specifically configured to determine an information gain of each QoS parameter in the QoS parameter set according to a limit gradient lifting model; determining the value of the importance degree of each QoS parameter in the QoS parameter set according to the information gain of each QoS parameter in the QoS parameter set; and calculating the importance ratio of each QoS parameter according to the value of the importance degree of each QoS parameter in the QoS parameter set to obtain the weight of each QoS parameter.

In a third aspect, an embodiment of the present application provides an electronic device, including: a processor and a memory; the memory stores instructions executable by the processor; the processor is configured to execute the instructions, such that the electronic device implements the method according to the first aspect.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium having computer program instructions stored thereon; the computer program instructions, when executed by an electronic device, cause the electronic device to implement a method as described in the first aspect.

The beneficial effects of the second to fourth aspects can be referred to the description of the first aspect, and are not repeated.

Drawings

Fig. 1 is a schematic diagram of a mapping relationship between SLA parameters and QoS parameters of a network slice;

fig. 2 is a schematic flowchart of a service level agreement mapping method according to an embodiment of the present application;

fig. 3 is another schematic flow chart of a service level agreement mapping method according to an embodiment of the present application;

fig. 4 is a schematic flowchart of a service level agreement mapping method according to an embodiment of the present application;

fig. 5 is a schematic structural diagram of a service level agreement mapping apparatus according to an embodiment of the present application;

fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application. The terms "first", "second", and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

It is noted that the words "exemplary," "for example," and "such as" are used herein to mean serving as an example, instance, or illustration. Any embodiment or design described herein as "exemplary" or "e.g.," is not necessarily to be construed as preferred or advantageous over other embodiments or designs. Rather, use of the word "exemplary" or "such as" is intended to present concepts related in a concrete fashion.

An SLA (service level agreement) is an agreement between an enterprise or an individual (which may be referred to as a service provider) providing a service and a user (which may be referred to as a service user) using the service, in terms of quality of service, priority, obligation for responsibility, and the like. Different SLA parameters are determined by service provider and service consumer negotiations in SLAs for different services.

Taking the network slicing service as an example, in the network slicing SLA, the SLA parameters determined by the service provider and the service consumer through negotiation may include: accessibility, availability, utilization, maintainability, mobility, integrity, etc.

However, in the above SLA parameters, many values of the SLA parameters cannot be directly obtained, and the SLA parameters need to be mapped to other quality of service (QoS) parameters for comprehensive analysis and calculation. Other QoS parameters may be network delay, network packet loss rate, etc.

Currently, the mapping manner (or referred to as mapping relation, mapping mechanism, etc.) for mapping the SLA parameter to the QoS parameter is generally defined manually. For different service services and different bearer network types, the mapping manner of mapping the SLA parameters to the QoS parameters may be different.

For example, fig. 1 is a schematic diagram illustrating a mapping relationship between network slice SLA parameters and QoS parameters. As shown in fig. 1, in the network slice SLA, the SLA parameters determined by the service provider and the service consumer through negotiation may include: SLA parameters for accessibility, availability, utilization, maintainability, mobility, and integrity, etc. Wherein, the accessibility may be mapped to QoS parameters such as a network slice registered through an access and mobility management function (AMF), a single network slice registration success rate, and accessibility of a Data Radio Bearer (DRB) serving a User Equipment (UE). The utilization may be mapped to QoS parameters such as an average number of network slice Protocol Data Unit (PDU) sessions and network slice virtualization resource utilization. Sustainability can be mapped to QoS parameters such as sustainability of QoS flows. Mobility may be mapped to QoS parameters such as a next generation-radio access network (NG-RAN) handover success rate. Integrity may be mapped to QoS parameters for upstream and downstream throughput of network slices, upstream and downstream throughput of N3 interfaces, throughput of RAN UEs, and so on

In this background, an embodiment of the present application provides a service level agreement mapping method, which may divide an SLA parameter into a plurality of (e.g., at least two) service levels (or referred to as evaluation levels), and provide a mapping manner between the SLA parameter and a QoS parameter corresponding to the SLA parameter through means of machine learning and mathematical computation. The mapping mode between the SLA parameters and the QoS parameters provided in the service level agreement mapping method can be suitable for different service services and different bearer network types, and has universality. In addition, the mapping between the SLA parameters and the QoS parameters in the service level agreement mapping method is more objective.

