Network slice resource allocation method and device for power data network

Abstract

The embodiment of the invention provides a network slice resource allocation method and device for a power data network. The method comprises the following steps: receiving a power service slice request; wherein the power service slice request comprises slice state information, and the slice state information comprises a source node, a destination node and a performance demand condition; obtaining current link state information; wherein the link state information comprises a physical link set and performance parameters of each physical link in the physical link set; according to the slice state information and the link state information, determining an optimal physical link meeting the performance demand condition from the physical link set by using a pre-constructed link distribution model; and establishing a corresponding relationship among the source node, the destination node and the optimal physical link, and configuring a network slice of the power service. The performance requirements of different power businesses can be met, and the overall performance and quality of the power data network are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of network technology, and in particular, to a method and device for allocating network slice resources in a power data network. Background Art

[0002] The power data network is a comprehensive business transmission platform for power companies to achieve informatization. It is the basis for the interconnection of various computer application systems within the power company system and an important part of the power information infrastructure. Various power business applications running on the power data network involve industrial safety of the power grid, and the requirements for partition isolation and service level agreement (SLA) guarantees are getting higher and higher. The high-cost and low-efficiency traditional IP network communication technology can no longer meet the differentiated requirements of the power data network for latency, bandwidth, multi-protocol, security and reliability. Summary of the invention

[0003] In view of this, the purpose of an embodiment of the present application is to propose a network slice resource allocation method and device for a power data network.

[0004] Based on the above purpose, an embodiment of the present application provides a network slice resource allocation method for a power data network, including:

[0005] Receive a power service slicing request; wherein the power service slicing request includes slicing status information, and the slicing status information includes a source node, a destination node, and a performance requirement condition;

[0006] Acquire current link status information; wherein the link status information includes a physical link set and performance parameters of each physical link in the physical link set;

[0007] Determine, according to the slice state information and the link state information, an optimal physical link that meets the performance requirement from the physical link set using a pre-built link allocation model;

[0008] Establish a correspondence between the source node, destination node and optimal physical link, and configure the network slice for the power business.

[0009] Optionally, the link allocation model is implemented based on an actor network model, and the actor network model and the policy network model are trained simultaneously to obtain the link allocation model after training; wherein, the state of the actor network model includes the slice state information and the link state information, the action of the actor network model includes the selected physical link, and the reward function of the policy network model is constructed according to the performance requirement conditions.

[0010] Optionally, determining, according to the slice state information and the link state information, an optimal physical link that meets the performance requirement condition from a set of physical links using a pre-built link allocation model includes:

[0011] Processing the slice state information and the link state information into a feature vector suitable for processing by the link allocation model;

[0012] Inputting the feature vector into the link allocation model, and outputting the probability distribution of the action to be selected by the link allocation model;

[0013] According to the probability distribution of the actions to be selected, the action with the highest probability is selected, and the physical link corresponding to the action is used as the optimal physical link.

[0014] Optionally, the performance requirement conditions include the service type, latency and bandwidth of the power service, and the service type includes control services, information collection services and mobile application services; the reward function is:

[0015] R(s t ,a t )=(αR bw +βR delay )×γR pri (5)

[0016] Among them, R(s t ,a t ) is in state s t Next, perform action a t The reward value obtained, R bw is the bandwidth constraint of the power business, R delay is the delay constraint of the power business, R pri is the priority of power services. The priority of control services is higher than that of information collection services and mobile application services. α, β, and γ are weight parameters.

[0017] Optionally, the method further includes:

[0018] Obtain resource utilization of network slices;

[0019] When it is determined that the resource utilization rate is lower than a preset utilization rate threshold for a predetermined period of time, the network slice is recycled.

[0020] Optionally, the method further includes:

[0021] Get the remaining resources of the network slice;

[0022] When it is determined that the remaining resources are less than a preset resource threshold, the network slice is reconfigured based on the recovered network slice and the current link status information.

[0023] Optionally, the reconfiguring the network slice includes:

[0024] If the reconfiguration of the network slice fails, the priority of each currently configured network slice is calculated based on the preset service type weight and the resource utilization of the network slice;

[0025] Reclaim the network slices whose priority reaches the preset low condition, and reconfigure the network slices based on the recycled network slices and the current link status information.

