A hybrid service function chain pre-deployment method and system based on mobile perception

Abstract

This invention discloses a mobile-aware hybrid service function chain pre-deployment method and system. Based on mobile user request instructions and resource constraints of edge cloud nodes, this system utilizes node objectification and a dynamic auxiliary graph to implement H-SFC sequence constraints. The dynamic auxiliary graph reduces problem complexity and simplifies path finding. By considering that VNFs of the same type can share basic resources, resource sharing within and between H-SFC links is achieved. Furthermore, the system uses the fewest H-SFCs to cover more configuration file locations, reducing resource consumption and improving physical network throughput. This system ensures that when a user moves to the next location, the H-SFC at the current location can immediately provide service to the user, ensuring seamless migration of mobile user services while minimizing resource consumption.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a hybrid service function chain pre-deployment method and system based on mobile perception. Background Art

[0002] With the widespread adoption of mobile smart devices such as smartphones, an increasing number of computationally intensive mobile applications, such as interactive online games and self-driving cars, require timely data processing. Due to the limited computing, storage, and energy resources on mobile devices, offloading computationally intensive application data to a central cloud with abundant computing and storage resources can significantly reduce the computational load on mobile devices. However, since the central cloud is typically located far from end users, this can lead to unavoidable long response delays, degrading the user's quality of service experience. Therefore, for services with strict latency and high security requirements, MEC technology is generally adopted to offload application data to edge networks closer to mobile users for processing, thereby reducing processing latency. Typically, to meet business needs, in networks supporting Network Function Virtualization (NFV), flow requests need to be forwarded through a service function chain (SFC), consisting of a specific sequence of middleware implemented by virtual network functions (VNFs). In some mobile device applications, forward traffic originates from the mobile device's end node and is processed by a sequence of VNFs before being offloaded to the edge cloud for analysis. Backward traffic, on the other hand, is the data flow that returns the analysis results to the terminal after processing by the VNFs. Because forward and backward traffic may require different processing functions, the request flows require different forward and backward SFCs, namely, hybrid service function chains (H-SFCs). Since multiple VNF instances of the same type exist in the network and within the same H-SFC, resource sharing among VNFs of the same type within the network, or even within the same H-SFC, not only reduces computational and storage resource costs but also reduces link resource usage, improving resource utilization and reducing service latency. Furthermore, due to the high mobility of users, latency constraints may be violated when the served user exceeds the H-SFC coverage provided by the current edge node. Therefore, implementing resource-sharing-based hybrid service function chains for mobile users in MEC is of great significance. Most existing work migrates VNFs to new edge servers and redeploys them to meet the H-SFC latency requirements of the request flow. This results in excessive VNF deployment time and service interruptions during the migration process. Some work also reduces handover time by pre-deploying H-SFCs at each mobile user's location, but this increases resource costs. Summary of the Invention

[0003] The technical problem to be solved by the present invention is that the VNFs migration in the existing technology has the problems of long VNF deployment time during the migration process, easy service interruption, and high resource cost consumption. The purpose is to provide a hybrid service function chain pre-deployment method and system based on mobile perception. According to the request instructions of mobile users and the resource constraints of edge cloud nodes, node objectification and dynamic auxiliary graphs are used to implement the sequence constraints of H-SFC, and the dynamic auxiliary graph is used to reduce the complexity of the problem and simplify the path search, ensuring that when the user moves to the next location, the H-SFC at the current location can immediately provide services to the user, ensuring seamless migration of mobile user services while minimizing resource consumption.