In the embodiment of the present application, the QoS parameter may also be referred to as an evaluation index, and a plurality of (e.g., at least two) QoS parameters may form an evaluation index combination.

The service level agreement mapping method provided by the embodiment of the present application is exemplarily described below.

The executing body of the service level agreement mapping method provided in the embodiment of the present application may be a server, a computer, or other electronic devices with data processing capability, or a network element in a communication network, which is not limited herein.

Fig. 2 is a flowchart illustrating a service level agreement mapping method according to an embodiment of the present application. As shown in fig. 2, the method may include S201-S204.

S201, an SLA evaluation index set and an SLA evaluation grade set corresponding to the SLA parameters are obtained.

Wherein, the SLA evaluation index set may include at least one evaluation index (i.e., QoS parameter), and the SLA evaluation level set may include a plurality of evaluation levels.

Alternatively, the set of SLA evaluation metrics and the set of SLA evaluation levels may be derived from an entered SLA. The signed SLA comprises an SLA evaluation index set and an SLA evaluation grade set corresponding to the SLA parameters. Which evaluation indexes are included in the SLA evaluation index set corresponding to the SLA parameters and several evaluation levels are included in the SLA evaluation level set, which can be determined manually (e.g., by an administrator).

As described above, the evaluation indexes in the SLA evaluation index set may be QoS parameters. For example, the QoS parameters may include delay, throughput, packet loss rate, and the like, and the type and number of the QoS parameters included in the SLA evaluation index set are not limited herein.

Illustratively, taking a certain SLA parameter 1 in the network slicing service as an example, in one possible design, according to a signed network slicing SLA, an SLA evaluation index set and an SLA evaluation level set corresponding to the obtained SLA parameter 1 may be as shown in table 1 below.

TABLE 1

As shown in table 1, the evaluation indexes included in the SLA evaluation index set corresponding to SLA parameter 1 may include: delay (delay), throughput (throughput), and loss rate (loss). The evaluation level set corresponding to the SLA parameter 1 divides the evaluation levels into five evaluation levels, namely level 1, level 2, level 3, level 4 and level 5.

Wherein, the evaluation index combination corresponding to the grade 1 comprises: time delay of less than "<"means less than) 1 millisecond (ms), a throughput of 1 gigabit per second (Gbps), and a packet loss rate of less than 10^-5(ii) a The evaluation index combinations corresponding to the level 2 include: time delay less than 5ms, throughput of 1Gbps, and packet loss rate less than 10^-5(ii) a The evaluation index combinations corresponding to the level 3 include: delay less than 10ms, throughput of 500 megabits per second (Mbps), and packet loss rate less than 10^-5(ii) a The evaluation index combination corresponding to the rank 4 includes: the time delay is less than 20ms, the throughput is 300Mbps, and the packet loss rate is less than 10^-4(ii) a The evaluation index combination corresponding to the rank 5 includes: the time delay is less than 25ms, the throughput is 150Mbps, and the packet loss rate is less than 10^-3。

In this embodiment of the application, the SLA evaluation index set corresponding to the SLA parameter acquired in S201 may be shown as V below, and the SLA evaluation level set may be shown as R below.

V＝{v₁，v₂，...，v_n}；

R＝{r₁，r₂，...，r_m}。

Wherein v is₁To v_nThe evaluation indexes are the centralized evaluation indexes of the SLA evaluation indexes; r is₁To r_mThe evaluation levels in the SLA evaluation level set. n is an integer greater than 0 and m is an integer greater than 1.

With reference to the example shown in table 1, the SLA evaluation index set corresponding to the SLA parameter 1 acquired in S201 may be shown as V below, and the SLA evaluation level set may be shown as R below.

V＝{delay，throughput，loss}；

R ═ level 1, level 2, level 3, level 4, level 5.

It should be noted that the SLA evaluation index set and the SLA evaluation level set shown in table 1 above are only exemplary. The method and the device have the advantages that the number and the type of the evaluation indexes included in the SLA evaluation index set corresponding to the SLA parameters and the number of the evaluation levels included in the SLA evaluation level set are not limited.

S202, establishing an evaluation matrix according to the SLA evaluation index set and the SLA evaluation grade set.

