CN115209431A - Triggering method, device, equipment and computer storage medium - Google Patents
Triggering method, device, equipment and computer storage medium Download PDFInfo
- Publication number
- CN115209431A CN115209431A CN202110397103.6A CN202110397103A CN115209431A CN 115209431 A CN115209431 A CN 115209431A CN 202110397103 A CN202110397103 A CN 202110397103A CN 115209431 A CN115209431 A CN 115209431A
- Authority
- CN
- China
- Prior art keywords
- node
- nodes
- slice
- current
- reconstruction
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000003860 storage Methods 0.000 title claims abstract description 19
- 239000013598 vector Substances 0.000 claims abstract description 81
- 238000005070 sampling Methods 0.000 claims description 45
- 238000012549 training Methods 0.000 claims description 16
- 238000013508 migration Methods 0.000 claims description 11
- 230000005012 migration Effects 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 238000004891 communication Methods 0.000 claims description 9
- 230000002596 correlated effect Effects 0.000 claims description 6
- 230000001419 dependent effect Effects 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 16
- 230000000875 corresponding effect Effects 0.000 description 11
- 230000001960 triggered effect Effects 0.000 description 9
- 238000012544 monitoring process Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 238000013507 mapping Methods 0.000 description 6
- 239000011159 matrix material Substances 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- 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
- H04W16/22—Traffic simulation tools or models
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/02—Arrangements for optimising operational condition
Abstract
The invention discloses a triggering method, which comprises the following steps: and acquiring a slice example of the network slice, inputting the slice example into the trained Node2Vec model, outputting a vector of a Node in the slice example, and determining whether to trigger the reconstruction of the slice example according to the vector of the Node. The embodiment of the invention also discloses a triggering device, equipment and a computer storage medium, which improve the decision efficiency of whether the network slice needs to be reconstructed or not and further improve the reconstruction efficiency of the network slice for reconstruction.
Description
Technical Field
The present invention relates to a reconfiguration triggering technology for network slices, and in particular, to a triggering method, apparatus, device, and computer storage medium.
Background
At present, network slicing is an important content in 5G construction, and the technology meets the differentiated requirements of the vertical industry on network services by constructing a plurality of logically independent proprietary networks on the same physical basic platform. Different from the traditional single network management mode, the network slicing technology provides a larger selection space for customizing individual requirements, and provides a more convenient, efficient, safe and low-cost operation and maintenance scheme for operators to bear diversified network services.
The third Generation Partnership Project (3 gpp,3rd Generation Partnership Project) specifies four mandatory phases of the slice lifecycle, namely a preparation phase, an instantiation, configuration, activation phase, a runtime phase, and a logoff phase. In the preparation stage, all types of requirements of services to be carried by the bearer network need to be specified, and different slicing templates are customized according to different types of requirements. In the instantiation, configuration and activation stages, the service requirement needs to be converted into the network performance requirement, a corresponding slice template is selected, the association of the virtual topology and the physical bearing topology is realized, the corresponding network resource is configured, and the instantiation of the slice based on the template is realized. In the runtime phase, traffic needs to be carried onto instantiated slices. In the offline phase, instantiated slices that have been serviced need to be deleted and the relevant underlying resources need to be reclaimed.
In the operation stage, slice conditions can be monitored by means of an intelligent network operation and maintenance method, instantiated slices are reconstructed according to changes of business requirements, and self-adaptation of slice structures and resource to requirement changes is achieved, so that slice flexibility is improved, and service quality of a network is improved.
In the related art, the network slice is usually identified and calculated based on a matrix to trigger the reconstruction of the network slice, however, the method has a large calculation amount, so that the triggering method has high complexity, and the network slice has poor reconstruction triggering efficiency; therefore, the technical problem that the conventional network slice reconstruction triggering method is low in efficiency exists.
Disclosure of Invention
In view of this, the present invention provides a triggering method, an apparatus, a device, and a computer storage medium, so as to solve the technical problem in the prior art that the network slice reconfiguration triggering method has low efficiency.
The technical scheme of the invention is realized as follows:
in a first aspect, an embodiment of the present invention provides a triggering method, where the method includes:
acquiring a slice example of the network slice;
inputting the slice example into a trained Node2Vec model, and outputting to obtain a vector of a Node in the slice example; wherein the vector elements of the node include: a quantity representing a node structure and a quantity representing a node load;
determining whether to trigger reconstruction of the slice instance according to the vector of the node.
In the above method, the trained Node2Vec model is obtained as follows:
acquiring an initial node and a next hop node of the initial node from a slice example to be trained, determining the next hop node as a current node, and setting an initial value of a sampling frequency i as 1;
acquiring adjacent nodes of a current node;
when i is smaller than the sampling step length, calculating the bias walk probability of the current node and the adjacent node according to a preset bias walk probability algorithm; the preset bias walk probability is positively correlated with the network performance parameter of the current service when the slice example is adopted at the current moment;
sorting the calculated offset walk probability from big to small, selecting the adjacent nodes which are arranged at the top M, determining the selected adjacent nodes as the current nodes, updating i to i +1, and returning to execute the adjacent nodes which obtain the current nodes; wherein M is a positive integer less than the number of adjacent nodes;
and when i is larger than or equal to the sampling step length, inputting at least two groups of Node sequences obtained by sampling into a preset Node2Vec model for training to obtain the trained Node2Vec model.
In the above method, the network performance parameter of the current service includes one or more of the following:
the average occupied bandwidth of the current service, the average message delay of the current service and the average packet loss rate of the current service. .