[0004] The present invention is achieved through the following technical solutions:

[0005] A first aspect of the present invention provides a hybrid service function chain pre-deployment method based on mobility perception, comprising the following specific steps:

[0006] S1. Obtain the request instruction of the mobile user and determine the adapted link and node information based on the resource constraints of the edge cloud node;

[0007] S2. Based on the node information, objectify the nodes and construct an edge weighted graph;

[0008] S3. Based on the edge-weighted graph, adjust the link and node objects and assign weighted edges to construct a dynamic auxiliary weighted graph that satisfies the order constraint;

[0009] S4. Based on the mobile user's request instruction, obtain the mobile user's profile, determine the highest probability location where the mobile user may stay in the profile, and combine it with the dynamic auxiliary weighted graph to determine a set of all paths that meet the H-SFC delay requirement required by the mobile user;

[0010] S5. Determine the primary link based on the path set, update the edge weights of the dynamic auxiliary weighted graph and the location information in the configuration file, and jump to step S4 until all access locations in the configuration file are deployed. Use the updated deployment path as the secondary path for this mobile user.

[0011] The present invention combines the mobile user's request instructions with the resource constraints of the edge cloud node to determine the adapted link and node information, construct an edge weighted graph, and construct a dynamic auxiliary graph that meets the sequence constraints based on the sequence requirements of the H-SFC requested by the user. This method determines the highest-probability location where the mobile user may stay in the configuration file, determines the set of all paths that meet the H-SFC delay requirements required by the mobile user, and determines the main link and secondary path. This ensures that when the user moves to the next location, the H-SFC at the current location can immediately provide services to the user, ensuring seamless migration of mobile user services while minimizing resource consumption.

[0012] Furthermore, the request instruction of the mobile user includes the sequence requirement, delay constraint, bandwidth requirement and mobile profile of the H-SFC.

[0013] Furthermore, objectifying each functional node specifically includes:

[0014] Determine the VNF resources that can be shared in the node and the remaining resources of the node;

[0015] The VNF resources include: VNF deployment resources and VNF processing resources;

[0016] The sharing of the VNF resources occurs within the H-SFC and between multiple H-SFCs.

[0017] Furthermore, the step of constructing an edge-weighted graph specifically includes:

[0018] Perform weighted processing on each node based on all VNFs required by H-SFC in the request instruction;

[0019] If a node contains a VNF required by the H-SFC and the node's remaining resources are greater than the processing resource requirements of the corresponding VNF in the H-SFC, the VNF is split into a virtual link and a relative cost weight and delay weight are assigned to the virtual link.

[0020] If the node does not have the VNF required by H-SFC, but the node resources meet the deployment resource and processing resource requirements of the required VNF, the required VNF is first deployed on the functional node, and then the VNF is split into a virtual link, and the relative cost weight and delay weight are assigned to the virtual link respectively.

[0021] Furthermore, the specific steps of constructing a dynamic auxiliary weighted graph that satisfies the order constraint include:

[0022] S31. Based on the sequence requirement of the H-SFC requested by the user, all nodes that can provide the VNFs required by the H-SFC are selected from the weighted graph, and the selected nodes are arranged into a sequence graph consisting of multiple columns according to the VNF type requirements and sequence requirements in the H-SFC.

[0023] S32. In the sequence diagram, each node in the previous column of nodes is connected to all nodes in the next column to form a link;

[0024] S33, using the Dijkstra algorithm based on the original physical network to calculate the relative cost weight and delay weight between two nodes in the sequence graph;

[0025] S34. If the delay weight exceeds the requested delay tolerance, the node is removed, and otherwise each value is assigned to the corresponding link.

[0026] Furthermore, the nodes in each column are composed of functional nodes that can provide the same type of VNF.

[0027] Furthermore, determining a set of all paths that meet the H-SFC delay requirement of the mobile user includes:

[0028] Use the depth-first search algorithm to calculate the set of all paths that meet the required H-SFC delay requirements;

[0029] Depth-first search is used to calculate the delay and relative cost consumption at the same time, and the node resource status is updated in real time during the search.

[0030] Furthermore, the profile of the mobile user is the probability that the mobile user stays at different edge cloud nodes in time segment t.

