CN112738723B - Network resource allocation method and device and computer readable storage medium - Google Patents
Network resource allocation method and device and computer readable storage medium Download PDFInfo
- Publication number
- CN112738723B CN112738723B CN201910962559.5A CN201910962559A CN112738723B CN 112738723 B CN112738723 B CN 112738723B CN 201910962559 A CN201910962559 A CN 201910962559A CN 112738723 B CN112738723 B CN 112738723B
- Authority
- CN
- China
- Prior art keywords
- information
- network
- enhancement
- resource allocation
- intention
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/02—Services making use of location information
- H04W4/029—Location-based management or tracking services
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/02—Resource partitioning among network components, e.g. reuse partitioning
- H04W16/10—Dynamic resource partitioning
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/18—Network planning tools
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/12—Wireless traffic scheduling
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/12—Wireless traffic scheduling
- H04W72/1263—Mapping of traffic onto schedule, e.g. scheduled allocation or multiplexing of flows
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/50—Allocation or scheduling criteria for wireless resources
- H04W72/52—Allocation or scheduling criteria for wireless resources based on load
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/50—Allocation or scheduling criteria for wireless resources
- H04W72/535—Allocation or scheduling criteria for wireless resources based on resource usage policies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/50—Allocation or scheduling criteria for wireless resources
- H04W72/54—Allocation or scheduling criteria for wireless resources based on quality criteria
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Abstract
The disclosure relates to a network resource allocation method, a network resource allocation device and a computer readable storage medium, and relates to the technical field of communication. The method of the present disclosure comprises: determining position information according to the intention information of the user; determining the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information according to the wish information; determining the prediction information of the end-to-end network state corresponding to the position information according to the end-to-end historical network state corresponding to the position information; and determining end-to-end resource allocation information corresponding to the position information according to the enhancement information of each end-to-end network element corresponding to the position information, the enhancement information of the link and the corresponding prediction information of the end-to-end network state.
Description
Technical Field
The present disclosure relates to the field of communications technologies, and in particular, to a method and an apparatus for allocating network resources, and a computer-readable storage medium.
Background
In a 5G network, different service requirements present differentiation on network deployment and resource requirements, which brings new challenges to network deployment and operation of operators, so that the complexity of network planning and the difficulty of accurate resource allocation docking are increased sharply.
The 5G network is more flexible to deploy, and it is expected that flexible allocation of resources can be achieved according to the wishes of users, but there is no specific implementation scheme at present.
Disclosure of Invention
One technical problem to be solved by the present disclosure is: how to allocate resources according to the will of the user and improve the accuracy of resource allocation.
According to some embodiments of the present disclosure, a method for allocating network resources is provided, including: determining position information according to the intention information of the user; determining the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information according to the wish information; determining the prediction information of the end-to-end network state corresponding to the position information according to the end-to-end historical network state corresponding to the position information; and determining end-to-end resource allocation information corresponding to the position information according to the enhancement information of each end-to-end network element corresponding to the position information, the enhancement information of the link and the corresponding prediction information of the end-to-end network state.
In some embodiments, determining, according to the willingness information, the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information includes: under the condition that the user is a first user, inputting the intention information into a first sub-model of the intention translation model, and outputting network management intention information corresponding to the intention information; and inputting the network management intention information into a second sub-model of the intention translation model, and outputting the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information.
In some embodiments, determining, according to the willingness information, the enhancement information of the end-to-end network elements and the enhancement information of the link corresponding to the location information includes: and under the condition that the user is a second user, inputting the intention information into a second sub-model of the intention translation model, and outputting the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information.
In some embodiments, inputting the willingness information into a first sub-model of the willingness translation model, and outputting network management willingness information corresponding to the willingness information includes: in the first submodel, carrying out named entity recognition on the willingness information, and determining named entities of all participles in the willingness information; and converting each participle into a corresponding preset network management word according to the named entity of each participle and a preset translation dictionary, thereby generating network management intention information.
In some embodiments, determining end-to-end resource allocation information corresponding to the location information according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information and the prediction information of the corresponding end-to-end network state includes: and inputting the enhancement information of each network element from end to end corresponding to the position information, the enhancement information of the link and the prediction information of the corresponding network state from end to end into a deep reinforcement learning model to obtain the end-to-end resource allocation information corresponding to the output position information.
In some embodiments, in the deep reinforcement model, a resource allocation action is determined according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the input position information, and the prediction information of the corresponding end-to-end network state; inputting the resource allocation action into a simulator to determine a network quality parameter as an environment reward; and adjusting the resource allocation action according to the environment reward, and determining the resource allocation action which enables the environment reward to be maximum as end-to-end resource allocation information corresponding to the output position information.