In the method, the bias migration probability pi from the current node to the adjacent node is calculated by adopting the following formula vx ：
π vx ＝α pq (t,x)·w vx
Where v denotes the current node, t denotes the neighbor node of the current node, W 0 Represents the average occupied bandwidth, T, of the current service at the current time 0 Average message delay representing the current service at the current time, E 0 Represents the average packet loss rate, k, of the current service at the current time W Is W 0 Is estimated by the level k T Is T 0 Estimated level of (k) E Is E 0 Estimated level of (d) tx Representing the shortest hop number required for node t to hop to node x.
In the above method, the determining whether to trigger reconstruction of the slice instance according to the vector of the node includes:
calculating the Euclidean distance between any two nodes in the slice example according to the vectors of the nodes;
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes meets the preset condition.
In the above method, when the euclidean distance between any two nodes satisfies a preset condition, triggering the reconstruction of the slice instance includes:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is smaller than the preset minimum Euclidean distance between the nodes.
In the above method, when the euclidean distance between any two nodes satisfies a preset condition, triggering the reconstruction of the slice instance includes:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is larger than the preset maximum Euclidean distance between the nodes.
In the above method, the method further comprises:
and when the Euclidean distance between any two nodes does not have a distance smaller than the preset minimum Euclidean distance between the nodes, and the Euclidean distance between any two nodes does not have a distance larger than the preset maximum Euclidean distance between the nodes, forbidding to trigger the reconstruction of the slicing example.
In a second aspect, the present invention provides a triggering device, the device comprising:
the acquisition module is used for acquiring a slice example of the network slice;
the processing module is used for inputting the slice example into the trained Node2Vec model and outputting the vector of the Node in the slice example; wherein the vector elements of the node include: a quantity representing a node structure and a quantity representing a node load;
and the triggering module is used for determining whether to trigger the reconstruction of the slice example according to the vector of the node.
In the above apparatus, the apparatus is further configured to:
obtaining the trained Node2Vec model by adopting the following method:
acquiring an initial node and a next hop node of the initial node from a slice example to be trained, determining the next hop node as a current node, and setting an initial value of a sampling frequency i as 1;
acquiring adjacent nodes of a current node;
when i is smaller than the sampling step length, calculating the bias migration probability of the current node and the adjacent node according to a preset bias migration probability algorithm; the preset bias migration probability is positively correlated with the network performance parameter of the current service when the slice example is adopted at the current moment;
sorting the calculated bias walk probability from big to small, selecting M adjacent nodes arranged at the top, determining the selected adjacent nodes as current nodes, updating i to i +1, and returning to execute the adjacent nodes for acquiring the current nodes; wherein M is a positive integer less than the number of adjacent nodes;
and when i is larger than or equal to the sampling step length, inputting at least two groups of Node sequences obtained by sampling into a preset Node2Vec model for training to obtain the trained Node2Vec model.
In the above apparatus, the network performance parameter of the current service comprises one or more of:
the average occupied bandwidth of the current service, the average message delay of the current service and the average packet loss rate of the current service.
In the above apparatus, the apparatus is further configured to: calculating to obtain the bias migration probability pi from the current node to the adjacent node by adopting the following formula vx ：
π vx ＝α pq (t,x)·w vx
Where v denotes the current node, t denotes the neighbor node of the current node, W 0 Represents the average occupied bandwidth, T, of the current service at the current time 0 Average message delay representing the current service at the current time, E 0 Average packet loss rate, k, representing the current traffic at the current time W Is W 0 Estimated level of (k) T Is T 0 Estimated level of (k) E Is E 0 Estimated level of d tx Representing the shortest hop number required for node t to hop to node x.
In the above apparatus, the triggering module is specifically configured to:
calculating the Euclidean distance between any two nodes in the slice example according to the vectors of the nodes;
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes meets the preset condition.
In the above apparatus, when the euclidean distance between any two nodes satisfies a preset condition, the triggering module triggers the reconstruction of the slice instance, including:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is smaller than the preset minimum Euclidean distance between the nodes.
In the above apparatus, when the euclidean distance between any two nodes satisfies a preset condition, the triggering module triggers the reconstruction of the slice instance, which includes:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is larger than the preset maximum Euclidean distance between the nodes.
In the above apparatus, the apparatus is further configured to:
and when the Euclidean distance between any two nodes does not have a distance smaller than a preset minimum Euclidean distance between the nodes, and the Euclidean distance between any two nodes does not have a distance larger than a preset maximum Euclidean distance between the nodes, forbidding to trigger the reconstruction of the slice example.
In a third aspect, an embodiment of the present invention further provides an apparatus, where the apparatus includes: the trigger device comprises a processor and a storage medium storing instructions executable by the processor, wherein the storage medium depends on the processor to execute operations through a communication bus, and when the instructions are executed by the processor, the trigger device executes the triggering method of one or more of the embodiments.
The embodiment of the invention provides a computer storage medium, which stores executable instructions, and when the executable instructions are executed by one or more processors, the processors execute the triggering method of one or more embodiments.
The invention provides a triggering method, a triggering device, triggering equipment and a computer storage medium, wherein the method comprises the following steps: obtaining a slice example of a network slice, inputting the slice example into a trained Node2Vec model, and outputting to obtain a vector of a Node in the slice example, wherein vector elements of the Node comprise: determining a quantity for representing a node structure and a quantity for representing a node load, and triggering reconfiguration of a slice instance according to a vector of a node; that is to say, in the present invention, after a slice instance of a network slice is obtained, the slice instance is input into a trained Node2Vec model, and a vector of a Node of the slice instance can be obtained, because the obtained vector includes a quantity for representing a Node structure and a Node load, and the structure of the slice and the load condition of the slice can be reflected based on the vector of the Node, then, in determining whether to trigger reconstruction of the slice instance according to the vector of the Node, because the vector of the Node has two dimensions, the complexity of determining whether to trigger reconstruction of the slice instance is reduced by using the vector of the two dimensions, thereby improving the decision efficiency of whether the network slice needs to be reconstructed, and further improving the reconstruction efficiency of the network slice for reconstruction.