[0031] Furthermore, determining the primary link based on the path set specifically includes:

[0032] According to the location information in the updated configuration file, the remaining locations in the configuration file are linked to each path in the link set in sequence;

[0033] If the path meets the H-SFC delay requirement, this position is covered by the link; otherwise, the link set is traversed continuously;

[0034] Sum the user stay probabilities corresponding to each location in the location set covered by each path, and select the path corresponding to the maximum sum of probabilities as the main link;

[0035] If other locations in the configuration file are not overwritten, the link with the smallest relative cost in the path set is selected as the main link.

[0036] A second aspect of the present invention provides a hybrid service function chain pre-deployment system based on mobility perception, comprising:

[0037] The parameter acquisition module is used to obtain the request instructions of the mobile user and determine the adapted link and node information based on the resource constraints of the edge cloud node;

[0038] Node objectification and assignment module, used to objectify nodes based on node information and construct edge weighted graphs;

[0039] A dynamic auxiliary weighted graph module is constructed to adjust the link and node objects and assign weighted edges based on the edge weighted graph, so as to construct a dynamic auxiliary weighted graph that satisfies the order constraint;

[0040] The mobile user preselected path set acquisition module is used to obtain the mobile user's profile based on the mobile user's request instruction, determine the highest probability location of the mobile user in the profile, and combine it with the dynamic auxiliary weighted graph to determine the set of all paths that meet the mobile user's H-SFC delay requirements;

[0041] The mobile user deployment path acquisition module is used to determine the main link based on the path set, update the edge weights of the dynamic auxiliary weighted graph and the location information in the configuration file, jump to step S4, and continue until all the access locations of the configuration file are deployed. The updated deployment path is used as the secondary path for this mobile user.

[0042] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0043] 1. Based on mobile user request instructions and edge cloud node resource constraints, we leverage node objectification and a dynamic auxiliary graph to implement H-SFC sequence constraints. This dynamic auxiliary graph reduces problem complexity and simplifies path finding. By considering that VNFs of the same type can share basic resources, we enable resource sharing within and between H-SFC links. Furthermore, by using the fewest H-SFCs to cover more configuration file locations, we reduce resource consumption and improve physical network throughput. This ensures that when a user moves to the next location, the H-SFC at the current location can immediately provide service to the user, ensuring seamless migration of mobile user services while minimizing resource consumption.

[0044] 2. By linking node objects, the situation where H-SFC may pass through the same node is resolved, and the path algorithm is used to calculate the weights of each edge.

[0045] 3. By constructing a sequential dynamic auxiliary graph of requests, not only does it resolve the order constraints of H-SFC, but it can also use the dynamic auxiliary graph multiple times to calculate routing paths based on the configuration file;

[0046] 4. By considering load balancing, we use the M / M / 1 queuing model to represent the relative cost of resource consumption. When the path set for a certain location in the configuration file does not cover other locations, the path with the lowest relative cost is selected as the routing path for this location, ensuring both low resource consumption and balanced node load.

[0047] 5. By considering the shared basic resources of VNFs within and between links, and multiple stay locations can share one H-SFC to provide services, resource consumption is reduced, the request throughput is improved, and the delay constraint is guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the examples. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be considered as limiting the scope. A person of ordinary skill in the art can also derive other relevant drawings based on these drawings without inventive effort. In the drawings:

[0049] Figure 1 This is a schematic diagram of the overall solution flow in an embodiment of the present invention;

[0050] Figure 2 This is an example diagram of mobile user migration in an MEC network in an embodiment of the present invention;

[0051] Figure 3 is an example diagram of H-SFC in an embodiment of the present invention;

[0052] Figure 4 is a node object link graph in an embodiment of the present invention;

[0053] Figure 5 This is a sequential dynamic auxiliary weighted graph in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with examples and drawings. The exemplary embodiments of the present invention and their descriptions are only used to explain the present invention and are not intended to limit the present invention.