Fig. 3 is another flowchart of a service level agreement mapping method according to an embodiment of the present disclosure. As shown in fig. 3, the S202 may include:

s301, obtaining the membership degree of each evaluation index in the SLA evaluation index set to each evaluation level in the SLA evaluation level set.

Illustratively, the degree of membership of each evaluation index in the set of SLA evaluation indexes to each evaluation level in the set of SLA evaluation levels refers to: the number of the evaluation indexes which can satisfy the requirement of each evaluation grade.

In the examples of the present application, evaluation index v₁The degree of membership to each rating level in the SLA rating level set can be expressed as

Wherein v is₁₁Indicates satisfaction of v in level 1 in the actual evaluation index₁The number of required indexes; v. of₁₂Indicating satisfaction of v in level 2 in the actual evaluation index₁The number of required indexes; v. of_1mIndicating a satisfaction level in the actual evaluation indexm middle v₁The number of required indexes.

For example, taking the evaluation index delay described in table 1 above as an example, the membership degree of each evaluation level in the SLA evaluation level set shown in table 1 by delay may be expressed as delay^{Degree of membership of each grade}。

delay^{Degree of membership of each grade}＝(delay₁，delay₂，delay₃，delay₄，delay₅)

Wherein, delay₁The number of indexes meeting the requirement of grade 1delay in the actual evaluation indexes is represented; delay₂The number of indexes meeting the requirement of grade 2delay in the actual evaluation indexes is represented; delay₃The number of indexes meeting the requirement of grade 3delay in the actual evaluation indexes is represented; delay₄The number of indexes meeting the requirement of 4delay in the actual evaluation indexes is represented; delay₅The number of indexes satisfying the level 5delay requirement among the actual evaluation indexes is shown.

S302, determining the normalization result of the membership degree of each evaluation index to each evaluation grade according to the membership degree of each evaluation index to each evaluation grade.

In the examples of the present application, evaluation index v₁The normalization result of the membership degree of each evaluation level in the SLA evaluation level set can be calculated by the following formula (1).

In the formula (1), v_1jIndicates the evaluation index v₁Normalization of the membership of rank j results. v. of₁₁Indicates the evaluation index v₁Degree of membership to rank 1; v. of₁₂Indicates the evaluation index v₁Degree of membership to level 2; v. of_1mIndicates the evaluation index v₁Degree of membership to rank m.

The above-mentioned evaluation index v is used only as the evaluation index₁Determining an evaluation index v for the membership degree of each evaluation grade₁Normalizing the results of membership for each rating levelThe procedure of S302 is explained. For other evaluation indices (e.g.: v)₂To v_nEtc.), the normalization results of the membership degrees of other evaluation indexes to each evaluation level can be obtained in a similar manner.

The membership degree delay of each evaluation level in the SLA evaluation level set shown in table 1 by the evaluation index delay described in the above S301^{Degree of membership of each grade}For example, the result of normalization of the membership of delay to level 1 can be calculated by the above equation (1).

In the above formula, d₁Normalized for delay to membership of rank 1. delay1 through delay5 in turn represent the degree of membership of delay to level 1 through level 5.

Similarly, in the same manner as the normalization result of delay for the membership of level 1 is calculated, the normalization result d of delay for the membership of level 2 can be calculated₂Normalization of the membership of delay to level 3₃Normalization of the membership of delay to level 4₄And delay to rank 5 membership.

It should be noted that, the above-mentioned process of determining the normalized result of the membership degree of delay to each evaluation level according to the membership degree of delay to each evaluation level is only used for exemplarily describing S302. For other evaluation indexes (such as throughput, loss, etc.), the normalization result of the membership degree of the other evaluation indexes to each evaluation grade can be obtained in a similar manner.

And S303, establishing an evaluation matrix according to the normalization result of the membership degree of each evaluation index to each evaluation grade.