Drawings
Fig. 1 is a schematic flow chart of an alternative triggering method in an embodiment of the present invention;
FIG. 2 is a network topology diagram of an example of a slice labeled with bias weights in the related art;
fig. 3 is a schematic flowchart of an example of an alternative triggering method according to an embodiment of the present invention;
FIG. 4 is a schematic structural diagram of an alternative triggering device in an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an alternative apparatus according to an embodiment of the present invention.
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
Example one
An embodiment of the present invention provides a triggering method, and fig. 1 is a schematic flow diagram of an optional triggering in an embodiment of the present invention, and as shown in fig. 1, the triggering method may include:
s101: acquiring a slice example of the network slice;
at present, for network slicing, after a slicing service is started, a slicing template matched with a current service requirement is selected from a slicing template library, a slicing instance is created according to the selected slicing template, then an online slicing instance bears the service requirement, and after the slicing instance is online, the slicing instance is represented based on a matrix to determine whether to trigger the reconstruction of the slicing instance; then, in determining whether to trigger the reconstruction of the slice instance based on the matrix, since the dimension of the matrix is high, the matrix is adopted to represent the slice information, so that the computational complexity in determining whether to trigger the reconstruction of the slice instance is increased, and the cost for monitoring the slice instance is increased.
In order to reduce the complexity of determining whether to trigger the reconstruction of the slice instance so as to reduce the cost of monitoring the slice instance, the embodiment of the invention provides a triggering method for determining whether to trigger the reconstruction method of the slice instance.
It should be noted that the triggering method provided by the embodiment of the present invention is deployed after the slicing instance is online operated, that is, at the runtime stage, that is, after the slicing service is started, first a slicing template matching the current service requirement is selected from a slicing template library, then a slicing instance is created according to the selected slicing template, then the online slicing instance carries the service requirement, and after the slicing instance is online, the slicing instance is obtained.
The current service of the slicing example may be a shopping application, a mobile communication service, or a communication application, and this is not specifically limited in this embodiment of the present invention.
S102: inputting the slice example into a trained Node2Vec model, and outputting to obtain a vector of a Node in the slice example;
in the step of obtaining the slice example, a topological graph of the slice example is obtained, wherein the topological graph of the slice example comprises nodes in the slice example and connection relations between the nodes, and then the slice example is input into a trained Node2Vec model, so that vectors of the nodes in the slice example can be output and obtained.
The trained Node2Vec model is obtained by sampling to obtain sample data and then training the Node2Vec model by using the sample data, wherein the Node2Vec model is a model used for generating Node vectors in a network, the input of the model is a network structure, and the output of the model is a vector of each Node.
In the method, the slicing problem is introduced into the Node2Vec model, so that the slicing information is reduced from N dimension to 2 dimension, the space complexity of slice information representation and storage is effectively reduced, the slice management monitoring cost is reduced, and the network operation and maintenance benefit is improved.
Wherein the vector elements of the nodes include: quantities representing node structure and quantities representing node load; in other words, the obtained node vector can reflect the structure of the node and the load condition of the node, so that whether the reconstruction of the slice example is triggered is determined according to the node vector, and whether the reconstruction of the slice example is triggered is mainly determined according to the structure of the node and the load condition of the node.
Further, in order to obtain the trained Node2Vec model, in an alternative embodiment, the trained Node2Vec model is obtained as follows:
acquiring an initial node and a next hop node of the initial node from a slice example to be trained, determining the next hop node as a current node, and setting an initial value of a sampling frequency i as 1;
acquiring adjacent nodes of a current node;
when i is smaller than the sampling step length, calculating the bias migration probability of the current node and the adjacent node according to a preset bias migration probability algorithm;
sorting the calculated offset walk probability from big to small, selecting M adjacent nodes arranged at the top, determining the selected adjacent nodes as current nodes, updating i to i +1, and returning to execute the operation of obtaining the adjacent nodes of the current nodes;
and when the i is larger than or equal to the sampling step length, inputting at least two groups of Node sequences obtained by sampling into a preset Node2Vec model for training to obtain the trained Node2Vec model.
Specifically, a slice example to be trained is obtained, where the number of the slice examples to be trained may be one or more, and this is not particularly limited in this embodiment of the present invention.
In order to obtain sample data of model training, sampling is carried out in the following way: for the example of the slice to be trained, an initial node is randomly determined, then a next hop node of the initial node is randomly determined from adjacent nodes of the initial node, the next hop node is determined as a current node, and since the sampling step length is preset, the initial value of the sampling frequency i is set to be 1 in order to perform sampling according to the sampling step length.
And then acquiring adjacent nodes of the current node, judging the relation between the i and the sampling step length, and calculating the bias migration probability of the current node and the adjacent nodes according to a preset bias migration probability algorithm when the i is smaller than the sampling step length.
Fig. 2 is a network topology diagram of a slice example labeled with bias weight in the related art, as shown in fig. 2, t represents a start Node, v represents a next hop Node of the start Node, and t, x1, x2 and x3 represent neighboring nodes of the v Node, in the related art, a walk mode in a Node2Vec model adopts the following formula to calculate a bias walk probability α pq (t,x)：
Wherein, in connection with fig. 2, the parameters p and q are used for controlling the sampling, respectivelyThe process has tendencies of Breadth First Search (BFS) and Depth First Search (DFS), wherein p is a return parameter, the larger p is, the smaller the probability of obtaining the same node in the sampling process is, q is an in-out parameter, and if q is the in-out parameter>1, the sampling process will favor BFS if q is<1, the sampling process is biased towards DFS, d tx Representing the shortest hop number required for node t to hop to node v.