[0055] Example 1

[0056] like Figure 1 As shown, this embodiment provides a hybrid service function chain pre-deployment method based on mobility perception, including the following specific steps:

[0057] S1. Obtain the request instruction of the mobile user and determine the adapted link and node information based on the resource constraints of the edge cloud node;

[0058] S2. Based on the node information, objectify the nodes and construct an edge weighted graph;

[0059] S3. Based on the edge-weighted graph, adjust the link and node objects and assign weighted edges to construct a dynamic auxiliary weighted graph that satisfies the order constraint;

[0060] S4. Based on the mobile user's request instruction, obtain the mobile user's profile, determine the highest probability location where the mobile user may stay in the profile, and combine it with the dynamic auxiliary weighted graph to determine a set of all paths that meet the H-SFC delay requirement required by the mobile user;

[0061] S5. Determine the primary link based on the path set, update the edge weights of the dynamic auxiliary weighted graph and the location information in the configuration file, and jump to step S4 until all access locations in the configuration file are deployed. Use the updated deployment path as the secondary path for this mobile user.

[0062] Combining the mobile user's request instructions and the resource constraints of the edge cloud node, the adapted link and node information is determined, an edge weighted graph is constructed, and a dynamic auxiliary graph that satisfies the sequence constraints is constructed based on the sequence requirements of the H-SFC requested by the user. This determines the highest-probability location where the mobile user may stay in the configuration file, the set of all paths that meet the H-SFC latency requirements required by the mobile user, and the main link and secondary path. This ensures that when the user moves to the next location, the H-SFC at the current location can immediately provide services to the user, ensuring seamless migration of mobile user services while minimizing resource consumption.

[0063] In some possible embodiments, the initial node network diagram is preliminarily screened to obtain link information and node information that meet the requirements. At the same time, since the nodes that meet the requirements in the obtained node information may not necessarily be able to form a connection relationship, the adapted links are obtained to improve the effectiveness of the acquired data and reduce the complexity of subsequent data processing.

[0064] In some possible embodiments, a mobile user's request instruction includes H-SFC sequence requirements, delay constraints, bandwidth requirements, and a mobility profile. A mobility profile is a collection of all possible location probabilities for a mobile user. When each user moves within the coverage area of a base station, this location is referred to as the user's location at that moment. Each user has a limited number of locations, and the probability of moving to each location varies. Therefore, timely H-SFC processing data must be provided at the mobile location to ensure seamless migration of the mobile user and avoid service interruptions.

[0065] like Figure 2 As shown, areas 1, 2, and 3 represent the locations where mobile users stay. These locations constitute the mobility profile. H-SFC instances provide latency-satisfactory services to mobile users at different locations. Due to the high mobility of mobile users, VNF instance deployment latency is very long. Temporarily deploying H-SFC instances can cause service interruptions. Therefore, H-SFC instances must be deployed in advance at the mobile locations.

[0066] In the network, an H-SFC instance is deployed in area 1. Area 2 can meet latency requirements by providing services through the H-SFC instance in area 1, so there is no need to deploy a separate H-SFC instance in area 2. Therefore, areas 1 and 2 share the same H-SFC instance, which can also be represented as covering both areas 1 and 2. The H-SFC latency for area 3 accessing area 1 exceeds the H-SFC latency, requiring the redeployment of a new H-SFC instance.

[0067] like Figure 3 As shown in Figure 1, the H-SFC consists of an ordered set of VNFs f, divided into forward SFC and backward SFC. When the forward SFC and backward SFC share the same VNF, computing resources can be conserved by leveraging the VNF's shared infrastructure. Each MEC contains deployed VNF instances and remaining computing resources. The remaining computing resources can be used to deploy new VNF instances.

[0068] like Figure 4 As shown in the figure, based on the requested H-SFC requirements and node resource availability, deployable VNFs and already deployed VNFs are split into directed virtual links. If the node's computing resources can serve as a destination, the destination is converted into a directed virtual link.

[0069] In some possible embodiments, objectifying each functional node specifically includes:

[0070] Determine the VNF resources that can be shared in the node and the remaining resources of the node;

[0071] VNF resources include: VNF deployment resources and VNF processing resources;

[0072] VNF resources are shared within an H-SFC and between multiple H-SFCs.