In the present example, the evaluation index v obtained in S302 is assumed₁The normalization results of the membership degrees from the level 1 to the level m are v in sequence₁₁、v₁₂、...、v_1m，v₂Membership to level 1 to level mThe normalized result of the degree is v in turn₂₁、v₂₂、...、v_2m，...，v_nThe normalization results of the membership degrees from the level 1 to the level m are v in sequence_n1、v_n2、...、v_nm。

According to the evaluation index v₁Normalization result of membership degree from level 1 to level m, evaluation index v₂Normalization results of membership degrees from level 1 to level m₂For the normalized results of the membership degrees of the grades 1 to m, the evaluation matrix B is established as follows:

illustratively, also taking the example that the SLA evaluation level set shown in table 1 includes levels 1 to 5, and the SLA evaluation index set includes delay, throughput, and loss as examples, assuming that the normalization results of delay on the membership degrees of levels 1 to 5 obtained in S302 are d in sequence₁、d₂、d₃、d₄、d₅The normalization of the degree of membership of grade 1 to grade 5 by the throghput results in turn tp₁、tp₂、tp₃、tp₄、tp₅The normalized result of loss on the membership of level 1 to level 5 is l in order₁、l₂、l₃、l₄、l₅。

The evaluation matrix B established based on the normalization result of delay for the membership degree of level 1 to level 5, the normalization result of throughput for the membership degree of level 1 to level 5, and the normalization result of loss for the membership degree of level 1 to level 5 may be as follows:

s203, determining the membership degree of the SLA evaluation index set to each evaluation level in the SLA evaluation level set according to the evaluation matrix and the weight of each evaluation index in the SLA evaluation index set.

Fig. 4 is a schematic flowchart of a service level agreement mapping method according to an embodiment of the present application. As shown in fig. 4, S203 described above may include S401-S403.

S401, obtaining the weight of each evaluation index in the SLA evaluation index set.

In the embodiment of the application, a large number of training sets corresponding to an SLA evaluation index set can be learned by using an eXtreme gradient boosting (XGBoost) model to obtain the weight of each evaluation index, and the weights of each evaluation index can form a weight set.

Illustratively, also taking the example that the SLA evaluation level set shown in table 1 includes level 1 to level 5, and the SLA evaluation index set includes delay, throughput, and loss as an example, the training set corresponding to the SLA evaluation index set may include a large number of training samples corresponding to delay, throughput, and loss, respectively.

For example, training sample 1: v₁The evaluation grade of the training sample 1 is grade 1; training sample 2: v₂The evaluation grade of the training sample 1 is grade 2; training sample 3: v₃The evaluation grade of the training sample 3 is grade 3; training sample 4: v₄The evaluation grade of the training sample 4 is grade 5; training sample 5: v₅The evaluation grade of the training sample 5 is grade 5; training sample 6: v₆The evaluation grade of the training sample 6 is grade 1; training sample 7: v₇The evaluation grade of the training sample 7 is grade 2; training sample 8: v₈The evaluation grade of the training sample 8 is grade 3; training sample 9: v₉The evaluation grade of the training sample 9 is grade 4; training sample 10: v₁₀The evaluation grade of the training sample 10 is grade 5; training sample 11: v₁₁＝{loss＜10^-5The evaluation grade of the training sample 11 is grade 1; training sample 12: v₁₂＝{loss＜10^-5The evaluation grade of the training sample 12 is grade 2; trainingSample 13: v₁₃＝{loss＜10^-5The evaluation grade of the training sample 13 is grade 3; training sample 14: v₁₄＝{loss＜10^-4The evaluation grade of the training sample 14 is grade 4; training sample 15: v₁₅＝{loss＜10^-3The rating of the training sample 15 is grade 5.

The extreme gradient boosting model is used for learning a large number of training samples corresponding to the delay, the throughput and the loss respectively, so that the weight of the delay, the weight of the throughput and the weight of the loss can be obtained.

For example, the obtained training samples 1 to 15 and the evaluation levels corresponding to the training samples 1 to 15 are taken as examples, the X6Boost model may include a prediction model. After the SLA evaluation index set is input into the XGboost model, the XGboost model can respectively obtain the score corresponding to each evaluation index in the SLA evaluation index set according to the prediction model.

The prediction model can be expressed as the following formula (1):

in the formula (2), V_iRepresenting the ith training sample, i is an integer greater than 0, and i is less than or equal to the total number of training samples; f. of_k(V_i) Representing the corresponding scores of the ith training sample on the kth classification and regression tree; d is the total number of the classification and regression trees, D is an integer greater than 0, and k is an integer greater than 0 and less than D;representing the corresponding prediction score of the training sample. The practical meaning of the formula (2) is that the score corresponding to the ith training sample obtained by the prediction model is the sum of the score corresponding to the ith training sample on the 1 st classification and regression tree to the score corresponding to the Dth classification and regression tree.

Score f of ith training sample on kth classification and regression tree_k(V_i) Can be expressed as the following formula (3):

in the formula (3), f_k(V_i) Representing the corresponding scores of the ith training sample on the kth classification and regression tree;the tree structure represents the classification and regression tree where the ith training sample is;representing the first training sample in a classification and regression tree structure ofThe score of the leaf node in the classification and regression tree; r^TA set of vectors representing leaf nodes; r^dA set of structures representing classification and regression trees; t represents a classification and regression tree structure as R^dThe classification of any structure and the number of leaf nodes in the regression tree. The practical meaning of formula (3) is that the prediction model obtains the structure of the ith training sample in one treeThe corresponding score of the classification and regression tree of (1) is the sum of the scores of the T leaf nodes of the ith training sample on the tree.

The objective function of the prediction model is:

in the formula (4), y_iRepresents the actual score, y, corresponding to the ith training sample_iThe evaluation level can be obtained by conversion of the evaluation level corresponding to the ith training sample, and the embodiment of the application is not limited to the specific rule of conversion; m represents the number of training samples, and m is an integer greater than 0; l denotes predictionThe model predicts the training errors after m training samples;canonical term, Ω (f), representing D classes and regression trees_k) Representing the complexity of the kth classification and regression tree.

The complexity of the kth classification and regression tree can be calculated by equation (5):

in the formula (5), T_kRepresenting the number of leaf nodes in the kth classification and regression tree; gamma denotes T_kThe coefficient of (a); omega_k，jRepresents the score (which may also be a weight) of the kth classification and the jth leaf node in the regression tree; λ represents the pair ω_k，jPenalty of (2). The practical meaning of the formula (5) is to control the number of the leaf nodes of the classification and regression tree and the weight of each leaf node, and prevent the overfitting caused by the excessive number of the leaf nodes of the classification and regression tree and the excessive weight of a single leaf node.

The objective function of the t-th iteration is calculated by the prediction result of the t-1 th iteration and the classification and regression tree of the t-th inclusion model, because each iteration generates a tree, the objective function in the formula (4) is changed, then the objective function is expanded by a second-order taylor series, the objective function calculates partial derivative of omega to obtain a weight value which can enable the objective function to be minimum, then the omega is brought back to the objective function, and the minimum objective function value is solved (the smaller the objective function value is, the lower the complexity of the classification and regression tree is, the stronger the generalization ability is):

in the formula (6), the first and second groups,representing the optimal fraction of the jth leaf node after the tth iteration; i is_jRepresents the set of training samples at the jth leaf node, I_j＝{i|q(V_i)＝j}；g_iIndicates the predicted score of l for the t-1 st timeThe first derivative of (a); h is_iIndicates the predicted score of l for the t-1 st timeSecond derivative of, T_tRepresents the total number of classification and regression trees for the t-th iteration.

g_iAnd h_iCan be calculated by equation (8) and equation (9), respectively:

the prediction model recursively selects the optimal features of the tree structure from the root node by adopting a greedy algorithm, and the training sample is segmented according to the optimal features. Let I_LSet of samples to the left of the cut point, I_RIs the set of samples to the right of the cut point, and I ═ I_L∪I_R. And calculating the information gain of each segmentation scheme, wherein the segmentation with the maximum information gain is the optimal segmentation of the node:

in the formula (10), L_splitIndicates the gain of the information after the division,representing the sum of the scores of the left sub-tree after splitting,representing the sum of the scores of the right subtree after splitting,represents the score value of the parent node before splitting and gamma represents the complexity cost introduced by adding a new leaf node. L is_splitIf < 0, the division is discarded.

The importance (numerical value of the degree of importance) of each evaluation index can be obtained by calculating the information gain of each evaluation index:

in the formula (11), the reaction mixture,represents the importance of the ith training sample, S_kRepresents the set of split points for each tree, and s represents a split point.

By calculating the importance ratio of each training sample, the weight of each evaluation index to the corresponding evaluation grade can be obtained, and the calculation process can be shown as formula (12):

in the formula (12), a_iThe weight of the i-th evaluation index is represented.

For example, also taking the example that the SLA evaluation level set shown in table 1 includes level 1 to level 5, and the SLA evaluation index set includes delay, throughput, and loss as an example, it is assumed that the weight of each kind of evaluation index obtained according to formula (12) in step S402 may be used to obtain a weight set a of the evaluation index set to the evaluation level set, which is {0.5, 0.3, 0.2 }. The weight set a indicates that the evaluation index of which the evaluation index type is delay has a weight of 0.5 for the evaluation level; the weight of the evaluation index with the evaluation index type of throughput to the evaluation grade is 0.3; the weight of the evaluation index having the evaluation index type of loss to the evaluation level is 0.2.

It can be understood that, as the number of types of evaluation indexes in the training samples increases, the number of weights in the weight set a may also increase, and the number of weights in the embodiment of the present application is not limited.

The process of determining the weight of each evaluation index in the SLA evaluation index set is described above, and after the weight of each evaluation index in the SLA evaluation index set is obtained, the process of determining the membership degree of the SLA evaluation index set to each evaluation level in the SLA evaluation level set according to the evaluation matrix and the weight of each evaluation index in the SLA evaluation index set may include steps S402 and S403.

S402, generating a weight matrix according to the weight of each evaluation index in the SLA evaluation index set.

And S403, determining the membership degree of the SLA evaluation index set to each evaluation level in the SLA evaluation level set according to the weight matrix and the evaluation matrix.

In the embodiment of the present application, in S402, the evaluation index v is used₁Weight of a₁Evaluation index v₂Weight of a₂A_nWeight of a_nThe composed weight matrix can be represented as a:

A＝{0₁0₂...a_n}

the evaluation matrix established in S202 is:

in S403, fuzzy synthesis calculation may be performed on the evaluation matrix B and the weight matrix a by the following formula (13) to determine the membership of the evaluation SLA evaluation index set to each evaluation level in the SLA evaluation level set.

In S403, the evaluation matrix B and the weight matrix a may be subjected to fuzzy synthesis calculation by the following formula (13) to determine the membership of the SLA evaluation index set shown in table 1 to each evaluation level in the SLA evaluation level set.

In equation (13), result represents the degree of membership,representing fuzzy synthetic relations, the computation of fuzzy synthetic relations being similar to the computation of the product of matrices, result₁Represents an SLA evaluation index set (v)₁，v₂，...，v_n) Degree of membership to level 1, result₂Set of price indices (v)₁，v₂，...，v_n) Degree of membership to level 2, result_mSet of price indices (v)₁，v₂，...，v_n) Degree of membership to rank m.

For example, also taking the case where the SLA evaluation level set shown in table 1 above includes level 1 to level 5, and the SLA evaluation index set includes delay, throughput, and loss as examples, assume that the weight of delay is a₁The weight of the throughput is a₂The weight of loss is a₃The evaluation matrix established in S202 is:

the weight matrix composed of the weight according to delay, the weight according to throughput, and the weight according to loss in S402 may be represented as a:

A＝{a₁，a₂，a₃}

in S403, the evaluation matrix B and the weight matrix a may be subjected to fuzzy synthesis calculation by the following formula (13) to determine the degree of membership of the SLA evaluation index set shown in table 1 to each evaluation level in the SLA evaluation level set.

And S204, determining the membership degree of the SLA evaluation index set to each evaluation level in the SLA evaluation level set, wherein the evaluation level corresponding to the maximum membership degree is the evaluation level of the SLA evaluation level set.

In the embodiment of the present application, the maximum membership degree can be calculated by equation (14):

r＝Max(result₁，result₂，...，result_m) Formula (14)

In the formula (14), r represents an evaluation level corresponding to the maximum membership degree in the SLA evaluation level set. Max stands for fetch₁、result₂、...、result_mMaximum value of (2).

For example, also in the case where the SLA evaluation level set shown in table 1 includes levels 1 to 5, and the SLA evaluation index set includes delay, throughput, and loss, assuming that the membership degree result of the evaluation index combination of delay, throughput, and loss obtained by the formula (13) in S204 is (0.1,0.2,0.2,0.3,0.5) for the evaluation levels of level 1, level 2, level 3, level 4, and level 5, the membership degree of the evaluation index combination for the level 5 is the largest and is 0.5, and therefore the evaluation level of the evaluation index combination is level 5.

The service level agreement mapping method provided by the embodiment of the application can divide the SLA parameter into a plurality of (e.g., at least two) service levels (or referred to as evaluation levels), and provide a mapping manner between the SLA parameter and the QoS parameter corresponding to the SLA parameter by means of machine learning and mathematical computation. The mapping mode between the SLA parameters and the QoS parameters provided in the service level agreement mapping method can be suitable for different service services and different bearer network types, and has universality. The service provider is convenient to manage and arrange the services provided by the service provider.

In addition, the mapping between the SLA parameters and the QoS parameters in the service level agreement mapping method is more objective.

In addition, the service level agreement mapping method provided by the embodiment of the application can also map the service level agreement in a reverse direction, and the QoS parameter can also be mapped to the SLA parameter in a reverse direction through the mapping mode between the SLA parameter and the QoS parameter provided by the service level agreement mapping method, so that a service user can conveniently evaluate the service used by the service user.

Fig. 5 is a schematic structural diagram of a service level agreement mapping apparatus according to an embodiment of the present application. As shown in fig. 5, the apparatus includes an obtaining module 501 and a processing module 501

An obtaining module 501, configured to obtain a quality of service (QoS) parameter set and an SLA evaluation level set corresponding to a Service Level Agreement (SLA) parameter;

the processing module 502 is configured to establish an evaluation matrix according to the QoS parameter set and the SLA evaluation level set, determine a membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the evaluation matrix and a weight of each QoS parameter in the QoS parameter set, determine a maximum membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set, and determine an evaluation level corresponding to the maximum membership degree as an evaluation level corresponding to the QoS parameter set.

In one possible design, the obtaining module 501 is further configured to obtain a weight of each QoS parameter in the QoS parameter set; the processing module 502 is further configured to generate a weight matrix according to the weight of each QoS parameter in the QoS parameter set, and determine the membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to the weight matrix and the evaluation matrix.

In another possible design, the obtaining module 501 is further configured to obtain a fuzzy synthetic relationship calculation result of the weight matrix and the evaluation matrix; the processing module 502 is further configured to determine a membership degree of the QoS parameter set to each evaluation level in the SLA evaluation level set according to a fuzzy synthetic relationship calculation result of the weight matrix and the evaluation matrix.

In yet another possible design, the obtaining module 501 is further configured to obtain a membership degree of each QoS parameter in the QoS parameter set to each evaluation level in the SLA evaluation level set; the processing module 502 is further configured to determine a normalization result of the membership degree of each QoS parameter to each evaluation level in the SLA evaluation level set according to the membership degree of each QoS parameter to each evaluation level, and establish an evaluation matrix according to the normalization result of the membership degree of each QoS parameter to each evaluation level.

In another possible design, the obtaining module 501 is further configured to obtain the number of QoS parameters that satisfy each evaluation level; the processing module 502 is further configured to obtain a membership degree of each QoS parameter to each evaluation level in the SLA evaluation level set according to the number of QoS parameters satisfying each evaluation level.

Optionally, the obtaining module 501 is specifically configured to determine an information gain of each QoS parameter in the QoS parameter set according to a limit gradient lifting model; determining the value of the importance degree of each QoS parameter in the QoS parameter set according to the information gain of each QoS parameter in the QoS parameter set; and calculating the importance ratio of each QoS parameter according to the value of the importance degree of each QoS parameter in the QoS parameter set to obtain the weight of each QoS parameter.

Fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present application, and as shown in fig. 6, the electronic device includes: a processor 601 and a memory 602; the memory stores instructions executable by the processor; the processor is configured to execute the instructions, causing the electronic device to implement the method as described in the preceding embodiments.

In an exemplary embodiment, the present application further provides a computer-readable storage medium having stored thereon computer program instructions; the computer program instructions, when executed by an electronic device, cause the electronic device to implement a method as described in the preceding embodiments.

The computer readable storage medium may be a non-transitory computer readable storage medium, which may be, for example, a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

The above description is only an embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions within the technical scope of the present disclosure should be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.