In the embodiment of the present invention, in order to implement triggering of slice instance reconfiguration, a preset offset walk probability is provided, where the preset offset walk probability is positively correlated with a network performance parameter of a current service when a slice instance is adopted at a current time, that is, a value of the offset walk probability increases with an increase in the network performance parameter of the current service of the slice instance and decreases with a decrease in the network performance parameter of the current service of the slice instance, so that the offset walk probability is correlated with the network performance parameter of the current service.
Sorting the calculated bias walk probability from big to small, and selecting M adjacent nodes arranged at the top as current nodes, wherein M is a positive integer smaller than the number of the adjacent nodes; therefore, the training sample obtained by sampling is the Node sequence with better network performance parameters of the current service, the quality of the training sample is improved, and the Node2Vec model which meets the requirements better can be trained.
And after the current node is updated by selecting the adjacent node, updating i to i +1, sampling for the next time, and returning to execute the acquisition of the adjacent node of the current node.
In addition, when i is larger than or equal to the sampling step length, the sampling is completed, then at least two groups of Node sequences jumped in the sampling are used as training samples and input into a preset Node2Vec model for training, and the trained Node2Vec model is obtained.
In the embodiment of the invention, the bias parameter w is introduced in the sampling stage of the trained Node2Vec model vx So that the sampled node sequence sample can train a mapping model fusing a slice structure and a load state to finish slicing knotsAnd fusion mapping of the configuration and load information can realize multi-factor oriented reconfiguration decision.
In order to determine whether to trigger the reconfiguration of the slice instance more accurately, in an optional embodiment, the network performance parameter of the current service includes one or more of the following:
the average occupied bandwidth of the current service, the average message delay of the current service and the average packet loss rate of the current service.
Specifically, the network performance parameter of the current service may include one or more of an average occupied bandwidth of the current service, an average message delay of the current service, and an average packet loss rate of the current service, where this is not specifically limited in this embodiment of the present invention.
Further, in order to determine whether to trigger the reconfiguration of the slice instance more accurately, the network performance parameters of the current service may include an average occupied bandwidth of the current service, an average message delay of the current service, and an average packet loss rate of the current service, and in an optional embodiment, the offset wandering probability pi from the current node to the adjacent node is calculated by using the following formula vx ：
π vx ＝α pq (t,x)·w vx (2)
Where v denotes the current node, t denotes the neighbor node of the current node, W 0 Represents the average occupied bandwidth, T, of the current service at the current time 0 Average message delay representing the current service at the current time, E 0 Represents the average packet loss rate, k, of the current service at the current time W Is W 0 Estimated level of (k) T Is T 0 Is estimated by the level k E Is E 0 The estimated level of (c).
Wherein, in practical application, the k is W ，k T And k E The sum is 1, and the occupied bandwidth and the message time can be adjusted by adjusting the three estimation grade parametersThe importance of delay and packet loss rate in the sampling decision.
Here, formula (3) is introduced on the basis of the above formula (1) in the original bias walk probability, that is, the original bias walk probability is multiplied by w vx So that pi is obtained by adopting a preset bias walk probability algorithm vx The structure of the Node and the load condition of the Node are integrated, so that the network performance of the current service is considered in the obtained bias wandering probability, the selected adjacent Node is the network Node with better network performance, the quality of a training sample is favorably improved, and a more accurate Node2Vec model is favorably trained.
S103: and determining whether to trigger the reconstruction of the slice example according to the vector of the node.
After determining the node vector of the slice instance, in order to determine whether to trigger the reconstruction of the slice instance, whether to trigger the reconstruction of the slice instance may be determined according to the node vector, for example, one component in the node vector may be used to determine whether to trigger the reconstruction of the slice instance, another component in the node vector may be used to determine whether to trigger the reconstruction of the slice instance, or two components of the node vector may be used to determine whether to trigger the reconstruction of the slice instance, which is not specifically limited herein.
In order to determine whether to trigger the reconfiguration of the slice instance, in an alternative embodiment, S102 may include:
calculating the Euclidean distance between any two nodes in the slice example according to the vectors of the nodes;
and when the Euclidean distance between any two nodes meets a preset condition, triggering the reconstruction of the slice example.
Specifically, according to the vectors of the nodes, the Euclidean distance between any two nodes in the slicing example is calculated by using a distance formula between the two nodes, whether the Euclidean distance between any two nodes meets a preset condition or not is judged, the reconstruction of the slicing example is triggered only when the Euclidean distance meets the preset condition, and the reconstruction of the slicing example is forbidden when the Euclidean distance does not meet the preset condition.
Further, in order to trigger the reconstruction of the slice instance, in an optional embodiment, when the euclidean distance between any two nodes satisfies a preset condition, the triggering of the reconstruction of the slice instance includes:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is smaller than the preset minimum Euclidean distance between the nodes.
Specifically, a preset minimum euclidean distance between nodes is preset in the triggering method, the minimum euclidean distance represents an acceptable minimum vector distance in a vector space, and the practical meaning is that a load upper limit which can be borne between two corresponding nodes in the current slice example is present.
Further, in order to trigger the reconstruction of the slice instance, in an optional embodiment, when the euclidean distance between any two nodes satisfies a preset condition, the triggering of the reconstruction of the slice instance includes:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is larger than the preset maximum Euclidean distance between the nodes.