In some embodiments, determining the prediction information of the end-to-end network state corresponding to the location information according to the end-to-end historical network state corresponding to the location information includes: and inputting the end-to-end historical network state corresponding to the position information into the long-time memory network to obtain the prediction information of the end-to-end network state corresponding to the output position information.
According to another embodiment of the present disclosure, an apparatus for allocating network resources is provided, which includes: the position determining module is used for determining position information according to the intention information of the user; the intention translation module is used for determining the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information according to the intention information; the prediction module is used for determining the prediction information of the end-to-end network state corresponding to the position information according to the end-to-end historical network state corresponding to the position information; and the decision module is used for determining end-to-end resource allocation information corresponding to the position information according to the enhancement information of each end-to-end network element corresponding to the position information, the enhancement information of the link and the corresponding prediction information of the end-to-end network state.
In some embodiments, the willingness translation module is configured to, if the user is a first user, input willingness information into a first sub-model of the willingness translation model, and output network management willingness information corresponding to the willingness information; and inputting the network management intention information into a second sub-model of the intention translation model, and outputting the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information.
In some embodiments, the intention translation module is configured to, if the user is a second user, input intention information into a second sub-model of the intention translation model, and output enhancement information of each end-to-end network element and enhancement information of a link corresponding to the location information.
In some embodiments, the willingness translation module is configured to perform named entity recognition on the willingness information in the first submodel, and determine named entities of respective participles in the willingness information; and converting each participle into a corresponding preset network management word according to the named entity of each participle and a preset translation dictionary, thereby generating network management intention information.
In some embodiments, the decision module is configured to input, to the deep reinforcement learning model, end-to-end enhancement information of each network element and link corresponding to the location information, and prediction information of a corresponding end-to-end network state, to obtain end-to-end resource allocation information corresponding to the output location information.
In some embodiments, the decision module is configured to determine, in the deep reinforcement model, a resource allocation action according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the input location information, and the prediction information of the corresponding end-to-end network state; inputting the resource allocation action into a simulator to determine a network quality parameter as an environment reward; and adjusting the resource allocation action according to the environment reward, and determining the resource allocation action which enables the environment reward to be maximum as end-to-end resource allocation information corresponding to the output position information.
In some embodiments, the prediction module is configured to input the end-to-end historical network state corresponding to the location information into the long-term memory network, and obtain the prediction information of the end-to-end network state corresponding to the output location information.
According to still other embodiments of the present disclosure, an apparatus for allocating network resources is provided, which includes: a processor; and a memory coupled to the processor for storing instructions that, when executed by the processor, cause the processor to perform the method for network resource deployment of any of the foregoing embodiments.
According to still further embodiments of the present disclosure, there is provided a computer-readable non-transitory storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method for network resource allocation of any of the foregoing embodiments.
According to the method and the device, the enhancement information of each end-to-end network element and the enhancement information of a link at a corresponding position are determined according to the intention information of a user, the prediction information of the end-to-end network state at the corresponding position is determined according to the historical network state, and then the end-to-end resource allocation information is determined according to the enhancement information of each network element and the enhancement information of the link and the prediction information of the end-to-end network state. According to the resource allocation method and the resource allocation device, the resources are allocated according to the wishes of the users, meanwhile, the historical network states are predicted, and the resources are allocated according to the prediction information, so that the resource allocation strategy not only meets the wishes of the users, but also can meet the change of the network states, and the resource allocation accuracy is improved.
Other features of the present disclosure and advantages thereof will become apparent from the following detailed description of exemplary embodiments thereof, which proceeds with reference to the accompanying drawings.
Drawings
In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present disclosure, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 illustrates a flow diagram of a method for network resource deployment in accordance with some embodiments of the present disclosure.
Fig. 2 is a flow chart illustrating a network resource allocation method according to another embodiment of the disclosure.
Fig. 3 illustrates a schematic structure diagram of a network resource allocation apparatus according to some embodiments of the present disclosure.
Fig. 4 is a schematic structural diagram of a network resource allocating apparatus according to another embodiment of the present disclosure.
Fig. 5 is a schematic structural diagram of a network resource allocating apparatus according to still other embodiments of the disclosure.
Detailed Description
The technical solutions in the embodiments of the present disclosure will be described clearly and completely with reference to the drawings in the embodiments of the present disclosure, and it is obvious that the embodiments described are only some embodiments of the present disclosure, rather than all embodiments. The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the disclosure, its application, or uses. All other embodiments, which can be derived by a person skilled in the art from the embodiments disclosed herein without making any creative effort, shall fall within the protection scope of the present disclosure.
The present disclosure provides a method for allocating network resources, which can meet the user's will and improve the accuracy of resource allocation, and some embodiments of the method for allocating network resources of the present disclosure are described below with reference to fig. 1.