[0026] Optionally, the reconfiguring the network slice further includes:

[0027] If the number of network slice reconfiguration failures exceeds the preset failure threshold, all network slices will be reconfigured.

[0028] Optionally, the method further includes:

[0029] Deploy network slices in the production control area and information management area based on the hard isolation method;

[0030] Based on the soft isolation method, network slices for different power services in the information management area and network slices for different power services in the production control area are deployed.

[0031] The embodiment of the present application also provides a network slice resource allocation device for a power data network, including:

[0032] A receiving module, configured to receive a power service slicing request; wherein the power service slicing request includes slicing status information, and the slicing status information includes a source node, a destination node, and performance requirement conditions;

[0033] An acquisition module, used to acquire current link status information; wherein the link status information includes a physical link set and a performance parameter of each physical link in the physical link set;

[0034] An allocation module, configured to determine an optimal physical link that meets the performance requirement from the physical link set using a pre-built link allocation model according to the slice state information and the link state information;

[0035] A configuration module is used to establish a correspondence between the source node, the destination node and the optimal physical link, and configure the network slice for the power business.

[0036] From the above description, it can be seen that the network slice resource allocation method and device of the power data network provided in the embodiment of the present application receives the power service slice request, obtains the current link status information, and determines the optimal physical link that meets the performance requirement conditions from the physical link set based on the slice status information and link status information using the pre-built link allocation model, establishes the corresponding relationship between the source node, the destination node and the optimal physical link, and configures the network slice of the power service. The present application can meet the performance requirements of different power services and improve the overall performance and quality of the power data network. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 A schematic diagram of a method flow of an embodiment of the present application;

[0039] Figure 2 A schematic diagram of a reinforcement learning model according to an embodiment of the present application;

[0040] Figure 3 This is a block diagram of the device structure of an embodiment of the present application;

[0041] Figure 4 This is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0042] In order to make the objectives, technical solutions and advantages of the present disclosure more clearly understood, the present disclosure is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0043] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be the usual meanings understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the embodiments of the present application do not represent any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing in front of the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connecting" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to represent relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0044] In the related technologies, the electric power data network is divided into the production control area and the information management area. The related electric power services between the two areas cannot communicate directly. Various types of electric power services, such as information collection, mobile application and control, are running in the network. Different types of electric power services have different requirements for partition isolation, network resource requirements, service levels, etc.

[0045] In order to meet the needs of different types of power services and improve the resource utilization of the power data network, the embodiment of the present application allocates optimal links for different power services based on the reinforcement learning method, configures corresponding network slices, and realizes physical hard isolation or logical soft isolation to meet the performance requirements of different power services. It can also dynamically optimize resource configuration according to resource status and improve network performance.

[0046] The technical solution of the present application is further described in detail below through specific embodiments.

[0047] like Figure 1 As shown, an embodiment of the present application provides a network slice resource allocation method for a power data network, including:

[0048] S101: receiving a power service slicing request; wherein the power service slicing request includes slicing status information, and the slicing status information includes a source node, a destination node, and performance requirement conditions;

[0049] In this embodiment, when a network slice is configured for a specific power service, a power service slice request is received, and the power service slice request includes slice status information. The slice status information includes a source node, a destination node, and performance requirements such as the service type of the power service, the latency upper limit and bandwidth of the power service.

[0050] S102: Acquire current link status information; wherein the link status information includes a physical link set and performance parameters of each physical link in the physical link set;

[0051] In this embodiment, the current link state information is obtained so as to configure a suitable network slice for the power business in combination with the current network resources. The link state information includes a physical link set, which includes two physical nodes of each physical link, performance parameters such as bandwidth and latency.

[0052] In some approaches, the underlying physical network topology can be abstracted as an undirected weighted graph G p =(N p ,L p ), where N p Represents the set of nodes of the physical network, L p Represents the set of physical links of the underlying physical network. Each physical link is a four-tuple, including the two physical nodes of the link, the available bandwidth of the link, and the transmission delay. Assume u,v∈N p For two directly connected physical nodes, the four-tuple of the physical link (u, v) is represented by K p (u,v)=(U,V,B,D), where B represents the bandwidth of the physical link (u,v) and D represents the transmission delay of the physical link (u,v).

[0053] S103: Determine an optimal physical link that meets the performance requirement from a set of physical links using a pre-built link allocation model according to the slice state information and the link state information;

[0054] S104: Establish a correspondence between the source node, the destination node and the optimal physical link, and configure the network slice for the power business.