Specifically, the preset maximum euclidean distance between nodes is preset in the triggering method, the maximum euclidean distance represents the acceptable maximum vector distance in the vector space, and the practical meaning is that the lower limit of the acceptable resource occupancy rate between two corresponding nodes in the current slicing example, so that when the distance greater than the preset maximum euclidean distance between the nodes exists in the euclidean distance between any two nodes, it indicates that the resource occupancy rate between two nodes in the slicing example is greater than the lower limit of the acceptable resource occupancy rate between two corresponding nodes in the current slicing example, and it is seen that the current slicing example cannot guarantee the stable operation of the current service, and therefore, the reconfiguration of the slicing example is triggered.
In order to achieve accurate triggering of the reconfiguration of the slice instance, in an optional embodiment, the method further includes:
and when the Euclidean distance between any two nodes does not have a distance smaller than the preset minimum Euclidean distance between the nodes, and the Euclidean distance between any two nodes does not have a distance larger than the preset maximum Euclidean distance between the nodes, the reconstruction of the slicing example is forbidden to be triggered.
When the Euclidean distance between any two nodes does not have a distance smaller than the preset minimum Euclidean distance between the nodes, the fact that the load between the two nodes in the slicing example is lower than the upper limit of the load which can be borne by the two corresponding nodes in the current slicing example is indicated, and when the Euclidean distance between any two nodes does not have a distance larger than the preset maximum Euclidean distance between the nodes, the fact that the resource occupancy rate between the two nodes in the slicing example is larger than the lower limit of the acceptable resource occupancy rate between the two corresponding nodes in the current slicing example is indicated.
The triggering method described in one or more of the above embodiments is described below by way of example.
Fig. 3 is a schematic flowchart of an example of an optional triggering method provided in an embodiment of the present invention, and as shown in fig. 3, the slice reconfiguration triggering method of this example is deployed after an online operation of a slice example and before an offline of the slice example, and realizes an evolution from a static slice to an elastic slice in a standard slice life cycle; the triggering method can comprise the following steps:
s301: after the slicing service is started, firstly, selecting a slicing template from a slicing template library; wherein the selected slice template is matched with the current business requirement;
s302: creating a slice instance according to the selected slice template;
s303: and the slice example is operated on line to bear the service requirement.
Specifically, after the slice instance is online, the slice reconstruction triggering strategy provided by the invention is started.
S304: judging whether the slicing service is finished or not; if yes, executing S305, if no, executing S306;
s305: off-line slicing examples; executing S311;
s306: sampling by adopting an S strategy, namely adopting a section example to be trained to obtain a training sample;
s307: training a Node2Vec model by using a training sample to obtain a trained Node2Vec model;
s308: inputting the slice example into a trained Node2Vec model, and mapping the slice example to a vector space to obtain a vector of a Node in the slice example;
specifically, the slice reconfiguration triggering strategy is to sample the current slice based on an S sampling strategy (which is equivalent to the sampling manner described in one or more embodiments above), train the Node2Vec model using the collected samples, and then input the slice into the trained model to obtain the mapping result of the slice structure and the load state in the low-dimensional vector space.
It should be noted that the Node2Vec model is trained by using the training data obtained by the S strategy, the model can map the slice and the slice load from the high-dimensional information space to the two-dimensional vector space, and each virtual Node obtains its corresponding two-dimensional vector representation. Because the information of the load state of the virtual link is fused in the offset walk strategy of the Node2Vec sampling, the Euclidean distance between the vectors can represent not only the connection similarity of the topological structure, but also the load state between two virtual nodes.
S309: calculating Euclidean distances between nodes, and judging whether a distance exceeding a threshold exists in the Euclidean distances; if yes, executing S310, and if not, executing S304;
s310: reconstructing slices, namely triggering the reconstruction of slice instances;
s311: and recovering the resources and ending the service.
Specifically, after the mapping is completed, a threshold check is started to see if there are vector pairs that are outside the threshold range. If the current slice exists, reconstructing the current slice, replacing the current running slice with the reconstructed slice, and continuously executing the monitoring of the reconstruction trigger strategy on the reconstructed slice; and if not, continuing to monitor the current slice by the reconstruction triggering strategy. And if the slicing service is finished, taking the slicing example off the line, recycling all resources distributed to the slicing and finishing the service.
For the threshold check, specifically, if a virtual link connection exists between two nodes or the traffic load between them is heavy, it may cause the euclidean distance between their corresponding vector pairs in the vector space to become close, based on the variation relationship, the slice state monitoring algorithm of this example will set up two thresholds, threshold T min Representing the acceptable minimum vector distance in the vector space, wherein the practical meaning is the upper limit of the load which can be borne between two corresponding nodes in the current slicing example, and when the situation that the vector space is smaller than T appears is monitored min At euclidean distance of (e), slice reconstruction will be triggered. Threshold value T max The maximum vector distance acceptable in the vector space is represented, the practical meaning is that the acceptable resource occupancy rate lower limit between two corresponding nodes in the slicing example is represented, and when the situation that more than T appears in the vector space is monitored max At euclidean distance of (c), slice reconstruction will be triggered. The difference in triggering slice reconstruction in both cases is that the former allocates more bearer resources for the new slice instance than the existing slice instance at the time of reconstruction, while the latter allocates less bearer resources than the existing slice instance.