Fig. 1 is a flow chart of some embodiments of a method for allocating network resources according to the present disclosure. As shown in fig. 1, the method of this embodiment includes: steps S102 to S108.
In step S102, location information is determined according to user' S intention information.
The intention information may be input by the user through the terminal, for example, by voice, video, text, etc. For voice or video, etc., it can be converted into wishlist text by the existing technology. The user-entered desire information needs to contain location-related information. For example, a named entity recognition method may be employed to extract location information from user's intention information. The named entity recognition method may adopt the prior art, and is not described herein again. For example, the intention information of the user is "new population 50 ten thousand in beijing", and the extracted location information is "beijing". For another example, the user intention information is "cell a load exceeds the threshold value 20%", and the extracted position information is "cell a". The willingness information may include location-related information and specific demand information, which relates to the addition and scheduling of resources.
In step S104, the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information are determined according to the willingness information.
In some embodiments, determining end-to-end network elements from the location information comprises: each network element in the access network, the bearer network and the core network. For example, the network element of the access network includes a base station and the like, and a specific network element may be known according to actual deployment. According to the position information, the access network element corresponding to the position information can be determined, and then the network elements of each access network connected to the core network device and each bearer network passed by the link are determined. If the willingness information includes service category information, such as AR/VR/MR (augmented reality/virtual reality/mixed reality), cloud office, cloud game, etc., for example, the location information and the service category information may be combined to determine end-to-end network elements related to the service category.
And further, inputting the willingness information into a machine learning model to obtain the output enhanced information of each network element and the enhanced information of the link. The enhanced information of each network element, for example, increases the number of corresponding devices for each network element, or the capability enhanced information of each network element. The number of corresponding devices is increased, for example, by 10 base stations. The capability enhancement information of the network element, for example, the enhancement of the port capability, for example, the uplink port of an Edge Router (Edge Router) device generally adopts a 50GE link, and the capacity can be expanded to n × 50GE as required. The enhancement information of the link includes, for example, an increase in the number of links between the respective network elements.
And inputting the machine learning model to be trained by using the intention information marked with the corresponding enhancement information of each network element and the enhancement information of the link as training samples, calculating a loss function according to the output result and the corresponding mark, and adjusting the parameters of the machine learning model until a convergence condition is met to obtain the trained machine learning model. The machine learning model is, for example, a neural network model. After the willingness information is input into the machine learning model, the willingness information can be firstly converted into corresponding feature vectors, and the output enhancement information of each network element and the enhancement information of the link are further determined according to the feature vectors. And then, inputting a new willingness information sample into the trained machine learning model, so that the enhanced information of each network element and the enhanced information of the link can be directly output.
Different types of users can cause different forms of input intention information, for example, intention information input by ordinary users is more spoken, and intention information input by network operation and maintenance personnel contains professional terms for network management. Therefore, different interfaces can be set for different users to acquire the willingness information, so that the willingness information is processed differently. In some embodiments, in a case that the user is a first user, inputting the willingness information into a first sub-model of the willingness translation model, and outputting network management willingness information corresponding to the willingness information; and inputting the network management intention information into a second sub-model of the intention translation model, and outputting the enhanced information of each end-to-end network element and the enhanced information of the link corresponding to the position information. The first user is, for example, a general user, a non-network operation and maintenance person.
And under the condition that the user is a second user, inputting the intention information into a second sub-model of the intention translation model, and outputting the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information. The second user is, for example, a network operation and maintenance person, and the intention information of the second user is network management intention information, and may be directly input into the second submodel.
Furthermore, for example, in the first submodel, named entity recognition is performed on the willingness information, and named entities of respective participles in the willingness information are determined. And converting each participle into a corresponding preset network management word according to the named entity of each participle and a preset translation dictionary, thereby generating network management intention information. For example, "new population" may translate to "network expansion," etc. Or, for example, in the first submodel, the intention information is segmented, each segmented word is semantically recognized, a pre-translated word close to the segmented word semantics in the pre-translated dictionary is determined, and then the pre-network management word corresponding to the pre-translated word is determined. The preset transfer dictionary comprises preset network management words and pre-translation words corresponding to the preset network management words.
The second submodel is, for example, a neural network model, and may use network management intention information labeled with corresponding enhancement information of each network element and enhancement information of a link as a training sample, input the second submodel to be trained, calculate a loss function according to an output result and a corresponding label, adjust parameters of the second submodel until a convergence condition is satisfied, and obtain the trained second submodel. When the method is applied, the network management intention information is directly input into the second submodel, and the enhanced information of each network element and the enhanced information of the link can be obtained.
