Air-space-ground network service failure recovery method based on matching game

Abstract

The invention discloses an air-space-ground network service failure recovery method based on a matching game, and the method comprises the steps: dividing a time period into a plurality of time slots, and converting an air-space-ground network into a quasi-static network at each time slot; establishing a service function chain deployment model based on the reconfigurable time extension graph in a space-air-ground network scene; based on the service function chain deployment model, providing service function chain deployment and service recovery problems under the condition of resource failure, including faults of unmanned aerial vehicles and satellite nodes, faults of node-related links and strategies for processing resource faults; and processing unpredictable resource faults based on a matching game algorithm so as to redeploy the service function chain influenced by the faults. According to the invention, the efficient redeployment completion condition of the service function chain can be obtained in as less time as possible under the condition of resource failure.

Description

Technical Field

[0001] The present invention relates to the technical field of network resource allocation, and in particular to an air-ground-space network service failure recovery method based on matching game. Background Art

[0002] The space-ground network has been a hot topic in recent years and is a key component of the sixth generation of communication technology. The space-ground network is mainly composed of satellites, drones, high-altitude platforms and ground stations. The space-ground network can provide a large number of information services in environmental monitoring, traffic management, industry and agriculture, disaster relief, etc. However, as a multi-layer, highly heterogeneous network, the space-ground network has diverse resource types and complex resource structures, such as the highly dynamic periodic motion of satellite nodes, the flexibility and controllability of drone nodes, and the different resource capabilities of multiple network nodes. In addition, in order to complete a specific task, the resources of different nodes are limited. Therefore, different networks cannot share resources, resulting in poor task processing capabilities and low resource utilization. In addition, problems involving network failures and resource failures affect the execution and transmission of tasks in the space-ground network. These problems result in the inability to smoothly transmit or process tasks. This leads to an increase in task completion time and affects network efficiency. Therefore, effective and robust task recovery is a necessary means to deal with resource failures.

[0003] As mentioned above, network function virtualization technology is introduced in the air-space-ground network. By decoupling the software and hardware devices, the technical differences between various devices are shielded, thereby realizing the interconnection and resource sharing of the air-space-ground network. Among them, the virtual network function is a software application that implements the network function. A task contains multiple virtual network functions connected in sequence, which can be represented as a service function chain. Therefore, the deployment of the service function chain provides a new mechanism for the elastic resource management of network function virtualization, but it also brings the following new challenges.

[0004] For example, how to achieve a unified representation of resources across different domains and time periods is a key challenge that hinders the ability to create effective deployment models for service function chains. In addition, the heterogeneity and temporal changes of resources in different domains exacerbate the difficulty of standardizing resource representation, which is critical for the deployment of service function chains.

[0005] Highly dynamic networks are prone to network failures and resource failures. These instabilities may affect the deployment of service function chains, resulting in business interruptions and performance degradation. In order to maintain service function chain deployment in such a turbulent environment, it is critical to manage and mitigate these risks.

[0006] It is a challenge to design a robust recovery mechanism to handle resource failures and ensure the redeployment of service function chains. In the event of resource failures, it is critical to implement an effective fast recovery strategy to minimize time consumption and maintain service continuity. In addition, resource reallocation and effective redeployment strategies should also be considered. Summary of the invention

[0007] The purpose of the present invention is to provide an air-ground-space network service failure recovery method based on matching game. Aiming at the deployment and redeployment of tasks in heterogeneous multi-layer networks in the face of resource failure, a service failure recovery model based on matching game algorithm is designed, taking into account various constraints and other situations, in order to minimize the objective function problem of the time consumption for completing the deployment of all service function chains. The present invention can flexibly adjust the transmission path of the service function chain and replace the processing node in response to sudden resource failures, and the node resource utilization rate is high. In the case of resource failure, efficient service function chain redeployment completion can be obtained in the shortest possible time.

[0008] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0009] A method for recovering from air-ground-space network service failure based on matching game, the method comprising the following steps:

[0010] Step 1: Divide the air-ground network scenario into three layers, including low-Earth orbit satellites, drones, and ground stations; divide the time period into multiple time slots, and transform the air-ground network into a quasi-static network in each time slot. The drone nodes and satellite nodes are quasi-static in each time slot;

[0011] Step 2: Based on the quasi-static network in step 1, a service function chain deployment model based on a reconfigurable time extension graph in the air-ground network scenario is established;

[0012] Step 3: Based on the service function chain deployment model, the service function chain deployment and service recovery problems in the case of resource failure are proposed, including the failure of drone and satellite nodes, the failure of node-related links, and the strategy for handling resource failures;

[0013] Step 4: A matching game-based algorithm handles unpredictable resource failures to redeploy the service function chain affected by the failure.