That is to say, this example provides a Node2 Vec-based slice state monitoring method, which maps a slice structure and a slice load from a high-dimensional information space to a two-dimensional vector space, thereby reducing the spatial complexity required for storing the slice structure and the slice state, reducing the spatial cost for monitoring the slice structure and the slice state, and improving the efficiency for managing the slice structure and the slice state, and adopts a two-dimensional vector space-based slice reconstruction trigger decision, so that in a two-dimensional vector diagram output by a Node2Vec algorithm, vectors and nodes in a slice topology have a one-to-one correspondence relationship, and the vector relationship reflects the Node2Vec algorithmThe method has the advantages that the topological structure and the load state between nodes in the slicing Chilean are realized, the multi-dimensional information fusion decision of the slicing structure and the load is realized by taking the relation operation result between vectors as the judgment standard triggered by reconstruction, the decision accuracy is improved, the sampling strategy S is adopted, and the offset parameter w is introduced in the Node2Vec model sampling stage vx The sampled Node sequence sample can train a mapping model fusing a slice structure and a load state, and an implementation scheme is provided for the application of the Node2Vec model in the field of slice representation and monitoring.
The invention provides a triggering method, which comprises the following steps: obtaining a slice example of a network slice, inputting the slice example into a trained Node2Vec model, and outputting to obtain a vector of a Node in the slice example, wherein vector elements of the Node comprise: determining a quantity for representing a node structure and a quantity for representing a node load, and triggering reconfiguration of a slice instance according to a vector of a node; that is to say, in the present invention, after a slice instance of a network slice is obtained, the slice instance is input into a trained Node2Vec model, and a vector of a Node of the slice instance can be obtained, because the obtained vector includes a quantity for representing a Node structure and a Node load, and the structure of the slice and the load condition of the slice can be reflected based on the vector of the Node, then, in determining whether to trigger reconstruction of the slice instance according to the vector of the Node, because the vector of the Node has two dimensions, the complexity of determining whether to trigger reconstruction of the slice instance is reduced by using the vector of the two dimensions, thereby improving the decision efficiency of whether the network slice needs to be reconstructed, and further improving the reconstruction efficiency of the network slice for reconstruction.
Example two
Based on the same inventive concept, an embodiment of the present invention provides a triggering device, and fig. 4 is a schematic structural diagram of an optional triggering device in the embodiment of the present invention, as shown in fig. 4, the triggering device includes: an acquisition module 41, a processing module 42 and a triggering module 43;
the acquiring module 41 is configured to acquire a slice example of a network slice;
the processing module 42 is used for inputting the slice example into the trained Node2Vec model and outputting the vector of the Node in the slice example; wherein the vector elements of the nodes include: quantities representing node structure and quantities representing node load;
a triggering module 43, configured to determine whether to trigger a reconfiguration of the slice instance according to the vector of the node.
In an alternative embodiment, the apparatus is further configured to:
the trained Node2Vec model is obtained by adopting the following method:
acquiring an initial node and a next hop node of the initial node from a slice example to be trained, determining the next hop node as a current node, and setting an initial value of sampling times i as 1;
acquiring adjacent nodes of a current node;
when i is smaller than the sampling step length, calculating the bias walk probability of the current node and the adjacent node according to a preset bias walk probability algorithm; the preset bias migration probability is positively correlated with the network performance parameter of the current service when the slice example is adopted at the current time;
sorting the calculated offset walk probability from big to small, selecting M adjacent nodes arranged at the top, determining the selected adjacent nodes as current nodes, updating i to i +1, and returning to execute the operation of obtaining the adjacent nodes of the current nodes; wherein M is a positive integer less than the number of adjacent nodes;
and when the i is larger than or equal to the sampling step length, inputting at least two groups of Node sequences obtained by sampling into a preset Node2Vec model for training to obtain the trained Node2Vec model.
In an alternative embodiment, the network performance parameters of the current service include one or more of:
the average occupied bandwidth of the current service, the average message delay of the current service and the average packet loss rate of the current service.
In an alternative embodiment, the apparatus is further configured to:
calculated by the above formula (1) to formula (3)Bias walk probability pi from current node to adjacent node vx 。
In an alternative embodiment, the triggering module 43 is specifically configured to:
calculating the Euclidean distance between any two nodes in the slice example according to the vectors of the nodes;
and when the Euclidean distance between any two nodes meets a preset condition, triggering the reconstruction of the slice example.
In an optional embodiment, when the euclidean distance between any two nodes satisfies the preset condition, the triggering module 43 triggers the reconfiguration of the slice instance, which includes:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is smaller than the preset minimum Euclidean distance between the nodes.
In an optional embodiment, when the euclidean distance between any two nodes satisfies a preset condition, the triggering module 43 triggers the reconfiguration of the slice instance, including:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is larger than the preset maximum Euclidean distance between the nodes.
In an alternative embodiment, the apparatus is further configured to:
and when the Euclidean distance between any two nodes does not have a distance smaller than the preset minimum Euclidean distance between the nodes, and the Euclidean distance between any two nodes does not have a distance larger than the preset maximum Euclidean distance between the nodes, the reconstruction of the slicing example is forbidden to be triggered.
In practical applications, the obtaining module 41, the Processing module 42 and the triggering module 43 may be implemented by a processor located on a device, specifically, implemented by a Central Processing Unit (CPU), a Microprocessor Unit (MPU), a Digital Signal Processor (DSP), a Field Programmable Gate Array (FPGA), or the like.
Fig. 5 is a schematic structural diagram of an alternative apparatus provided in an embodiment of the present invention, and as shown in fig. 5, an embodiment of the present invention provides an apparatus 500, including:
a processor 51 and a storage medium 52 storing instructions executable by the processor 51, wherein the storage medium 52 depends on the processor 51 to perform operations through a communication bus 53, and when the instructions are executed by the processor 51, the triggering method of the first embodiment is performed.