In some embodiments, the network elements from end to end are determined according to the location information, or the network elements from end to end are determined according to the location information and the service category information. Further, according to the service increment key words in the wish information, the corresponding preset template is searched, and the enhanced information of each network element and the enhanced information of the link are determined. The business increment key is, for example, 50 million people, 20% load or 500G flow, and the business increment key can be identified by using an algorithm such as semantic identification. The preset template comprises a service increment range, corresponding enhancement information of each network element and corresponding enhancement information of a link, the service increment range corresponding to the service increment is determined according to the service increment key words, and further the enhancement information of each network element and the enhancement information of the link can be directly determined. For example, the service increment range corresponding to 50 million increase of population is 45 to 55 million increase of users, and the enhancement information of each network element such as 100 stations and the like and the enhancement information of the link are added corresponding to the base station.
In step S106, the prediction information of the end-to-end network state corresponding to the location information is determined according to the end-to-end historical network state corresponding to the location information.
End-to-end network states include, for example: the traffic information between the network elements may further include at least one item of information such as the number of connected users and the resource utilization rate of each network element. And under the condition that the intention information of the user comprises the service type information, the end-to-end network state is the end-to-end network state corresponding to the service type. For example, for a VR service, the prediction information of the network state related to the end-to-end VR service corresponding to the location information may be determined according to the historical network state related to the end-to-end VR service corresponding to the location information.
In some embodiments, the end-to-end historical network state corresponding to the location information is input to a long-short-term memory network (LSTM), and prediction information of the end-to-end network state corresponding to the output location information is obtained. The LSTM network is an existing model and will not be described herein.
In step S108, end-to-end resource allocation information corresponding to the location information is determined according to the end-to-end enhancement information of each network element and the link enhancement information corresponding to the location information, and the corresponding end-to-end prediction information of the network state.
The end-to-end resource allocation information may include: at least one of the number of devices of each network element, the capability enhancement information of each network element, the scheduling of related resources between each network element, the increase of links, and the change of the amount of related resources. The resources include, for example, spectrum resources, bandwidth resources, virtual resources, cloud resources, and the like. For example, the end-to-end resource allocation information includes a proposal for building and networking a DC cloud resource pool, a proposal for expanding the wireless site scale and deployment optimization, a deployment strategy of MEC (mobile edge computing) or UPF (user plane function), a deployment strategy of Virtual Network Function (VNF) of core network, a proposal for initial, medium and target layout of ip ran network and OTN network of local network, and the like.
In some embodiments, the end-to-end enhancement information of each network element and the link corresponding to the location information, and the corresponding end-to-end prediction information of the network state are input into a deep reinforcement learning model, so as to obtain end-to-end resource allocation information corresponding to the output location information.
Further, in the deep reinforcement model, a resource allocation action is determined according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the input position information and the prediction information of the corresponding end-to-end network state. And inputting the resource allocation action into a simulator to determine a network quality parameter as an environment reward. And adjusting the resource allocation action according to the environment reward, and determining the resource allocation action which enables the environment reward to be maximum as end-to-end resource allocation information corresponding to the output position information. Network quality parameters include, for example: at least one of network load, link congestion condition, user QoS, network resources, and utilization rates of CPU and memory in the DC cloud resource pool.
GIS geographic information (such as wireless base stations, comprehensive offices, edge DCs and core network DCs, and the condition of optical cable routing), network operation data (network access flow in different directions, flow proportion of different service types, link bandwidth and flow statistics), local basic resources (local positions, machine room space and power capacity) and the like can be input into the deep reinforcement learning model and used for the simulator. The deep reinforcement learning model can also consider factors such as load balancing, redundant disaster tolerance and the necessity of backup.
In some embodiments, incremental information for the end-to-end network state is determined based on the predicted information for the end-to-end network state and the current network state. And determining the prediction enhancement information of each end-to-end network element and the prediction enhancement information of the link corresponding to the position information according to the increment information. For example, a traffic increase level is determined according to traffic increase information between end-to-end network elements, and then corresponding preset resource increase information is determined according to the traffic increase level between the network elements.
Further, according to the end-to-end predicted enhancement information of each network element and the predicted enhancement information of the link corresponding to the position information, the enhancement information of each network element and the enhancement information of the link, and a preset scheduling strategy, the end-to-end resource allocation information corresponding to the position information is determined. The preset scheduling policy includes, for example: and preferentially scheduling resources from the adjacent equipment, preferentially performing at least one of equipment capability enhancement and the like. The prediction enhancement information and the prediction enhancement information of the link of each network element, and the enhancement information of the link of each network element may include enhancement information of specific resources, for example, at least one of bandwidth resources, spectrum resources, virtual resources, and cloud resources, and the prediction enhancement information of the link of each network element, and the enhancement information of the link of each network element are weighted to obtain the overall enhancement information and the overall enhancement information of the link of each network element. Scheduling strategies such as scheduling resources from adjacent equipment and preferentially performing equipment capacity enhancement can be preferentially scheduled from the adjacent equipment and preferentially performing equipment capacity enhancement, and corresponding resources are increased by adding corresponding equipment under the condition that the resources and the equipment capacity enhancement of the adjacent equipment are not enough to meet the integral enhancement information of each network element and the integral enhancement information of a link.