[0014] In step 1, the air-ground network scenario is divided into three layers, including low-Earth orbit satellites, drones, and ground stations. A reconfigurable time extension graph is proposed, and the time period is divided into T small time slots t, t∈T, where t represents the length of each time slot, and the drone nodes and satellite nodes are quasi-static within time slot t.

[0015] Step 2 further includes:

[0016] Step 2-1: Create a network framework where ε=ε g ∪ε u ∪ε s Represent the ground nodes ε g 、UAV node ε u and LEO satellite node ε s A collection of; Represents the link, represents the set of links between two nodes, They correspond to the links from ground to drone G2U, satellite to ground S2G, drone to ground U2G, drone to drone U2U, drone to satellite U2S and satellite to satellite S2S respectively; i t ∈ε represents the i-th node in time slot t, j t ∈ε represents the jth node in time slot t; Represents the same node i t The storage link of the service function chain from the previous time slot t to the adjacent time slot t+1;

[0017] Set the service function chain to Where k is the serial number of the service function chain, is the total number of service function chains; represents the mth virtual network function in the kth service function chain, l k represents the total number of virtual network functions in the kth service function chain;

[0018] Step 2-2: Establish the transmission channel model of the air-ground network:

[0019] According to Shannon's formula, the maximum data rate from ground to drone G2U, drone to ground U2G, drone to drone U2U, and satellite to satellite S2G is expressed as:

[0020]

[0021] in express Respectively represent the maximum data rates of G2U, U2G, U2U, and S2G; B represents B gu , B ug , B uu , B sg , representing the bandwidth of G2U, U2G, U2U, and S2G respectively; Ω t express They represent the signal-to-noise ratios of G2U, U2G, U2U, and S2G, respectively;

[0022] For U2S and S2S channels, the maximum data rates are:

[0023]

[0024] in express Represent the maximum data rates of U2S and S2S, P ij , and Represent the transmit power, transmit gain and receive antenna gain respectively. represents the free space related fading, L l represents the total line loss, (E b / N0) r represents the ratio of the energy received per bit to the noise density, k B , T s and S m represent the Boltzmann constant, system noise temperature and maximum tilt range respectively;

[0025] Step 2-3: Establish the energy consumption model of the air-ground-space network, including:

[0026] The energy consumption model of the drone is:

[0027]

[0028] in, Represents the energy loss of the drone, represents the path energy loss, represents the transmission energy loss;

[0029] The energy loss model of the satellite is:

[0030] in, represents the energy loss of the satellite, represents the receiving energy loss, represents the transmission energy loss, Represents the operating energy loss.

[0031] Step 3 further includes:

[0032] Step 3-1: For each virtual network function in the service function chain, only one node can be selected to be deployed in a time slot:

[0033]

[0034] in Represents a virtual network function deployed on a node;

[0035] All service function chains need to complete the processing and deployment of all included virtual network functions as follows:

[0036]

[0037] For each service function chain, virtual network functions are deployed in order:

[0038]

[0039] in Represents a virtual network function The time slot, Represents a virtual network function The time slot, Represents the computing resources consumed by the virtual network function, Represents the computing power of the node;

[0040] The flow constraints are:

[0041]

[0042] in Represents the service function chain through the link (i t ,j t ), o t and d t Represent the source node and target node in the service function chain deployment respectively. The above three formulas represent the flow conservation of the service function chain at the source node, the flow conservation at the intermediate node, and the flow conservation at the target node. ;

[0043] When each service function chain is transmitted on the channel, the transmission resources consumed cannot exceed the capacity of the channel:

[0044]

[0045] where Δ k Represents the amount of data in the service function chain, Represents a link (i t ,j t )’s channel capacity;

[0046] When processing a service function chain on a node, the computing resources cannot exceed the computing capacity of the node:

[0047]

[0048] in Represents node i t computing power;

[0049] When the service function chain is not processed on the node and transmitted on the channel, the service function chain chooses to be stored on the node. The amount of data stored on the node cannot exceed the storage capacity of the node:

[0050]

[0051] in The virtual network functions representing the service function chains are temporarily stored on the nodes. Represents node i t Storage resource capacity;

[0052] The sum of the total computational energy consumption at a node and the energy consumption of node operations cannot exceed the energy capacity:

[0053]

[0054] in Represents the energy consumption of unit computing resources on the node, Represents the energy capacity of the node;

[0055] Step 3-2: For the service function chain, the time consumption includes transmission delay, processing delay, storage delay, delay caused by redeployment, and delay caused by redeployment of other service function chains:

[0056]

[0057] in represents the transmission delay, Represents processing delay, represents storage delay, represents the delay caused by redeployment, Represents the delay caused by the redeployment of other service function chains;

[0058]

[0059] in Represents the redeployment of the service function chain, Represents the latency of redeploying virtual network functions in a service function chain.