It should be noted that, in practical applications, the various components in the terminal are coupled together by a communication bus 53. It will be appreciated that the communication bus 53 is used to enable communications among the components of the connection. The communication bus 53 includes a power bus, a control bus, and a status signal bus in addition to a data bus. But for clarity of illustration the various buses are labeled in figure 5 as communication bus 53.
Embodiments of the present invention provide a computer storage medium storing executable instructions, and when the executable instructions are executed by one or more processors, the processors execute the triggering method according to one embodiment.
The computer-readable storage medium may be a magnetic random access Memory (FRAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), a Flash Memory (Flash Memory), a magnetic surface Memory, an optical Disc, or a Compact Disc Read-Only Memory (CD-ROM).
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. 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 a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention.
Claims (11)
1. A method of triggering, comprising:
acquiring a slice example of the network slice;
inputting the slice example into a trained Node2Vec model, and outputting to obtain a vector of a Node in the slice example; wherein the vector elements of the node include: quantities representing node structure and quantities representing node load;
determining whether to trigger reconstruction of the slice instance according to the vector of the node.
2. The method of claim 1 wherein the trained Node2Vec model is obtained as follows:
acquiring an initial node and a next hop node of the initial node from a slice example to be trained, determining the next hop node as a current node, and setting an initial value of a sampling frequency i as 1;
acquiring adjacent nodes of a current node;
when i is smaller than the sampling step length, calculating the bias walk probability of the current node and the adjacent node according to a preset bias walk probability algorithm; the preset bias walk probability is positively correlated with the network performance parameter of the current service when the slice example is adopted at the current moment;
sorting the calculated bias walk probability from big to small, selecting M adjacent nodes arranged at the top, determining the selected adjacent nodes as current nodes, updating i to i +1, and returning to execute the adjacent nodes for acquiring the current nodes; wherein M is a positive integer less than the number of adjacent nodes;
and when i is larger than or equal to the sampling step length, inputting at least two groups of Node sequences obtained by sampling into a preset Node2Vec model for training to obtain the trained Node2Vec model.
3. The method of claim 2, wherein the network performance parameters of the current traffic comprise one or more of:
the average occupied bandwidth of the current service, the average message delay of the current service and the average packet loss rate of the current service.
4. A method according to claim 2 or 3, characterized in thatThen, the bias migration probability pi from the current node to the adjacent node is calculated by adopting the following formula vx ：
π vx ＝α pq (t,x)·w vx
Where v denotes the current node, t denotes the neighbor node of the current node, W 0 Represents the average occupied bandwidth, T, of the current service at the current time 0 Average message delay representing the current service at the current time, E 0 Represents the average packet loss rate, k, of the current service at the current time W Is W 0 Is estimated by the level k T Is T 0 Is estimated by the level k E Is E 0 Estimated level of d tx Representing the shortest hop number required for node t to hop to node x.
5. The method of claim 1, wherein the determining whether to trigger reconstruction of the slice instance according to the vector of nodes comprises:
calculating the Euclidean distance between any two nodes in the slice example according to the vectors of the nodes;
and when the Euclidean distance between any two nodes meets a preset condition, triggering the reconstruction of the slice example.
6. The method of claim 5, wherein triggering the reconstruction of the slice instance when the Euclidean distance between any two nodes satisfies a preset condition comprises:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is smaller than the preset minimum Euclidean distance between the nodes.
7. The method of claim 5, wherein triggering the reconstruction of the slice instance when the Euclidean distance between any two nodes satisfies a preset condition comprises:
and triggering the reconstruction of the slice example when the Euclidean distance between any two nodes is larger than the preset maximum Euclidean distance between the nodes.
8. The method of claim 5, further comprising:
and when the Euclidean distance between any two nodes does not have a distance smaller than a preset minimum Euclidean distance between the nodes, and the Euclidean distance between any two nodes does not have a distance larger than a preset maximum Euclidean distance between the nodes, forbidding to trigger the reconstruction of the slice example.
9. A trigger device, the device comprising:
the acquisition module is used for acquiring a slice example of the network slice;
the processing module is used for inputting the slice example into the trained Node2Vec model and outputting the vector of the Node in the slice example; wherein the vector elements of the node include: a quantity representing a node structure and a quantity representing a node load;
and the triggering module is used for determining whether to trigger the reconstruction of the slice example according to the vector of the node.
10. An apparatus, characterized in that the apparatus comprises:
a processor and a storage medium storing instructions executable by the processor to perform operations dependent on the processor via a communication bus, the instructions, when executed by the processor, performing the triggering method of any of the preceding claims 1 to 8.