[0073] In some possible embodiments, constructing an edge-weighted graph specifically includes:

[0074] Perform weighted processing on each node based on all VNFs required by H-SFC in the request instruction;

[0075] If a node contains a VNF required by the H-SFC and the node's remaining resources are greater than the processing resource requirements of the corresponding VNF in the H-SFC, the VNF is split into a virtual link and a relative cost weight and delay weight are assigned to the virtual link.

[0076] If the node does not have the VNF required by H-SFC, but the node resources meet the deployment resource and processing resource requirements of the required VNF, the required VNF is first deployed on the functional node, and then the VNF is split into a virtual link, and the relative cost weight and delay weight are assigned to the virtual link respectively.

[0077] In some possible embodiments, resource requirements Relative cost weight Delay Weight The expressions are:

[0078]

[0079]

[0080]

[0081] and VNF f i The deployment resource consumption and processing resource consumption of e′ is a virtual link, E′ is the set of virtual links, <f i,n ,f′ i,n > represents VNF f on node n i Split into f i,n and f′ i,n The link formed by connecting two virtual points, Indicates VNF type i i Whether the instance has been deployed on server u, if VNF i If the instance already exists on server u, then Otherwise 0; cpu d is the resource consumption of the destination node, <d r,n ,d′ r,n > indicates a virtual link formed by two virtual points split from the destination node; b r is the bandwidth resource requirement of the request flow r, <b′ n-1 ,b n > represents two nodes b′ n-1 ,b n Virtual links between is the remaining rate of computing resources on node u, C u is the computing resource capacity of node u; B e is the bandwidth resource capacity between links, is the bandwidth surplus rate of link e; is the processing delay on server u, is the transmission delay of link e.

[0082] like Figure 5 As shown, S ris the set of locations where the configuration file stays, and the H-SFC order is {f1, f2, d, f1}. Arrange the probability of the configuration file's locations from high to low as the first column of the dynamic auxiliary graph. Then, based on the H-SFC order requirements, select nodes where f1 has been deployed or can be deployed as the second column. Deploy f2, d, and f1 following the same steps as above. The nodes between the two columns are connected by links, and the link weights are calculated using the shortest path algorithm. s1 is the location with the highest probability in the current configuration file. The two red links A and B are instances of H-SFC that meet the delay requirements found by s1 in the dynamic auxiliary graph through depth-first search, and are called s1's path set.

[0083] In some possible embodiments, the specific steps of constructing a dynamic auxiliary weighted graph that satisfies the order constraint include:

[0084] S31. Based on the sequence requirement of the H-SFC requested by the user, all nodes that can provide the VNFs required by the H-SFC are selected from the weighted graph, and the selected nodes are arranged into a sequence graph consisting of multiple columns according to the VNF type requirements and sequence requirements in the H-SFC.

[0085] S32. In the sequence diagram, each node in the previous column of nodes is connected to all nodes in the next column to form a link;

[0086] S33, using the Dijkstra algorithm based on the original physical network to calculate the relative cost weight and delay weight between two nodes in the sequence graph;

[0087] S34. If the delay weight exceeds the requested delay tolerance, the node is removed, and otherwise each value is assigned to the corresponding link.

[0088] In some possible embodiments, the nodes in each column are composed of functional nodes that can provide the same type of VNF.

[0089] In some possible embodiments, determining a set of all paths that meet the H-SFC delay requirement of the mobile user includes:

[0090] Use the depth-first search algorithm to calculate the set of all paths that meet the required H-SFC delay requirements;

[0091] Depth-first search is used to calculate the delay and relative cost consumption at the same time, and the node resource status is updated in real time during the search.

[0092] In some possible embodiments, the profile of the mobile user is the probability that the mobile user stays at different edge cloud nodes in time segment t.