[0055] In this embodiment, after obtaining the slice status information and link status information, the optimal physical link that meets the performance requirements of the power business is selected from the physical link set based on the pre-built link allocation model, and a mapping relationship between the source node, the destination node and the two nodes of the physical link is established. The network slice for the power business is configured, and the network slice is used to carry the corresponding power business to meet the application requirements of the power business and improve network performance.

[0056] In some embodiments, the link allocation model is implemented based on the actor network model, and the actor network model and the policy network model are trained simultaneously to obtain the link allocation model after training; wherein, the state of the actor network model includes slice state information and link state information, the action of the actor network model includes the selected physical link, and the reward function of the policy network model is constructed according to the performance requirement conditions.

[0057] In this embodiment, according to the acquired slice state information and link state information, the optimal physical link is selected based on the reinforcement learning model. The reinforcement learning model includes an actor network model and a policy network model that combine value learning and policy learning, namely, an Actor network and a Critic network. Among them, the actor network model is used to select an action according to the current state, usually represented by a parameterized function (such as a neural network) and outputs the probability distribution of one or more actions according to the current state. The policy network model is used to evaluate the quality of selecting a specific action, which is achieved by estimating the state value function V(s) or the action value function Q(s,a). The estimation of the value function is used to guide the actor network model to update in order to make better action selections. The actor network model and the policy network model are updated through synchronous training to obtain a trained actor network model, namely a link allocation model, so as to use the link allocation model to determine the optimal physical link according to the current slice state information and link state information.

[0058] In some implementations, the state space of the reinforcement learning model includes slice state information and link state information, and the state at the current time t can be expressed as:

[0059] s t ={S R (t),S G (t)} (1)

[0060] S R (t)={v×f′| v∈|R|} (2)

[0061] S G (t)={i×f | i∈|E|} (3)

[0062] Among them, S R is a collection of slice status information, v represents slice request, and the types include information collection, control and mobile application, f′=(f1′f2′…f k ′) T S represents the k characteristic information dimensions of the power business. The characteristic information of the power business includes occupied bandwidth, latency, and currently selected physical link. R is the set of slice requests. G is the feature set of the physical link, f=(f1f2…f m ) T , represents the m characteristic dimension information of the i-th physical link. The characteristic information of the physical link includes the link remaining bandwidth, link initial bandwidth, link delay, and whether the link has been mapped. E is the set of physical links.

[0063] The action space is defined as the physical links and bandwidths to be selected, and the action selected at time t is expressed as:

[0064] at = {L t ,B t} (4)

[0065] The reward is the feedback provided by the environment after the action is selected to determine whether the action is correct. If the action is correct, the environment will provide positive feedback. If the action is wrong, the environment will provide negative feedback. This optimizes the selection of actions based on the reward value. In this embodiment, the reward function is constructed based on the performance requirements of the power business and is expressed as:

[0066] R(s t ,a t )=(αR bw +βR delay )×γR pri (5)

[0067] Among them, R(s t ,a t ) is in state s t Next, perform action a t The reward value obtained, α, β, γ are weight parameters, R bw is the bandwidth constraint of the power business, R delay is the delay constraint of the power business, R pri The priority of the power business is that the priority of the control business is higher than that of the information collection business and the mobile application business.

[0068] For the bandwidth constraint of the power business, in order to ensure the quality of service, the selected physical link must not only meet the bandwidth requirements of the power business, but also consider a certain degree of reservation. If the physical link meets or exceeds the reserved bandwidth requirements, a positive reward will be given; if it does not meet the reserved bandwidth requirements, a negative reward will be given. The bandwidth constraint can be expressed as:

[0069]

[0070] Where b is the actual available bandwidth of the physical link, b th The reserved bandwidth threshold for power services.

[0071] In some methods, considering the high importance of control services, the reserved bandwidth ratio is set higher, that is, the reserved bandwidth is larger. If the bandwidth requirements are met (including the bandwidth requirements and reserved bandwidth ratio required for the power business), a larger positive reward can be obtained, and if not, a larger negative reward can be obtained; the reserved bandwidth ratio of information collection services is set smaller, and a smaller positive reward can be obtained if the bandwidth requirements are met, and a smaller negative reward can be obtained if they are not met; the reserved bandwidth ratio of mobile application services can be set between the reserved bandwidth ratios of control services and information collection services.