In the method of the embodiment, the enhancement information of each network element and the enhancement information of the link from end to end at the corresponding position are determined according to the intention information of the user, the prediction information of the network state from end to end at the corresponding position is determined according to the historical network state, and the resource allocation information from end to end is determined according to the enhancement information of each network element and the enhancement information of the link and the prediction information of the network state from end to end. The method of the embodiment not only carries out resource allocation according to the will of the user, but also carries out prediction by referring to the historical network state, and carries out resource allocation according to the prediction information, so that the resource allocation strategy not only accords with the user's will, but also can meet the change of the network state, and the accuracy of resource allocation is improved.
The method is suitable for 5G application scenes, is suitable for various specific service scenes, such as AR/VR/MR, cloud office, cloud game and other services, and can determine resource allocation information corresponding to the services according to prediction and willingness information of related network states of the emerging services. The method disclosed by the invention not only predicts based on historical data, considers the requirements of business willingness, and can input more comprehensive information according to the characteristics of future network clouding trend, shared and isolated resources among 5G network slices and the factors of network element level redundancy disaster tolerance and backup. The output end-to-end infrastructure resource allocation strategy can be applied to the scene of future cloud network fusion construction, the human input of network planning can be reduced, and the resource deployment efficiency is improved.
Other embodiments of the network resource allocation method of the present disclosure are described below with reference to fig. 2.
Fig. 2 is a flowchart of another embodiment of a network resource allocation method according to the present disclosure. As shown in fig. 2, the method of this embodiment includes: steps S202 to S210.
In step S202, intention information of the user is acquired.
The intention information of the first user and the second user can be obtained through different interfaces, and further, the intention information is analyzed by referring to different methods in the foregoing embodiments.
In step S204, location information and service type, and enhancement information of each end-to-end network element and enhancement information of a link corresponding to the location information and the service type are determined according to the intention information of the user by using the intention translation model.
The service type is, for example, the existing service type in 2G/3G/4G, or the new service type in 5G.
In step S206, the prediction model is used to determine the prediction information of the end-to-end network state corresponding to the location information and the service type based on the end-to-end historical network state corresponding to the location information and the service type.
In step S208, the decision model is used to determine end-to-end resource allocation information corresponding to the location information and the service type according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information and the service type, and the prediction information of the corresponding end-to-end network state.
In step S210, at least one of the willingness translation model, the prediction model and the decision model is periodically updated according to the resource deployment execution result.
Performing actual resource allocation according to the resource allocation information in step S208, to obtain whether the resource allocation information is suitable, for example, after 10 base stations are added, it is found that the resource allocation information is still insufficient, and 12 base stations are needed to meet the requirement, or the difference between the predicted traffic and the actual traffic is large, and finally the allocation method meeting the requirement is used as a resource allocation execution result, and the training data is regenerated to update at least one of the will translation model, the prediction model, and the decision model.
Some application examples of the present disclosure are described below.
In an application example, the intention information of the first user, for example, a general user, triggers the infrastructure resource allocation process. For example, if a certain city is listed as an emerging developing city and is about 50 thousands of new population to be admitted, the first submodel of the willingness transfer model translates willingness information into network management willingness information related to network capacity expansion, and then the enhanced information of the scale of the gNB, the scale of the IPRAN (IP radio access network) and the scale of the DC cloud resource pool server is output by the second submodel. And the prediction model predicts the service flow and the resource demand according to the network historical data. Based on the output of the willingness transfer model and the prediction model, an end-to-end resource allocation strategy is output by using the decision model, such as the construction scale of the core DC and the edge DC, the scale and the distribution of the wireless gNB, the VNFs deployment strategy of the core network control plane, the deployment strategy of the core UPF and the edge UPF, the network layout of the local network IPRAN and the like. And the resource allocation strategy is issued to the corresponding arrangement management and control layer. And arranging an execution strategy of the management and control layer, and scheduling and deploying corresponding infrastructure resources. And the arranging and controlling layer periodically feeds back the resource allocation strategy execution result to the perception engine so as to optimize the model. The perception engine can be used for perceiving historical or current network states, loads of different service areas at different moments, congestion conditions of different links, qoS (quality of service) experience of users, utilization conditions of CPUs (central processing units) and memories in the DC cloud resource pool and the like, and acquiring and managing data required by each model.