[0060] The optimization problem is:

[0061]

[0062] The parameters parameter

[0063] parameter parameter

[0064] Step 3-3: When a node failure causes the link connected to it to be unable to transmit, the affected service function chain searches for available drones and satellites and redeploys the virtual network function near the failed node.

[0065] Step 4 further includes:

[0066] Initialize service function chain deployment under normal circumstances and service function chain recovery in case of resource failure;

[0067] When the service function chain is not affected by the faulty node, it is deployed according to the initialization path;

[0068] When multiple service function chains arrive at the same node at the same time and are ready to be processed, the service function chains are sorted from small to large according to the amount of data; the node selects the service function chains in turn until its capacity limit; the remaining service function chains wait and are stored on the node, and are evaluated and processed in the next time slot; when a node failure affects the service function chain, resource business recovery and service function chain redeployment are completed through a matching algorithm; when all service function chains are completed, the total time consumed is calculated.

[0069] Furthermore, in step 4, the process of initializing the service function chain deployment includes the following steps:

[0070] Get the initial positions of the drone and satellite nodes, as well as the data volume, start node, and end node of the service function chain; for each service function chain, find the drone node closest to the start and end points; set the distance from the start node to itself to 0, and the distance to all other nodes to infinity;

[0071] Create a collection C for recording unvisited nodes uv , computing node i s To set C uv The distance d between midpoints sc , and the distance d sc The shortest node is set to the current node i c ; For node i c Each neighbor node i n , calculate from node i s Start, go to neighbor node i n , and passes through the current node i c The total distance d new ; if d new Less than d sc , then update the path and distance of the neighboring nodes;

[0072] Repeat the above steps until all nodes are visited and a set of the shortest paths from the start node to the destination node is obtained.

[0073] Furthermore, in step 4, the process of redeploying the service function chain based on the matching game algorithm includes the following steps:

[0074] Set the service function chain as the matching proposer and the node as the matching receiver, and build the preference list of the service function chain and the node; the preference list L of the service function chain S It is the sorting of the time consumed from the current node to the destination node, that is:

[0075]

[0076] in Indicates that the service function chain starts from the current node To any next neighbor node The time consumed, Represents a slave node To the destination node The time taken to travel the shortest distance;

[0077] The node's preference list L N The order of the service function chain based on the selection of this node is expressed as:

[0078]

[0079] Among them, a, b, and c are weighted parameters, a has the largest value and c has the smallest value. Indicates that the service function chain is being processed; Indicates that the service function chain is being stored; Δ k The amount of data representing the service function chain;

[0080] In each cycle, a non-matching service function chain is randomly selected According to the preference list L S , select L S The previous node if If the capacity is insufficient, the service function chain Assigned to Otherwise, the current matching list Select the priority list The lowest ranking service function chain and with Compare; if express exist L N Ranked higher than but Add to matching list delete

[0081] When all service function chains that need to be redeployed are matched with nodes or the node capacity is exhausted, the loop ends, the matching information of the redeployed service function chain and the node is obtained, and the service function chain is transmitted to the matching node through the corresponding link of the current time slot.

[0082] Compared with the prior art, the present invention has the following beneficial effects:

[0083] First, the air-ground-space network service failure recovery method based on matching game of the present invention aims at the problem of deployment and redeployment of tasks in heterogeneous multi-layer networks facing resource failures, and designs a service failure recovery model based on matching game algorithm. It takes into account various constraints and other situations to solve the objective function problem of minimizing the time consumption for completing the deployment of all service function chains.

[0084] Second, the air-space-ground network service failure recovery method based on matching game of the present invention uses matching game algorithm to solve the problem, comprehensively considers the status information of the service function chain and the node, and can flexibly adjust the transmission path of the service function chain and replace the processing node in response to sudden resource failures, with high node resource utilization.