11. A computer storage medium having stored thereon executable instructions that, when executed by one or more processors, perform the triggering method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110397103.6A CN115209431B (en) | 2021-04-13 | 2021-04-13 | Triggering method, device, equipment and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110397103.6A CN115209431B (en) | 2021-04-13 | 2021-04-13 | Triggering method, device, equipment and computer storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115209431A true CN115209431A (en) | 2022-10-18 |
| CN115209431B CN115209431B (en) | 2023-10-27 |
Family
ID=83570403
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110397103.6A Active CN115209431B (en) | 2021-04-13 | 2021-04-13 | Triggering method, device, equipment and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115209431B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018184666A1 (en) * | 2017-04-04 | 2018-10-11 | Telefonaktiebolaget Lm Ericsson (Publ) | Training a software agent to control a communication network |
| CN108965024A (en) * | 2018-08-01 | 2018-12-07 | 重庆邮电大学 | A kind of virtual network function dispatching method of the 5G network slice based on prediction |
| CN110213780A (en) * | 2018-02-28 | 2019-09-06 | 中兴通讯股份有限公司 | Management method, management and the layout entity and storage medium of network slice |
| CN110881207A (en) * | 2019-10-31 | 2020-03-13 | 华为技术有限公司 | Network slice selection method and related product |
| WO2020057261A1 (en) * | 2018-09-17 | 2020-03-26 | 华为技术有限公司 | Communication method and apparatus |
| CN110958666A (en) * | 2019-11-20 | 2020-04-03 | 无锡北邮感知技术产业研究院有限公司 | Network slice resource mapping method based on reinforcement learning |
| CN111083744A (en) * | 2019-12-31 | 2020-04-28 | 北京思特奇信息技术股份有限公司 | Network slicing method, device, storage medium and equipment |
| CN111866953A (en) * | 2019-04-26 | 2020-10-30 | 中国移动通信有限公司研究院 | Network resource allocation method, device and storage medium |
| WO2020258920A1 (en) * | 2019-06-26 | 2020-12-30 | 华为技术有限公司 | Network slice resource management method and apparatus |
| CN112613230A (en) * | 2020-12-15 | 2021-04-06 | 云南电网有限责任公司 | Network slice resource dynamic partitioning method and device based on neural network |
-
2021
- 2021-04-13 CN CN202110397103.6A patent/CN115209431B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018184666A1 (en) * | 2017-04-04 | 2018-10-11 | Telefonaktiebolaget Lm Ericsson (Publ) | Training a software agent to control a communication network |
| CN110213780A (en) * | 2018-02-28 | 2019-09-06 | 中兴通讯股份有限公司 | Management method, management and the layout entity and storage medium of network slice |
| CN108965024A (en) * | 2018-08-01 | 2018-12-07 | 重庆邮电大学 | A kind of virtual network function dispatching method of the 5G network slice based on prediction |
| WO2020057261A1 (en) * | 2018-09-17 | 2020-03-26 | 华为技术有限公司 | Communication method and apparatus |
| CN111866953A (en) * | 2019-04-26 | 2020-10-30 | 中国移动通信有限公司研究院 | Network resource allocation method, device and storage medium |
| WO2020258920A1 (en) * | 2019-06-26 | 2020-12-30 | 华为技术有限公司 | Network slice resource management method and apparatus |
| CN110881207A (en) * | 2019-10-31 | 2020-03-13 | 华为技术有限公司 | Network slice selection method and related product |
| CN110958666A (en) * | 2019-11-20 | 2020-04-03 | 无锡北邮感知技术产业研究院有限公司 | Network slice resource mapping method based on reinforcement learning |
| CN111083744A (en) * | 2019-12-31 | 2020-04-28 | 北京思特奇信息技术股份有限公司 | Network slicing method, device, storage medium and equipment |
| CN112613230A (en) * | 2020-12-15 | 2021-04-06 | 云南电网有限责任公司 | Network slice resource dynamic partitioning method and device based on neural network |
Non-Patent Citations (3)
| Title |
|---|
| 唐伦;周钰;谭颀;魏延南;陈前斌;: "基于强化学习的5G网络切片虚拟网络功能迁移算法", 电子与信息学报, no. 03 * |
| 徐丹;王海宁;袁祥枫;朱雪田;: "基于AI的5G网络切片管理技术研究", 电子技术应用, no. 01 * |
| 秦爽;赵冠群;冯钢;: "基于多智体强化学习的接入网络切片动态切换", 电子科技大学学报, no. 02 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115209431B (en) | 2023-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108632365B (en) | Service resource adjusting method, related device and equipment | |
| CN111371603B (en) | Service instance deployment method and device applied to edge computing | |
| CN105471759A (en) | Network traffic scheduling method and apparatus for data centers | |
| CN111082960B9 (en) | Data processing method and device | |
| CN111104291B (en) | Environment monitoring method, device and system based on Internet of things and monitoring server | |
| CN112560204B (en) | Optical network route optimization method based on LSTM deep learning and related device thereof | |
| CN112764920A (en) | Edge application deployment method, device, equipment and storage medium | |
| CN110062375A (en) | Information processing method and device, computer storage medium | |
| CN113869521A (en) | Method, device, computing equipment and storage medium for constructing prediction model | |
| CN116450312A (en) | Scheduling strategy determination method and system for pipeline parallel training | |
| Esmat et al. | Deep reinforcement learning based dynamic edge/fog network slicing | |
| Nguyen et al. | Efficient virtual network embedding with node ranking and intelligent link mapping | |
| CN112925634A (en) | Heterogeneous resource scheduling method and system | |
| CN112130927B (en) | Reliability-enhanced mobile edge computing task unloading method | |
| CN111090401B (en) | Storage device performance prediction method and device | |
| CN115209431B (en) | Triggering method, device, equipment and computer storage medium | |
| CN106406082B (en) | System control method, device, controller and control system | |
| CN113050955A (en) | Self-adaptive AI model deployment method | |
| CN116149855A (en) | Method and system for optimizing performance resource cost under micro-service architecture | |
| CN115629883A (en) | Resource prediction method, resource prediction device, computer equipment and storage medium | |
| CN111639741B (en) | Automatic service combination agent system for multi-objective QoS optimization | |
| Bensalem et al. | Scaling Serverless Functions in Edge Networks: A Reinforcement Learning Approach | |
| JP7160205B2 (en) | Extraction device, extraction method and extraction program | |
| JP6732927B2 (en) | Housing form searching apparatus, housing form searching method and program | |
| CN112953844B (en) | Network traffic optimization method and device |
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 |