[0093] In some possible embodiments, determining the primary link based on the path set specifically includes:

[0094] According to the location information in the updated configuration file, the remaining locations in the configuration file are linked to each path in the link set in sequence;

[0095] If the path meets the H-SFC delay requirement, this position is covered by the link; otherwise, the link set is traversed continuously;

[0096] Sum the user stay probabilities corresponding to each location in the location set covered by each path, and select the path corresponding to the maximum sum of probabilities as the main link;

[0097] If other locations in the configuration file are not overwritten, the link with the smallest relative cost in the path set is selected as the main link.

[0098] Example 2

[0099] This embodiment provides a hybrid service function chain pre-deployment system based on mobility awareness, including:

[0100] The parameter acquisition module is used to obtain the request instructions of the mobile user and determine the adapted link and node information based on the resource constraints of the edge cloud node;

[0101] Node objectification and assignment module, used to objectify nodes based on node information and construct edge weighted graphs;

[0102] A dynamic auxiliary weighted graph module is constructed to adjust the link and node objects and assign weighted edges based on the edge weighted graph, so as to construct a dynamic auxiliary weighted graph that satisfies the order constraint;

[0103] The mobile user preselected path set acquisition module is used to obtain the mobile user's profile based on the mobile user's request instruction, determine the highest probability location of the mobile user in the profile, and combine it with the dynamic auxiliary weighted graph to determine the set of all paths that meet the mobile user's H-SFC delay requirements;

[0104] The mobile user deployment path acquisition module is used to determine the main link based on the path set, update the edge weights of the dynamic auxiliary weighted graph and the location information in the configuration file, jump to step S4, and continue until all the access locations of the configuration file are deployed. The updated deployment path is used as the secondary path for this mobile user.

[0105] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A hybrid service function chain pre-deployment method based on mobile perception, characterized in that: The specific steps include: S1. Obtain a request instruction from a mobile user and determine the adapted link and node information based on the resource constraints of the edge cloud node; the request instruction from the mobile user includes the H-SFC sequence requirement, delay constraint, bandwidth requirement, and mobile profile; S2. Based on the node information, objectify the nodes and construct an edge weighted graph; S3. Based on the edge-weighted graph, adjust the link and node objects and assign weighted edges to construct a dynamic auxiliary weighted graph that satisfies the order constraint; The specific steps of constructing a dynamic auxiliary weighted graph that satisfies the order constraint include: S31. Based on the sequence requirement of the H-SFC requested by the user, all nodes that can provide the VNFs required by the H-SFC are selected from the weighted graph, and the selected nodes are arranged into a sequence graph consisting of multiple columns according to the VNF type requirements and sequence requirements in the H-SFC. S32. In the sequence diagram, each node in the previous column of nodes is connected to all nodes in the next column to form a link; S33, using the Dijkstra algorithm based on the original physical network to calculate the relative cost weight and delay weight between two nodes in the sequence graph; S34. If the delay weight exceeds the requested delay tolerance, the node is removed, and otherwise the values are assigned to the corresponding links; S4. Based on the request instruction of the mobile user, a profile of the mobile user is obtained, and the highest probability location of the mobile user in the profile is determined. In combination with the dynamic auxiliary weighted graph, a set of all paths that meet the H-SFC delay requirement of the mobile user is determined. Determining all the path sets that meet the H-SFC delay requirements of the mobile user includes: calculating all the path sets that meet the H-SFC delay requirements using a depth-first search algorithm; calculating the relative cost consumption while calculating the delay using the depth-first search, and updating the node resource status in real time during the search; S5. Determine the primary link based on the path set, update the edge weights of the dynamic auxiliary weighted graph and the location information in the configuration file, and execute step S4 until all access locations in the configuration file are deployed. Use the updated deployment path as the secondary path for this mobile user.

2. The method for pre-deploying a hybrid service function chain based on mobile perception according to claim 1, characterized in that: The objectification of the nodes specifically includes: Determine the VNF resources that can be shared in the node and the remaining resources of the node; The VNF resources include: VNF deployment resources and VNF processing resources; The sharing of the VNF resources occurs within the H-SFC and between multiple H-SFCs.

3. The method for pre-deploying a hybrid service function chain based on mobile perception according to claim 2, characterized in that: The constructing of the edge weighted graph specifically includes: Perform weighted processing on each node based on all VNFs required by H-SFC in the request instruction; If a node contains a VNF required by the H-SFC and the node's remaining resources are greater than the processing resource requirements of the corresponding VNF in the H-SFC, the VNF is split into a virtual link and a relative cost weight and delay weight are assigned to the virtual link. If the node does not have the VNF required by the H-SFC, but the node resources meet the deployment resource and processing resource requirements of the required VNF, the required VNF is first deployed on the functional node, and then the VNF is split into a virtual link, and the relative cost weight and delay weight are assigned to the virtual link respectively.

4. The method for pre-deploying a hybrid service function chain based on mobile perception according to claim 1, characterized in that: The nodes in each column are composed of functional nodes that can provide the same type of VNF.

5. The method for pre-deploying a hybrid service function chain based on mobile perception according to claim 1, characterized in that: The profile of the mobile user is the probability that the mobile user stays at different edge cloud nodes in a time segment.

6. The method for pre-deploying a hybrid service function chain based on mobile perception according to claim 1, characterized in that: Determining the primary link based on the path set specifically includes: According to the location information in the updated configuration file, the remaining locations in the configuration file are linked to each path in the link set in sequence; If the path meets the H-SFC delay requirement, this position is covered by the link; otherwise, the link set is traversed continuously; Sum the user stay probabilities corresponding to each location in the location set covered by each path, and select the path corresponding to the maximum sum of probabilities as the main link; If other locations in the configuration file are not overwritten, the link with the smallest relative cost in the path set is selected as the main link.

7. A hybrid service function chain pre-deployment system based on mobile perception, characterized in that: include: A parameter acquisition module is used to obtain the mobile user's request instructions and determine the adapted link and node information in combination with the resource constraints of the edge cloud node; the mobile user's request instructions include the H-SFC sequence requirements, delay constraints, bandwidth requirements, and mobile profile; Node objectification and assignment module, used to objectify nodes based on node information and construct edge weighted graphs; A dynamic auxiliary weighted graph module is constructed to adjust the link and node objects and assign weighted edges based on the edge weighted graph, so as to construct a dynamic auxiliary weighted graph that satisfies the order constraint; The specific steps of constructing a dynamic auxiliary weighted graph that satisfies the order constraint include: Based on the sequence requirements of the H-SFC requested by the user, all nodes that can provide the VNFs required by the H-SFC are selected from the weighted graph. The selected nodes are arranged into a sequence graph consisting of multiple columns according to the VNF type requirements and sequence requirements in the H-SFC. In the sequence diagram, each node in the previous column is connected to all nodes in the next column to form a link; Based on the original physical network, Dijkstra algorithm is used to calculate the relative cost weight and delay weight between two nodes in the sequence graph; If the delay weight exceeds the requested delay tolerance, the node is removed, and vice versa, each value is assigned to the corresponding link; The mobile user preselected path set acquisition module is used to obtain the mobile user's profile based on the mobile user's request instruction, determine the highest probability location of the mobile user in the profile, and combine it with the dynamic auxiliary weighted graph to determine the set of all paths that meet the mobile user's H-SFC delay requirements; Determining all the path sets that meet the H-SFC delay requirements of the mobile user includes: calculating all the path sets that meet the H-SFC delay requirements using a depth-first search algorithm; calculating the relative cost consumption while calculating the delay using the depth-first search, and updating the node resource status in real time during the search; The mobile user deployment path acquisition module is used to determine the primary link based on the path set, update the edge weights of the dynamic auxiliary weighted graph and the location information in the configuration file; based on the updated edge weights of the dynamic auxiliary weighted graph and the location information in the configuration file, determine the highest probability location where the mobile user may stay in the configuration file, and combine the dynamic auxiliary weighted graph to determine all paths that meet the H-SFC latency requirements required by the mobile user; until all access locations in the configuration file are deployed, the updated deployment path is used as the secondary path for this mobile user.

Images