[0085] Thirdly, the air-ground-space network service failure recovery method based on matching game of the present invention can achieve efficient service function chain redeployment completion in the shortest possible time under the condition of resource failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 This is a schematic diagram of the deployment scenario of the air-ground-space network service function chain of the present invention;

[0087] Figure 2 A schematic diagram of the redeployment of the space-time-space-earth-space network service function chain for failed nodes and links of the present invention;

[0088] Figure 3 This is a diagram showing the deployment status and time slot relationship experimental simulation results of the service function chain in the present invention;

[0089] Figure 4 This is a simulation result diagram of the relationship between the number of service function chains and the total time consumption in the present invention;

[0090] Figure 5 This is a diagram of the experimental simulation results of the relationship between the number of drones and the maximum completion time in the present invention.

[0091] Figure 6 This is a diagram of experimental simulation results of the ratio between the number of drones in the present invention and the number of service function chains being processed and unfinished in each time slot. DETAILED DESCRIPTION

[0092] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings.

[0093] like Figure 1-2 As shown, the present invention provides a method for recovering air-ground network service failure based on matching game, comprising the following steps:

[0094] Step 1: Divide the time period into multiple time slots, and transform the air-ground network into a quasi-static network in each time slot;

[0095] Step 2: According to step 1, a service function chain deployment model based on a reconfigurable time extension graph in an air-ground-space network scenario is established;

[0096] Step 3: A service function chain deployment and service recovery problem under resource failure is proposed, including the failure of drone and satellite nodes, the failure of node-related links, and the strategy for handling resource failures;

[0097] Step 4: A matching game-based algorithm is proposed to handle unpredictable resource failures, so that the service function chain affected by the failure can be deployed smoothly and quickly.

[0098] In the embodiment, step 1 divides the time period into multiple time slots, and converts the air-ground-space network into a quasi-static network in each time slot:

[0099] The air-ground network scenario is divided into three layers, including low earth orbit satellites, drones, and ground stations. First, a reconfigurable time extension graph is proposed, and the time period is divided into T small time slots t, t∈T, where τ represents the length of each time slot. When the time slot is small enough, it can be assumed that the drone nodes and satellite nodes are quasi-static within the time slot t.

[0100] In the embodiment, step 2 establishes a service function chain deployment model based on a reconfigurable time extension graph in an air-ground-space network scenario, including:

[0101] Step 2-1: Create a network framework where ε=ε g ∪ε u ∪ε s Represent the ground nodes ε g 、UAV node ε u and LEO satellite node ε s Collection, i t ∈ε represents the i-th node in time slot t; represents a link, where represents the link between two nodes, corresponding to ground to drone (G2U), satellite to ground (S2G), drone to ground (U2G), drone to drone (U2U), drone to satellite (U2S), satellite to satellite (S2S), and and Represents the same node i t The storage link of the service function chain from time slot t to the adjacent time slot t+1.

[0102] The service function chain is set to Where k is the serial number of the service function chain, in is the total number of service function chains. represents the mth virtual network function in the kth service function chain, l k Represents the total number of virtual network functions in the kth service function chain.

[0103] Step 2-2: Establish the transmission channel model of the air-ground network, including:

[0104] The maximum data rates of G2U, U2G, U2U, and S2G are expressed according to the Shannon formula:

[0105]

[0106] in Represents the maximum data rates of G2U, U2G, U2U, and S2G respectively. B=B gu ∪B ug ∪B uu ∪B sg Represents the bandwidth of G2U, U2G, U2U, and S2G. Represents the signal-to-noise ratio of G2U, U2G, U2U, and S2G.

[0107] For U2S and S2S channels, the maximum data rate is:

[0108]

[0109] in Represents the maximum data rate of U2S and S2S. ij , and They represent transmit power, transmit gain and receive antenna gain respectively. represents the relative attenuation in free space. L l Represents the total line loss. (E b / N0) r Represents the ratio of the energy received per bit to the noise density. B , T s and S m They represent the Boltzmann constant, system noise temperature and maximum tilt range respectively.

[0110] Step 2-3: Establish the energy consumption model of the air-ground-space network, including:

[0111] The energy consumption model of the drone is:

[0112]

[0113] in represents the path energy loss, Represents the transmission energy loss.

[0114] The energy loss model of the satellite is:

[0115]

[0116] in represents the receiving energy loss, represents the transmission energy loss, Represents the operating energy loss.

[0117] In the embodiment, step 3 proposes a service function chain deployment and service recovery problem in the case of resource failure, including failures of drones and satellite nodes, and failures of node-related links, and strategies for handling resource failures, including:

[0118] Step 3-1: Define that for each virtual network function in the service function chain, only one node can be selected to be deployed in a time slot:

[0119]

[0120] in Represents a virtual network function deployed on a node.