[0072] The delay constraint for the power business can be set as:

[0073]

[0074] Where, d is the actual delay, d th It is the delay threshold of the power service.

[0075] Regarding the priority of power services, the production control area of ​​the power data network deploys control services, which are the core services of power production. The system runs online in real time and has the highest priority. The information management area deploys information collection services and mobile application services to achieve the management of power dispatching and production, which have a lower priority than control services. As a result, high-priority services receive higher rewards, and low-priority services receive less rewards. In some ways, the priority of power services can be set as:

[0076]

[0077] In some embodiments, a proximal policy optimization algorithm is used to improve the training process of the actor network model and the policy network model, making the algorithm more stable and efficient. Figure 2 As shown in the figure, in one training round, the agent will use the existing strategy to interact with the environment K times and obtain K trajectories {s t a t ,P(π λold (a t |s t )),r t )},t=0,1,2,…,K,P(π θold (a t |s t )) means selecting action a from the action values ​​that obey the Gaussian distribution t The actor network model using the proximal policy optimization algorithm outputs the mean of the action, making the action value obey the Gaussian distribution, and then randomly samples an action a t , this method can effectively prevent excessive updates of the strategy. During the training process, use the old strategy To sample data to update the policy π θ , and use the clip function to prevent π θ The training instability caused by the large update of can be reduced, thereby maintaining a more robust learning process and performance.

[0078] In some implementations, the training process of the actor network model and the policy network model using the proximal policy optimization algorithm specifically includes:

[0079] 1) Initialize the initial number of iterations and the maximum number of iterations, and set the experience replay pool; 2) Create an agent object instance, which includes an actor network model and a strategy network model based on a graph neural network, and initialize the model parameters; 3) Use the graph neural network to obtain information related to the physical network topology structure, obtain link status information, receive power business slice requests, and obtain slice status information; 4) Construct the current state s based on the link status information and slice status information t ; 5) Process the current state into a feature vector suitable for graph neural network processing, input the feature vector into the actor network model, the agent generates an action, and outputs the probability distribution of the action; 6) Randomly sample the action a based on the probability distribution t And the probability of action selection 7) Perform action a in the environment t , get the updated state s after executing the action t+1 , according to formula (5) to calculate the execution action a t The reward value obtained, perform action a t After that, the environment is updated, that is, the delay, utilization, quality, etc. of the physical link change; 8) the current state s t 、Action a t , reward value r t and updated status t+1 As a state transfer sample (s t , a t , r t ,s t+1 ) is saved in the experience replay pool. Repeat the above steps 3) to 8) to construct a certain number of state transition samples until the experience replay pool is full.

[0080] The agent updates the model parameters based on the state transition samples in the experience replay pool, including:

[0081] 10) Update the state s in the sample t+1 In the input policy network model, the policy network model performs a series of linear combinations and nonlinear transformations on the input based on the model parameters through the activation function in the hidden layer to obtain the state value v′. According to the reward value at different times, the corresponding discounted reward R[t]=r[t]+δ1×r[t+1]+…+δ T-t ×r[T], we get a reward value set R=[R[0],R[1],…,R[t],…R[T]] consisting of reward values ​​at different times; where r[t+1] is the reward value at time t+1, T is the total number of time steps, δ1, δ T-t is the discount factor.

[0082] 11) Input the states corresponding to each moment in the sample into the strategy network model respectively, and process the input states through the linear combination of each layer of the strategy network model and the activation function ReLU to obtain the Vφ value corresponding to the state; According to the Vφ value and reward value set corresponding to each state, calculate the corresponding advantage function According to the advantage function corresponding to each state, the gradient is updated using the loss function of the policy network model. The loss function is expressed as:

[0083]

[0084] 12) Input the states and actions corresponding to each moment in the sample into the actor network model to obtain the probability ratio of the new and old strategies The loss function of the actor network model is used for updating, and the loss function is expressed as:

[0085]

[0086] Among them, ∈ is the cutoff hyperparameter, the value is 0.2, clip(r θ ,1-∈,1+∈) is used to convert r θ It is restricted to the interval [1-∈, 1+∈] to ensure its convergence.