In application example two, the willingness information of the second user, such as a network administrator, triggers the infrastructure resource deployment process. The intention information is, for example, a situation that the load of the cell a exceeds a threshold, and the second submodel of the intention transfer model outputs enhancement information of each end-to-end network element and enhancement information of the link corresponding to the cell a according to the intention information. And the prediction model predicts the service flow and the resource demand according to the network historical data. Based on the outputs of the willingness transfer model and the prediction model, an end-to-end resource allocation strategy is output by using the decision model, for example, the bandwidth of a certain link is increased/reduced, the CPU allocation of the edge DC cloud resource pool is rearranged, or resources are scheduled from other cells, and the like. And the resource allocation strategy is issued to the corresponding arrangement management and control layer. And arranging an execution strategy of the management and control layer, and scheduling and deploying corresponding infrastructure resources. The arrangement management and control layer periodically feeds back the execution result of the resource allocation strategy to the perception engine so as to carry out model optimization.
The present disclosure also provides a device for allocating network resources, which is described below with reference to fig. 3.
Fig. 3 is a block diagram of some embodiments of a network resource allocating apparatus according to the present disclosure. As shown in fig. 3, the apparatus 30 of this embodiment includes: a location determination module 310, a intent translation module 320, a prediction module 330, and a decision module 340.
A location determining module 310, configured to determine location information according to the intention information of the user.
And the intention translation module 320 is configured to determine, according to the intention information, the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information.
In some embodiments, the willingness translation module 320 is configured to, if the user is a first user, input willingness information into a first sub-model of the willingness translation model, and output network management willingness information corresponding to the willingness information; and inputting the network management intention information into a second sub-model of the intention translation model, and outputting the enhanced information of each end-to-end network element and the enhanced information of the link corresponding to the position information.
In some embodiments, the willingness translation module 320 is configured to, if the user is a second user, input willingness information into a second sub-model of the willingness translation model, and output enhanced information of each end-to-end network element and enhanced information of a link corresponding to the location information.
In some embodiments, the willingness translation module 320 is configured to perform named entity recognition on willingness information in the first sub-model, and determine named entities of respective participles in the willingness information; and converting each participle into a corresponding preset network management word according to the named entity of each participle and a preset translation dictionary, thereby generating network management intention information.
The prediction module 330 is configured to determine prediction information of an end-to-end network state corresponding to the location information according to an end-to-end historical network state corresponding to the location information.
In some embodiments, the prediction module 330 is configured to input the end-to-end historical network state corresponding to the location information into the long-term memory network, so as to obtain the prediction information of the end-to-end network state corresponding to the output location information.
The decision module 340 is configured to determine end-to-end resource allocation information corresponding to the location information according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information, and the prediction information of the corresponding end-to-end network state.
In some embodiments, the decision module 340 is configured to input, to the deep reinforcement learning model, end-to-end enhancement information of each network element and link corresponding to the location information, and prediction information of a corresponding end-to-end network state, so as to obtain end-to-end resource allocation information corresponding to the output location information.
In some embodiments, the decision module 340 is configured to determine, in the deep-drawn model, a resource allocation action according to the enhanced information of each end-to-end network element and the enhanced information of the link corresponding to the input location information, and the prediction information of the corresponding end-to-end network state; inputting the resource allocation action into a simulator to determine a network quality parameter as an environment reward; and adjusting the resource allocation action according to the environment reward, and determining the resource allocation action which enables the environment reward to be maximum as end-to-end resource allocation information corresponding to the output position information.
The network resource allocating apparatus in the embodiments of the present disclosure may be implemented by various computing devices or computer systems, and are described below with reference to fig. 4 and 5.
Fig. 4 is a block diagram of some embodiments of a network resource allocating apparatus according to the present disclosure. As shown in fig. 4, the apparatus 40 of this embodiment includes: a memory 410 and a processor 420 coupled to the memory 410, the processor 420 configured to execute the method for deploying network resources in any of the embodiments of the present disclosure based on instructions stored in the memory 410.
Fig. 5 is a block diagram of another embodiment of a network resource allocating apparatus according to the present disclosure. As shown in fig. 5, the apparatus 50 of this embodiment includes: memory 510 and processor 520 are similar to memory 410 and processor 420, respectively. An input output interface 530, a network interface 540, a storage interface 550, and the like may also be included. These interfaces 530, 540, 550 and the connections between the memory 510 and the processor 520 may be, for example, via a bus 560. The input/output interface 530 provides a connection interface for input/output devices such as a display, a mouse, a keyboard, and a touch screen. The network interface 540 provides a connection interface for various networking devices, such as a database server or a cloud storage server. The storage interface 550 provides a connection interface for external storage devices such as an SD card and a usb disk.
As will be appreciated by one skilled in the art, embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable non-transitory storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only exemplary of the present disclosure and is not intended to limit the present disclosure, so that any modification, equivalent replacement, or improvement made within the spirit and principle of the present disclosure should be included in the scope of the present disclosure.