[0121] Defining all service function chains requires the processing and deployment of all included virtual network functions:

[0122]

[0123] Definition For each service function chain, the virtual network functions need to be deployed in sequence, i.e. the next virtual network function must be processed after the deployment of the previous virtual network function is completed:

[0124]

[0125] in Represents the computing resources consumed by the virtual network function. Represents the computing power of the node.

[0126] Define the flow constraints as:

[0127]

[0128] Among them t and d t They represent the source node and target node in the service function chain deployment respectively. Represents the service function chain passing through the link.

[0129] Define that when each service function chain is transmitted on a channel, the transmission resources consumed cannot exceed the capacity of the channel:

[0130]

[0131] where Δ k Represents the quantity of service function chains.

[0132] When defining a service function chain on a node, the computing resources cannot exceed the computing capacity of the node:

[0133]

[0134] Defines that when the service function chain is not processed on this node and transmitted on the channel, the service function chain chooses to store on this node. The amount of data stored on the node cannot exceed the storage capacity of the node:

[0135]

[0136] in The representative service function chain is stored on the node.

[0137] It is defined that the sum of the total computational energy consumption of a node and the energy consumption of node operations cannot exceed the energy capacity:

[0138]

[0139] in Represents the energy consumption of unit computing resources on the node.

[0140] Step 3-2: Define the optimization problem to minimize the total time delay for processing and redeploying all service function chains. For a service function chain, the time consumption includes transmission delay, processing delay, storage delay, delay caused by redeployment, and delay caused by redeployment of other service function chains:

[0141]

[0142] in represents the transmission delay, Represents processing delay, represents storage delay, represents the delay caused by redeployment, Represents the delay caused by the redeployment of other service function chains.

[0143] So the optimization problem is:

[0144]

[0145] The parameters parameter

[0146] parameter parameter

[0147] The virtual network functions of the service function chain can be deployed on any spatial node or empty node. However, in reality, nodes may fail due to equipment failure, hardware damage, etc. Therefore, consider the failure of virtual network function processing caused by node failure, and the failure of the link connected to the corresponding failed node. Therefore, it is necessary to redeploy and schedule the service function chain on these failed nodes and links to restore the business. If there is a failure of a drone and a satellite node, the link connected to it cannot be transmitted. Therefore, the virtual network functions originally deployed on these two nodes need to be redeployed. The affected service function chain searches for available drones and satellites and redeploys the virtual network functions near the failed node. Therefore, the original deployment path is abandoned and a new deployment path is obtained. The service function chain can successfully complete the deployment task.

[0148] In the embodiment, step 4 proposes an algorithm based on matching game to handle unpredictable resource failures, so that the service function chain affected by the failure can be deployed smoothly and quickly, including:

[0149] The optimization problem in the case of resource failure is handled based on matching game. The entire algorithm for redeployment of service function chains in the case of resource failure is described as follows: first initialization, deployment of service function chains under normal circumstances, and recovery of service function chains in the case of resource failure. When the service function chain is not affected by the faulty node, it is deployed according to the initialization path. When multiple service function chains arrive at the same node at the same time and are ready to be processed, these service function chains are sorted from small to large according to the amount of data. Then, the node selects the service function chain in turn until its capacity limit. The remaining service function chains wait and are stored on the node for evaluation and processing in the next time slot. When the node failure affects the service function chain, the resource business recovery and redeployment of the service function chain are completed through the matching algorithm. When all service function chains are completed, the total time consumption is calculated.

[0150] First, initialize the service function chain deployment without considering node failure. The first step is to obtain the initial positions of the drone and satellite nodes, as well as the data volume, start node and end node of the service function chain. For each service function chain, find the drone node closest to the start point and the end point. First, set the distance from the start node to itself to 0 and the distance to all other nodes to infinity. Create a set C for recording unvisited nodes uv Then, calculate i s to C uv The distance between the midpoint and the distance d sc The shortest node is set to the current node i c For i c Each neighbor node i n , computing node i s to n Passing the current node i c The distance d new If d new Less than d sc , then update the path and distance of the neighboring nodes. Repeat the above steps until all nodes are visited. Finally, a set of the shortest paths from the starting node to the destination node is obtained.