[0087] Repeat the above steps 11)-12) until the predetermined number of times is reached. Determine whether the maximum number of iterations has been reached. If not, repeat the steps of constructing state transition samples and updating model parameters according to the samples. If the maximum number of iterations has been reached, end the training and obtain the optimal model parameters of the actor network model. The actor network model with the optimal model parameters set is used as a link allocation model to allocate the optimal link for the power business.

[0088] In some embodiments, according to the slice state information and the link state information, a pre-built link allocation model is used to determine the optimal physical link that meets the performance requirement from the physical link set, including:

[0089] Processing the slice state information and the link state information into feature vectors suitable for link allocation model processing;

[0090] The feature vector is input into the link allocation model, and the link allocation model outputs the probability distribution of the action to be selected;

[0091] According to the probability distribution of the actions to be selected, the action with the highest probability is selected, and the physical link corresponding to the action is taken as the optimal physical link.

[0092] In this embodiment, when the trained link allocation model is used to allocate the optimal link, the current slice state information and link state information are processed into a feature vector suitable for model processing, and the feature vector is input into the link allocation model. The link allocation model outputs the probability distribution of multiple actions to be selected. Since the actor network model adjusts parameters according to reward feedback, high-probability actions are more likely to obtain high reward values. Therefore, the action with the highest probability is selected from the multiple actions to be selected. The link corresponding to the action is the optimal physical link suitable for the power business. The optimal physical link is the optimal link suitable for the business type of the power business in the provided physical link set, and meets the performance requirements such as latency and bandwidth, while making full use of network resources. After configuring the network slice of the power business based on the selected optimal physical link, the network slice can carry the data flow of the power business that meets the access requirements.

[0093] In some embodiments, considering that changes in user traffic characteristics will affect the resource utilization efficiency of network slices, if the flow request type cannot adapt to the performance range of the network slice, it will lead to resource waste, and too dense traffic may reduce the performance of the network slice. Therefore, it is necessary to dynamically adjust the network slice in combination with the resource status of the network slice and the network resource status. The network slice resource allocation method also includes:

[0094] Obtain resource utilization of network slices;

[0095] When it is determined that the resource utilization rate is lower than the preset utilization threshold for a predetermined period of time, the network slice is reclaimed.

[0096] In this embodiment, by continuously acquiring and monitoring the resource utilization of the network slice, when it is determined that the resource utilization of the network slice is lower than the preset utilization threshold for a certain period of time, the network slice is recycled, that is, the network slice with lower resource utilization is recycled to avoid resource waste. The network slice resource utilization includes the average number of successfully established PDU sessions in the network slice and the utilization of virtualized resources allocated to the network slice.

[0097] In some embodiments, the network slice resource allocation method further includes:

[0098] Get the remaining resources of the network slice;

[0099] When it is determined that the remaining resources are less than the preset resource threshold, the network slice is reconfigured based on the recovered network slices and the current link status information.

[0100] In this embodiment, by continuously or periodically acquiring and monitoring the remaining resources of the network slice, it is determined whether the remaining resources are less than the set resource threshold, and if so, the network slice needs to be reconfigured. The remaining resources of the network slice refer to the remaining resources (including bandwidth) of each path from the source node to the destination node, that is, if the remaining resources of each path are less than the resource threshold, the network slice needs to be recycled first, and then based on the resources of the recycled network slice and the current link status information, the network slice is reconfigured according to the slice allocation method of this application.

[0101] In some embodiments, reconfiguring the network slice includes:

[0102] If the reconfiguration of network slices fails, the priority of each currently configured network slice is calculated based on the preset deployment area weight and the resource utilization of the network slice;

[0103] Reclaim the network slices whose priority reaches the preset low condition, and reconfigure the network slices based on the recycled network slices and the current link status information.

[0104] In this embodiment, during the process of reconfiguring network slices, if a new network slice is not successfully configured due to insufficient resources or other reasons, it is necessary to continue to recycle unimportant network slices with low resource utilization, and then reconfigure network slices based on the recycled network slices and current network resources. Among them, the priority of each network slice is calculated according to the deployment area weight and resource utilization of each network slice, and all network slices are sorted in order from high to low priority, and the network slices with a priority lower than a preset recycling threshold or a predetermined number of network slices ranked at the back of all network slices are recycled as unimportant network slices with low resource utilization.