Claims (14)
1. A network resource allocation method comprises the following steps:
determining position information according to the intention information of the user;
determining the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information according to the intention information;
determining the prediction information of the end-to-end network state corresponding to the position information according to the end-to-end historical network state corresponding to the position information;
determining end-to-end resource allocation information corresponding to the position information according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information and the prediction information of the corresponding end-to-end network state;
wherein the determining the prediction information of the end-to-end network state corresponding to the location information according to the end-to-end historical network state corresponding to the location information comprises:
and inputting the end-to-end historical network state corresponding to the position information into a long-time memory network to obtain the output end-to-end network state prediction information corresponding to the position information.
2. The formulation process according to claim 1,
the determining, according to the willingness information, the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information includes:
when the user is a first user, inputting the intention information into a first sub-model of an intention translation model, and outputting network management intention information corresponding to the intention information; and inputting the network management intention information into a second submodel of the intention translation model, and outputting the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information.
3. The formulation process according to claim 1,
the determining, according to the intention information, the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information includes:
and under the condition that the user is a second user, inputting the willingness information into a second sub-model of the willingness translation model, and outputting the enhanced information of each end-to-end network element and the enhanced information of the link corresponding to the position information.
4. The formulation process according to claim 2,
the inputting the willingness information into a first submodel of a willingness translation model, and the outputting the network management willingness information corresponding to the willingness information comprises:
in the first submodel, carrying out named entity recognition on the willingness information, and determining named entities of all participles in the willingness information;
and converting each participle into a corresponding preset network management word according to the named entity of each participle and a preset translation dictionary, thereby generating network management intention information.
5. The formulation process according to claim 1,
the determining, according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information and the prediction information of the corresponding end-to-end network state, the end-to-end resource allocation information corresponding to the location information includes:
and inputting the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information and the prediction information of the corresponding end-to-end network state into a deep reinforcement learning model to obtain the output end-to-end resource allocation information corresponding to the position information.
6. The formulation process according to claim 5,
in the deep reinforcement model, determining a resource allocation action according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the input position information and the prediction information of the corresponding end-to-end network state;
inputting the resource allocation action into a simulator to determine a network quality parameter as an environment reward;
and adjusting the resource allocation action according to the environment reward, and determining the resource allocation action which enables the environment reward to be maximum as end-to-end resource allocation information corresponding to the output position information.
7. An apparatus for allocating network resources, comprising:
the position determining module is used for determining position information according to the intention information of the user;
a wish translation module, configured to determine, according to the wish information, enhancement information of each end-to-end network element and enhancement information of a link corresponding to the location information;
the prediction module is used for determining the prediction information of the end-to-end network state corresponding to the position information according to the end-to-end historical network state corresponding to the position information;
a decision module, configured to determine end-to-end resource allocation information corresponding to the location information according to the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the location information, and the prediction information of the corresponding end-to-end network state;
the prediction module is used for inputting the end-to-end historical network state corresponding to the position information into a long-time memory network to obtain the output prediction information of the end-to-end network state corresponding to the position information.
8. The blending device of claim 7,
the system comprises a willingness translation module, a network management willingness information acquisition module and a willingness information output module, wherein the willingness translation module is used for inputting the willingness information into a first sub-model of a willingness translation model and outputting the network management willingness information corresponding to the willingness information under the condition that the user is a first user; and inputting the network management intention information into a second submodel of the intention translation model, and outputting the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information.
9. The blending device of claim 7,
and the intention translation module is used for inputting the intention information into a second sub-model of the intention translation model and outputting the enhancement information of each end-to-end network element and the enhancement information of the link corresponding to the position information under the condition that the user is a second user.
10. The blending apparatus of claim 8,
the intention translation module is used for carrying out named entity recognition on the intention information in the first submodel and determining named entities of all participles in the intention information; and converting each participle into a corresponding preset network management word according to the named entity of each participle and a preset translation dictionary, thereby generating network management intention information.
11. The blending device of claim 7,
the decision module is configured to input the end-to-end enhancement information of each network element and the link corresponding to the location information, and the corresponding end-to-end prediction information of the network state into a deep reinforcement learning model, so as to obtain end-to-end resource allocation information corresponding to the output location information.
12. The blending apparatus of claim 11,
the decision module is used for determining resource allocation actions according to the enhanced information of each end-to-end network element and the enhanced information of the link corresponding to the input position information and the corresponding end-to-end network state prediction information in the deep reinforcement model; inputting the resource allocation action into a simulator to determine a network quality parameter as an environment reward; and adjusting the resource allocation action according to the environment reward, and determining the resource allocation action which enables the environment reward to be maximum as end-to-end resource allocation information corresponding to the output position information.