[0151] When a node fails, all links connected to it also fail. Therefore, the service function chains transmitted on these links can only return to the previous node to find the next available node. Therefore, an algorithm based on matching game is designed to recover these service function chains to be redeployed, as follows:

[0152] In the matching game, there are two players. One player sends a match request, and the other chooses to accept or reject the match request. The desired result is that all players are satisfied with their opponents. The matching reaches a stable state, and no two players are willing to break the current matching result and form a new matching pair. Each party has a preference list of all players of the other party, that is, who it prefers to match. Set the service function chain as the match proposer and the node as the match acceptor. Therefore, there are two preference lists for the two matching parties, the service function chain and the node.

[0153] First, a preference list of service function chains and nodes is constructed. Due to the optimization problem, that is, the completion time of service function chain deployment should be as short as possible, the preference list L of the service function chain is S The time sequence from the current node to the next node Right now:

[0154]

[0155] in Indicates the current node arrive The time consumed to travel the distance. express To the destination node It is worth noting that since the optimization problem is related to the time consumed by the path, the premise for a node to be included in the priority list is that the current node and the next node can communicate and transmit the service function chain.

[0156] In order to minimize the total delay of the optimization problem and reduce time waste, the node tends to select the service function chain being processed on the node and the service function chain that is already waiting on the node. Therefore, the node's preference list L N The order of the service function chain based on the selection of this node is expressed as:

[0157]

[0158] Among them, a, b, and c are weighted parameters, with a being the largest and c being the smallest. Indicates that the service function chain is processing. Indicates that the service function chain is being stored. N The specific explanation is that the service function chain being processed is in the front, and the service function chain stored on the node is in the back. Other service function chains are arranged according to the data volume Δ k Sort from small to large. In each cycle, randomly select an unmatched service function chain According to the preference list L S , select L S The previous node if If the capacity is insufficient, the service function chain Assigned to Otherwise, the current matching list Select the priority list The lowest ranking service function chain and with For comparison. in express exist L N Ranked higher than but Add to matching list delete The cycle ends until all service function chains that need to be redeployed are matched with nodes or the node capacity is exhausted. Thus, the matching information between the service function chains that need to be redeployed and the nodes is obtained. The service function chain can be successfully transmitted to the matching node through the corresponding link of the current time slot.

[0159] The present invention uses a matching game algorithm to solve the air-ground network service failure recovery problem, and compares and analyzes it with three other methods, as follows:

[0160] The first algorithm involves sorting the service function chains in the preference list from large to small according to the data volume (FLTS), which is different from the algorithm proposed in the present invention. The second algorithm is that when multiple service function chains arrive at the same time, the node randomly selects the service function chain to be processed (RSSP). The third algorithm is that when a service function chain is affected by a node failure, the next node is randomly selected for transmission (RSNT). Specifically, the difference between RSSP and RSNT is that the node randomly selects the service function chain in RSSP, while the service function chain randomly selects the node in RSNT.

[0161] Figure 3 The figure is a simulation result diagram of the experimental deployment status and time slot relationship of the service function chain in the present invention. It can be seen that when the conditions of all variables and fault nodes are the same, the algorithm has different deployment processes for a specific number of service function chains. The algorithm proposed in the present invention requires the shortest time to complete the deployment of all service function chains, while RSNT requires the longest time. The time consumption of the other two algorithms FLTS and RSSP is between the algorithm proposed in the present invention and RSNT. It can be observed that RSNT has experienced multiple long time periods in which the ordinate value has not changed, that is, no new service function chain has been completed, which has a great impact on the overall time consumption.

[0162] Figure 4 The figure is a simulation result of the experiment on the relationship between the number of service function chains and the total time consumption in the present invention. Comparing the total time consumption when all service function chains are deployed, the algorithm proposed in the present invention consumes the least time, and RSNT consumes the most time. As the number of service function chains increases, the total time consumption increases. The time increase of the algorithm proposed in the present invention, FLTS and RSSP is less, while the time increase of RNST is greater.

[0163] Figure 5 The figure is the experimental simulation result of the relationship between the number of drones and the maximum completion time in the present invention. The maximum completion time of the service function chain is compared. The algorithm proposed in the present invention takes the shortest time, and the RSNT algorithm takes the longest time, and takes more time than other algorithms. Moreover, when the number of drones is small, the effect of RSNT is the worst, that is, the algorithm has poor adaptability to small-scale networks, while the algorithm proposed in the present invention can adapt to networks of various sizes.

[0164] Figure 6The figure is the experimental simulation result of the relationship between the number of drones and the number of service function chains being processed and unfinished in each time slot in the present invention. The ratio of the service function chains being processed in each time slot to all unfinished service function chains in the four algorithms is compared. RSSP has the highest utilization rate, and RSNT has the lowest utilization rate, which is much lower than the other three algorithms. This is because in RSSP, nodes randomly select the service function chains that arrive and do not reject any service function chain, so the proportion of service function chains processed in each time slot is relatively large. However, compared with Figure 4 and Figure 5 When compared together, it is noted that this random selection of service function chains does not help to reduce the overall time consumption. Therefore, the algorithm proposed in the present invention is the best in terms of time consumption and node utilization.

[0165] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0166] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0167] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0168] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0169] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0170] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for recovering air-ground network service failure based on matching game, characterized in that: The air-ground-space network service failure recovery method comprises the following steps: Step 1: Divide the air-ground network scenario into three layers, including low-Earth orbit satellites, drones, and ground stations; divide the time period into multiple time slots, and transform the air-ground network into a quasi-static network in each time slot. The drone nodes and satellite nodes are quasi-static in each time slot; Step 2: Based on the quasi-static network in step 1, a service function chain deployment model based on a reconfigurable time extension graph in the air-ground network scenario is established; Step 3: Based on the service function chain deployment model, the service function chain deployment and service recovery problems in the case of resource failure are proposed, including the failure of drone and satellite nodes, the failure of node-related links, and the strategy for handling resource failures; Step 4: A matching game-based algorithm handles unpredictable resource failures to redeploy the service function chains affected by the failures.

2. The method for recovering air-ground network service failure based on matching game according to claim 1, characterized in that: In step 1, the air-ground network scenario is divided into three layers, including low-Earth orbit satellites, drones, and ground stations. A reconfigurable time extension graph is proposed, and the time period is divided into T small time slots t, t∈T, where τ represents the length of each time slot, and the drone nodes and satellite nodes are quasi-static within the time slot t.

3. The method for recovering air-ground network service failure based on matching game according to claim 1, characterized in that: Step 2 further includes: Step 2-1: Create a network framework where ε=ε g ∪ε u ∪ε s Represent the ground nodes ε g 、UAV node ε u and LEO satellite node ε s A collection of; Represents the link, represents the set of links between two nodes, They correspond to the links from ground to drone G2U, satellite to ground S2G, drone to ground U2G, drone to drone U2U, drone to satellite U2S and satellite to satellite S2S respectively; i t ∈ε represents the i-th node in time slot t, j t ∈ε represents the jth node in time slot t; Represents the same node i t The storage link of the service function chain from the previous time slot t to the adjacent time slot t+1; Set the service function chain to Where k is the serial number of the service function chain, is the total number of service function chains; represents the mth virtual network function in the kth service function chain, l k represents the total number of virtual network functions in the kth service function chain; Step 2-2: Establish the transmission channel model of the air-ground network: According to Shannon's formula, the maximum data rate from ground to drone G2U, drone to ground U2G, drone to drone U2U, and satellite to satellite S2G is expressed as: in express Respectively represent the maximum data rates of G2U, U2G, U2U, and S2G; B represents B gu , B ug , B uu , B sg , representing the bandwidth of G2U, U2G, U2U, and S2G respectively; Ω t express They represent the signal-to-noise ratios of G2U, U2G, U2U, and S2G, respectively; For U2S and S2S channels, the maximum data rates are: in express Represents the maximum data rate of U2S and S2S respectively; P ij , and Represent the transmit power, transmit gain and receive antenna gain respectively. represents the free space related fading, L l represents the total line loss, (E b / N0) r represents the ratio of the energy received per bit to the noise density, k B , T s and S m represent the Boltzmann constant, system noise temperature and maximum tilt range respectively; Step 2-3: Establish the energy consumption model of the air-ground-space network, including: The energy consumption model of the drone is: in, Represents the energy loss of the drone, represents the path energy loss, represents the transmission energy loss; The energy loss model of the satellite is: in, represents the energy loss of the satellite, represents the receiving energy loss, represents the transmission energy loss, Represents the operating energy loss.

4. The method for recovering air-ground network service failure based on matching game according to claim 3, characterized in that: Step 3 further includes: Step 3-1: For each virtual network function in the service function chain, only one node can be selected to be deployed in a time slot: in Represents a virtual network function deployed on a node; All service function chains need to complete the processing and deployment of all included virtual network functions as follows: For each service function chain, virtual network functions are deployed in order: in Represents a virtual network function The time slot, Represents a virtual network function The time slot, Represents the computing resources consumed by the virtual network function, Represents the computing power of the node; The flow constraints are: in Represents the service function chain through the link (i t ,j t ), o t and d t They represent the source node and the target node in the deployment of the service function chain respectively. The three formulas of the flow constraint represent the flow conservation of the service function chain at the source node, the flow conservation at the intermediate node, and the flow conservation at the target node respectively. When each service function chain is transmitted on the channel, the transmission resources consumed cannot exceed the capacity of the channel: Where Δ k Represents the amount of data in the service function chain, Represents a link (i t ,j t )’s channel capacity; When processing a service function chain on a node, the computing resources cannot exceed the computing capacity of the node: in Represents node i t computing power; When the service function chain is not processed on the node and transmitted on the channel, the service function chain chooses to be stored on the node. The amount of data stored on the node cannot exceed the storage capacity of the node: in The virtual network functions representing the service function chains are temporarily stored on the nodes. Represents node i t Storage resource capacity; The sum of the total computational energy consumption at a node and the energy consumption of node operations cannot exceed the energy capacity: in Represents the energy consumption of unit computing resources on the node, represents the energy capacity of the node; Step 3-2: For the service function chain, the time consumption includes transmission delay, processing delay, storage delay, delay caused by redeployment, and delay caused by redeployment of other service function chains: in represents the transmission delay, Represents processing delay, represents storage delay, represents the delay caused by redeployment, Represents the delay caused by the redeployment of other service function chains; in Represents the redeployment of the service function chain, represents the latency of redeploying the virtual network functions in the service function chain; The optimization problem is: The parameters parameter parameter parameter Step 3-3: When a node failure causes the link connected to it to be unable to transmit, the affected service function chain searches for available drones and satellites and redeploys the virtual network function near the failed node.

5. The method for recovering air-ground network service failure based on matching game according to claim 1, characterized in that: Step 4 further includes: Initialize service function chain deployment under normal circumstances and service function chain recovery in case of resource failure; When the service function chain is not affected by the faulty node, it is deployed according to the initialization path; When multiple service function chains arrive at the same node at the same time and are ready to be processed, the service function chains are sorted from small to large according to the amount of data; the node selects the service function chains in turn until its capacity limit; the remaining service function chains wait and are stored on the node, and are evaluated and processed in the next time slot; when a node failure affects the service function chain, resource business recovery and service function chain redeployment are completed through a matching algorithm; when all service function chains are completed, the total time consumed is calculated.

6. The method for recovering air-ground network service failure based on matching game according to claim 5, characterized in that: In step 4, the process of initializing the service function chain deployment includes the following steps: Get the initial positions of the drone and satellite nodes, as well as the data volume, start node, and end node of the service function chain; for each service function chain, find the drone node closest to the start and end points; set the distance from the start node to itself to 0, and the distance to all other nodes to infinity; Create a collection C for recording unvisited nodes uv , computing node i s To set C uv The distance d between midpoints sc , and the distance d sc The shortest node is set to the current node i c ; For node i c Each neighbor node i n , calculate from node i s Start, go to neighbor node i n , and passes through the current node i c The total distance d new ; if d new Less than d sc , then update the path and distance of the neighboring nodes; Repeat the above steps until all nodes are visited and a set of the shortest paths from the start node to the destination node is obtained.

7. The method for recovering air-ground network service failure based on matching game according to claim 5, characterized in that: In step 4, the process of redeploying the service function chain based on the matching game algorithm includes the following steps: Set the service function chain as the matching proposer and the node as the matching receiver, and build the preference list of the service function chain and the node; the preference list L of the service function chain S It is the sorting of the time consumed from the current node to the destination node, that is: in Indicates that the service function chain starts from the current node To any next neighbor node The time consumed, Represents a slave node To the destination node The time taken to travel the shortest distance; The node's preference list L N The order of the service function chain based on the selection of this node is expressed as: Among them, a, b, and c are weighted parameters, a has the largest value and c has the smallest value. Indicates that the service function chain is being processed; Indicates that the service function chain is being stored; Δ k The amount of data representing the service function chain; In each cycle, a non-matching service function chain is randomly selected According to the preference list L S , select L S The previous node if If the capacity is insufficient, the service function chain Assigned to Otherwise, the current matching list Select the priority list The lowest ranking service function chain and with To compare; if express exist L N Ranked higher than but Add to matching list delete When all service function chains that need to be redeployed are matched with nodes or the node capacity is exhausted, the loop ends, the matching information of the redeployed service function chain and the node is obtained, and the service function chain is transmitted to the matching node through the corresponding link of the current time slot.

Images