[0105] In some embodiments, the priority of a network slice is calculated as follows:

[0106] P=w×U (11)

[0107] Among them, w is the deployment area weight, the deployment area weight of the network slice deployed in the production control area is greater than the deployment area weight of the network slice deployed in the information management area, and U is the resource utilization rate.

[0108] In some embodiments, reconfiguring the network slice further includes:

[0109] If the number of network slice reconfiguration failures exceeds the preset failure threshold, all network slices will be reconfigured.

[0110] In this embodiment, during the process of reconfiguring network slices, if the network slices are not successfully reconfigured by recovering them multiple times, it is necessary to reconfigure all network slices according to the method of this application, and optimize the network slice configuration and network resource utilization by reallocating network resources, so that the reconfigured network slices can meet the power business needs and improve the overall network performance and quality.

[0111] In some embodiments, the network slice resource allocation method further includes:

[0112] Deploy network slices in the production control area and information management area based on the hard isolation method;

[0113] Based on the soft isolation method, network slices for different power services in the information management area and network slices for different power services in the production control area are deployed.

[0114] In this embodiment, after configuring corresponding network slices for different power services, it is necessary to deploy network slices so that network slices for different power services can meet different power service requirements through soft isolation or hard isolation. Considering that the production controller area and the information management area cannot communicate directly, if there is business interaction, the information management area cannot actively access the production control area, and the production controller area can push data to the information management area and other network characteristics, when deploying network slices, the network slices of the production controller and the network slices of the information management area need to be hard isolated, and the network slices of different power services in the area can be soft isolated. That is, in response to the physical isolation requirements of the production control area and the management information area in the power data network, the network slices are deployed using the hard isolation method to achieve hard isolation of network slices between the two areas. In response to the logical isolation requirements between various power services in the production control area or the information management area, the network slices are deployed using the soft isolation method to achieve soft isolation of different power services in the production control area and soft isolation of different power services in the information management area.

[0115] In some methods, FlexE (Flex Ethernet) technology is used to perform hard isolation deployment of network slices. By adding a FlexE Shim layer between the physical layer and the data link layer, the Ethernet data link layer rate and the physical layer rate are decoupled, thereby flexibly adjusting the interface rate to ensure the data transmission stability and reliability of the network slices for control services.

[0116] The SRv6-based soft isolation technology is used to perform soft isolation deployment on network slices. By inserting the SegmentRouting Header (SRH) into the IPv6 packet header, an explicit SRv6 SID list is pushed in to define the network node or path list. During packet transmission, packets are processed and transmitted one by one according to the path list in the SRH until the destination address is reached. The SRv6 Locator is used as the unique identifier of the virtual slice, and the SRv6 SID prefixed with the Locator identifies the allocated network resources.

[0117] The network slice resource allocation method for the electric power data network provided in the embodiment of the present application selects the optimal physical link based on the reinforcement learning model according to the slice status information and link status information, and configures independent, physically or logically isolated network slices according to the selected physical links to meet the performance requirements of different electric power services. The configuration of the network slice is dynamically adjusted in combination with the resource status and network status of the electric power service, which can improve the configuration flexibility and adaptability and optimize the network resource configuration. The present application can improve the overall operating efficiency and security of the electric power data network and provide more reliable, efficient, secure and stable electric power data network services.

[0118] It should be noted that the method of the embodiment of the present application can be performed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present application, and the multiple devices will interact with each other to complete the described method.

[0119] It should be noted that the above is a description of a specific embodiment of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0120] like Figure 3 As shown, the embodiment of the present application also provides a network slice resource allocation device for a power data network, including:

[0121] A receiving module, used to receive a power service slicing request; wherein the power service slicing request includes slicing status information, and the slicing status information includes a source node, a destination node, and performance requirement conditions;

[0122] An acquisition module, used to acquire current link status information; wherein the link status information includes a physical link set and performance parameters of each physical link in the physical link set;

[0123] An allocation module, configured to determine an optimal physical link that meets the performance requirement from a set of physical links using a pre-built link allocation model according to the slice state information and the link state information;

[0124] The configuration module is used to establish the correspondence between the source node, the destination node and the optimal physical link, and configure the network slice for the power business.

[0125] For the convenience of description, the above devices are described in terms of functions divided into various modules. Of course, when implementing the embodiments of the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0126] The apparatus of the above-mentioned embodiment is used to implement the corresponding method in the above-mentioned embodiment, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0127] Figure 4 A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.

[0128] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0129] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0130] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0131] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0132] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0133] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0134] The electronic device of the above embodiment is used to implement the corresponding method in the above embodiment, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0135] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0136] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present disclosure (including the claims) is limited to these examples. Based on the concept of the present disclosure, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0137] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present application difficult to understand, the known power supply / ground connection with the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform to be implemented in the embodiments of the present application (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuits) to describe exemplary embodiments of the present disclosure, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0138] Although the present disclosure has been described in conjunction with specific embodiments of the present disclosure, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0139] The embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the scope of protection of the present disclosure.

Claims

1. A network slice resource allocation method for a power data network, characterized in that: include: Receive a power service slicing request; wherein the power service slicing request includes slicing status information, and the slicing status information includes a source node, a destination node, and a performance requirement condition; Acquire current link status information; wherein the link status information includes a physical link set and performance parameters of each physical link in the physical link set; Determine, according to the slice state information and the link state information, an optimal physical link that meets the performance requirement from the physical link set using a pre-built link allocation model; Establish a correspondence between the source node, destination node and optimal physical link, and configure the network slice for the power business.

2. The method according to claim 1, characterized in that The link allocation model is implemented based on an actor network model. The actor network model and the policy network model are trained simultaneously, and the link allocation model is obtained after training; wherein, the state of the actor network model includes the slice state information and the link state information, the action of the actor network model includes the selected physical link, and the reward function of the policy network model is constructed according to the performance requirement conditions.

3. The method according to claim 1, characterized in that: The determining, according to the slice state information and the link state information, an optimal physical link that meets the performance requirement condition from a set of physical links using a pre-built link allocation model includes: Processing the slice state information and the link state information into a feature vector suitable for processing by the link allocation model; Inputting the feature vector into the link allocation model, and outputting the probability distribution of the action to be selected by the link allocation model; According to the probability distribution of the actions to be selected, the action with the highest probability is selected, and the physical link corresponding to the action is used as the optimal physical link.

4. The method according to claim 2, characterized in that: The performance requirement conditions include the service type, delay and bandwidth of the power service, and the service type includes control service, information collection service and mobile application service; the reward function is: R(s t ,a t )=(αR bw +βR delay )×γR pri (5) Among them, R(s t ,a t ) is in state s t Next, perform action a t The reward value obtained, R bw is the bandwidth constraint of the power business, R delay is the delay constraint of the power business, R pri is the priority of power services. The priority of control services is higher than that of information collection services and mobile application services. α, β, and γ are weight parameters.

5. The method according to claim 1, characterized in that Also includes: Obtain resource utilization of network slices; When it is determined that the resource utilization rate is lower than a preset utilization rate threshold for a predetermined period of time, the network slice is recycled.

6. The method according to claim 1, characterized in that Also includes: Get the remaining resources of the network slice; When it is determined that the remaining resources are less than a preset resource threshold, the network slice is reconfigured based on the recovered network slice and the current link status information.

7. The method according to claim 6, characterized in that The reconfiguring of the network slice includes: If the reconfiguration of the network slice fails, the priority of each currently configured network slice is calculated based on the preset service type weight and the resource utilization of the network slice; Reclaim the network slices whose priority reaches the preset low condition, and reconfigure the network slices based on the recycled network slices and the current link status information.

8. The method according to claim 7, characterized in that The reconfiguring of the network slice further includes: If the number of network slice reconfiguration failures exceeds the preset failure threshold, all network slices will be reconfigured.

9. The method according to claim 1, characterized in that: Also includes: Deploy network slices in the production control area and information management area based on the hard isolation method; Based on the soft isolation method, network slices for different power services in the information management area and network slices for different power services in the production control area are deployed.

10. A network slice resource allocation device for a power data network, characterized in that: include: A receiving module, configured to receive a power service slicing request; wherein the power service slicing request includes slicing status information, and the slicing status information includes a source node, a destination node, and performance requirement conditions; An acquisition module, used to acquire current link status information; wherein the link status information includes a physical link set and a performance parameter of each physical link in the physical link set; An allocation module, configured to determine an optimal physical link that meets the performance requirement from the physical link set using a pre-built link allocation model according to the slice state information and the link state information; A configuration module is used to establish a correspondence between the source node, the destination node and the optimal physical link, and configure the network slice of the power business.

Images