13. An apparatus for allocating network resources, comprising:
a processor; and
a memory coupled to the processor for storing instructions that, when executed by the processor, cause the processor to perform the method of provisioning network resources of any of claims 1-6.
14. A computer-readable non-transitory storage medium having a computer program stored thereon, wherein the program, when executed by a processor, performs the steps of the method of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910962559.5A CN112738723B (en) | 2019-10-11 | 2019-10-11 | Network resource allocation method and device and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910962559.5A CN112738723B (en) | 2019-10-11 | 2019-10-11 | Network resource allocation method and device and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112738723A CN112738723A (en) | 2021-04-30 |
| CN112738723B true CN112738723B (en) | 2023-04-07 |
Family
ID=75588717
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910962559.5A Active CN112738723B (en) | 2019-10-11 | 2019-10-11 | Network resource allocation method and device and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112738723B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113891359B (en) * | 2021-10-11 | 2023-06-20 | 中国联合网络通信集团有限公司 | Data processing method, device and storage medium |
| CN117082009B (en) * | 2023-10-16 | 2024-02-27 | 天翼安全科技有限公司 | Cloud resource management method and management system based on software definition security |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106658405A (en) * | 2015-11-02 | 2017-05-10 | 维亚威解决方案英国有限公司 | Enhancing network topology information for a self-organizing network |
| CN109245916A (en) * | 2018-08-15 | 2019-01-18 | 西安电子科技大学 | A kind of the cloud access net system and method for intention driving |
| CN110198237A (en) * | 2019-05-27 | 2019-09-03 | 北京邮电大学 | A kind of wireless configuration method for being intended to driving network |
| US10409551B1 (en) * | 2016-06-21 | 2019-09-10 | Amazon Technologies, Inc. | Voice-driven monitoring of resources in a service provider network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10582393B2 (en) * | 2017-12-04 | 2020-03-03 | Verizon Patent And Licensing Inc. | Architecture for network slice deployment based on network resource utilization |
-
2019
- 2019-10-11 CN CN201910962559.5A patent/CN112738723B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106658405A (en) * | 2015-11-02 | 2017-05-10 | 维亚威解决方案英国有限公司 | Enhancing network topology information for a self-organizing network |
| US10409551B1 (en) * | 2016-06-21 | 2019-09-10 | Amazon Technologies, Inc. | Voice-driven monitoring of resources in a service provider network |
| CN109245916A (en) * | 2018-08-15 | 2019-01-18 | 西安电子科技大学 | A kind of the cloud access net system and method for intention driving |
| CN110198237A (en) * | 2019-05-27 | 2019-09-03 | 北京邮电大学 | A kind of wireless configuration method for being intended to driving network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112738723A (en) | 2021-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2020258920A1 (en) | Network slice resource management method and apparatus | |
| EP3573367B1 (en) | Resource allocation for ran slicing | |
| US10652360B2 (en) | Access scheduling method and apparatus for terminal, and computer storage medium | |
| CN107708214B (en) | Network utility maximization virtual resource allocation method based on Lyapunov | |
| CN113037786B (en) | Intelligent computing power scheduling method, device and system | |
| CN110287035B (en) | request scheduling method, device and equipment for hybrid edge computing and storage medium | |
| EP3089032A1 (en) | Management system, overall management node, and management method | |
| CN112738723B (en) | Network resource allocation method and device and computer readable storage medium | |
| EP3793206A1 (en) | Physical optical network virtualization mapping method and apparatus, and controller and storage medium | |
| CN113348651A (en) | Dynamic inter-cloud placement of sliced virtual network functions | |
| CN112887142B (en) | Method for generating web application slice in virtualized wireless access edge cloud | |
| CN112073237B (en) | Large-scale target network construction method in cloud edge architecture | |
| CN107360235B (en) | Task migration method based on reliability grading | |
| CN112925634A (en) | Heterogeneous resource scheduling method and system | |
| CN103973780A (en) | Resource allocation and overhead optimization method for cloud video | |
| CN109495565B (en) | High-concurrency service request processing method and device based on distributed ubiquitous computing | |
| Globa et al. | Architecture and operation algorithms of mobile core network with virtualization | |
| Saravanan et al. | Improving map reduce task scheduling and micro-partitioning mechanism for mobile cloud multimedia services | |
| CN106874108B (en) | Technology for minimizing number of micro clouds in mobile cloud computing | |
| CN102186182A (en) | Comprehensive network plan management system and method | |
| WO2018120961A1 (en) | Traffic simulation method and apparatus | |
| US11588882B2 (en) | Method, electronic device, and computer program product for application migration | |
| Chousainov et al. | Performance evaluation in single or multi-cluster C-RAN supporting quasi-random traffic | |
| CN111092755B (en) | Edge service migration simulation method based on resource occupation | |
| Hanczewski et al. | The analytical model of complex non-full-availability system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |