A Service Function Chain Orchestration Method and Device Applicable to Low Earth Orbit Satellite Networks

Abstract

The present invention provides a service function chain orchestration method and apparatus applicable to low-Earth orbit satellite networks, which relates to the technical field of low-Earth orbit satellite communications. The essence of the present invention is a distributed point-by-point service function chain orchestration scheme, which orchestrates service function chain (SFC) requests by sequentially embedding a series of virtual network functions (VNFs) into a series of adjacent satellite nodes. Therefore, the computing power requirement for the centralized controller is relatively low, and it is applicable to large-scale low-Earth orbit satellite networks without a stable centralized controller. During the distributed deployment of VNFs, first determine the set of candidate satellite nodes for deploying the next VNF, and then use the actor-critic network model integrating decision transformers to process the trajectories of the most recently deployed C VNFs to select a second satellite node for deploying the next VNF from the set of candidate satellite nodes. That is to say, this method can determine a more robust VNF deployment decision based on global context information, avoiding the problem of local optimality.

Description

Technical Field

[0001] The present invention relates to the technical field of low-earth orbit satellite communication, and in particular, to a service function chain orchestration method and device applicable to a low-earth orbit satellite network. Background Art

[0002] In recent years, low-earth orbit satellite networks have attracted extensive attention in the academic and industrial communities. Low-earth orbit satellite networks offer advantages such as wide coverage, large-scale access, rapid response, and cost-effective deployment. Therefore, their utilization rate has increased significantly exponentially. However, with the rapid expansion of the applications of low-earth orbit satellite networks, it has become increasingly important to meet service differentiation based on user needs. Each application has unique quality of service requirements in terms of data rate, latency, reliability, and coverage.

[0003] The advancement of service function chains has become a key technology to address this challenge. A service function chain is a network paradigm concept that combines virtual network functions such as firewalls, load balancers, and deep packet inspection in a predefined order to apply a set of network services. By orchestrating service function chains, traffic can be guided through the required network functions in a specific order to achieve service-specific quality of service requirements. Orchestrating service function chains involves managing and coordinating various virtual network functions to achieve the required service functions and ensure seamless operation. However, due to the exponential growth of the possible combinations and permutations of virtual network functions, finding an optimal service function chain orchestration strategy is a challenging task.

[0004] In the prior art, some researchers have proposed modeling the 5G network service function chain orchestration problem as an integer linear programming problem and deploying a rounding-based approximation algorithm on a centralized controller to generate solutions. However, traditional methods rely heavily on a centralized controller to generate optimal strategies. This algorithm has high computing power requirements for the controller and is not applicable to large-scale low-earth orbit satellite networks without a stable centralized controller. Summary of the Invention

[0005] The purpose of the present invention is to provide a service function chain orchestration method and device applicable to a low-earth orbit satellite network, so as to alleviate the technical problems existing in the existing service function chain orchestration methods, such as high computing power requirements for the centralized controller and inapplicability to large-scale low-earth orbit satellite networks without a stable centralized controller.

[0006] First aspect, the present invention provides a service function chain orchestration method applicable to a low-earth orbit satellite network, including: Step S102, receiving a service function chain (SFC) request and resetting the low-earth orbit satellite network environment; wherein, the SFC request includes: a source base station, a destination base station, and a plurality of virtual network functions (VNFs) orchestrated in a specified order; Step S104, calculating an orchestration direction unit vector of the SFC request according to the position coordinates of the source base station and the position coordinates of the destination base station; Step S106, randomly selecting a satellite node within the communication range of the source base station, and deploying the first VNF on the satellite node when it is determined that the satellite node meets the deployment conditions of the first VNF in the SFC request; Step S108, determining a set of candidate satellite nodes for deploying the next VNF based on the orchestration direction unit vector, the position coordinates of the first satellite node where the currently last deployed VNF is located, and the position coordinates of the neighbor satellite nodes of the first satellite node; Step S110, processing the trajectory points of the most recently deployed C VNFs using the actor-critic network model of the integrated decision transformer to select a second satellite node for deploying the next VNF from the set of candidate satellite nodes; wherein, each VNF trajectory point includes: the satellite node number where the VNF is deployed, the resource status of the satellite node, and the reward obtained from deploying the VNF; Step S112, calculating the SFC orchestration duration when it is determined that the second satellite node meets the deployment conditions of the next VNF; Step S114, when it is determined that the SFC orchestration duration is less than or equal to a preset duration threshold and all VNFs in the SFC request have not been deployed, return to Step S108 until all VNFs in the SFC request are deployed.

[0007] Optionally, determining a set of candidate satellite nodes for deploying the next VNF based on the orchestration direction unit vector, the position coordinates of the first satellite node where the currently last deployed VNF is located, and the position coordinates of the neighbor satellite nodes of the first satellite node includes: dividing the spherical area covered by the communication of the first satellite node into four quadrants based on the equal division principle and the position coordinates of the first satellite node to obtain four communication areas, and determining the center point coordinates of each communication area; calculating the unit vector of each communication area based on the center point coordinates of each communication area and the position coordinates of the first satellite node; calculating the dot product of the unit vector of each communication area and the orchestration direction unit vector, and sorting the four communication areas in descending order based on the dot product result; when it is determined that there are neighbor satellite nodes in the target communication area ranked first based on the position coordinates of the neighbor satellite nodes of the first satellite node, constructing a set of candidate satellite nodes based on all the neighbor satellite nodes in the target communication area; otherwise, sequentially select the communication areas ranked lower until a set of candidate satellite nodes is determined.

[0008] Optionally, calculating the SFC orchestration duration includes: using the formula Calculate the SFC scheduling duration; where, represents the SFC scheduling duration of the SFC request, represents the scheduling time consumption of the current round, represents the scheduling time consumption of the historical round; , L represents the set of physical network links, represents the set of virtual network links, represents the one-hop transmission delay from satellite node u to satellite node v at time slot t, represents the VNF virtual link deployed on the physical network link uv, represents the VNF virtual link not deployed on the physical network link uv, represents the processing delay of satellite node i, represents that the k-th VNF is deployed at satellite node i, represents that the k-th VNF is not deployed at satellite node i, represents the total number of satellites in the low-earth orbit satellite network, represents the set of all VNFs in the SFC request.

[0009] Optionally, after selecting the second satellite node for deploying the next VNF in the set of optional satellite nodes, it further includes: Step S111, based on the formula update the network parameters of the actor network in the actor-critic network model; based on the formula update the network parameters of the critic network; where, represents the network parameters of the actor network, represents the resource status of satellite node i at the current time slot t, represents that the VNF is deployed at satellite node i at the current time slot t, represents the old policy of the actor network , , represents the important weight of the relative difference between the old and new policies of the actor network, represents the clipping hyperparameter, represents the new policy of the actor network, , represents the advantage function calculated by the TD error, represents the discount factor, represents the reward obtained when the VNF is deployed at satellite node m at time slot x, and t - u + 1 = C, represents the value function of the critic network in the actor-critic network model, represents the network parameters of the critic network.

[0010] Optionally, after selecting a second satellite node for deploying the next VNF from the set of optional satellite nodes, the method further includes: Step S113, in the case where it is determined that the second satellite node does not meet the deployment conditions of the next VNF, releasing all the deployed VNFs in the low-earth orbit satellite network, returning to Step S106, and updating the SFC orchestration duration.

[0011] Optionally, after calculating the SFC orchestration duration, the method further includes: Step S115, in the case where it is determined that the SFC orchestration duration is greater than a preset duration threshold, determining that the SFC request orchestration fails.

[0012] In a second aspect, the present invention provides a service function chain orchestration device applicable to a low-earth orbit satellite network, including: a receiving and resetting module, configured to receive a service function chain SFC request and reset the low-earth orbit satellite network environment; where the SFC request includes: a source base station, a destination base station, and a plurality of virtual network functions VNFs orchestrated in a specified order; a first calculation module, configured to calculate an orchestration direction unit vector of the SFC request according to the position coordinates of the source base station and the position coordinates of the destination base station; a deployment module, configured to randomly select a satellite node within the communication range of the source base station, and in the case where it is determined that the satellite node meets the deployment conditions of the first VNF in the SFC request, deploy the first VNF on the satellite node; a determination module, configured to determine a set of optional satellite nodes for deploying the next VNF based on the orchestration direction unit vector, the position coordinates of the first satellite node of the currently last deployed VNF, and the position coordinates of the neighbor satellite nodes of the first satellite node; a selection module, configured to process the trajectory points of the most recently deployed C VNFs by using the actor-critic network model of the integrated decision transformer to select a second satellite node for deploying the next VNF from the set of optional satellite nodes; where each VNF trajectory point includes: the satellite node number of the deployed VNF, the resource status of the satellite node, and the reward obtained by deploying the VNF; a second calculation module, configured to calculate the SFC orchestration duration in the case where it is determined that the second satellite node meets the deployment conditions of the next VNF; a return module, configured to, in the case where it is determined that the SFC orchestration duration is less than or equal to the preset duration threshold and all the VNFs in the SFC request have not been deployed, return to call the determination module until all the VNFs in the SFC request are deployed.

[0013] Optionally, the determining module is specifically configured to: divide the spherical area covered by the communication of the first satellite node into four quadrants based on the equal division principle and the position coordinates of the first satellite node to obtain four communication areas, and determine the center point coordinates of each communication area; calculate the unit vector of each communication area based on the center point coordinates of each communication area and the position coordinates of the first satellite node; calculate the dot product of the unit vector of each communication area and the unit vector of the scheduling direction, and sort the four communication areas in descending order based on the dot product result; when it is determined that there are neighbor satellite nodes in the target communication area ranked first based on the position coordinates of the neighbor satellite nodes of the first satellite node, construct an optional satellite node set based on all the neighbor satellite nodes in the target communication area; otherwise, sequentially select the communication areas ranked later until an optional satellite node set is determined.

[0014] In a third aspect, the present invention provides an electronic device, including a memory and a processor. A computer program that can run on the processor is stored on the memory. When the processor executes the computer program, it implements the service function chain orchestration method for a low-earth orbit satellite network in any one of the foregoing embodiments.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium. The computer-readable storage medium stores computer instructions. When the computer instructions are executed by a processor, they implement the service function chain orchestration method for a low-earth orbit satellite network in any one of the foregoing embodiments.

[0016] The present invention provides a service function chain orchestration method for a low-earth orbit satellite network. The essence of this method is a distributed point-by-point service function chain orchestration scheme. It orchestrates the service function chain SFC request by sequentially embedding a series of virtual network functions VNFs into a series of adjacent satellite nodes. Therefore, the computing power requirement for the centralized controller is relatively low, and it is suitable for large-scale low-earth orbit satellite networks without a stable centralized controller. And during the distributed deployment of VNFs, first determine an optional satellite node set for deploying the next VNF, and then use the actor-critic network model of the integrated decision transformer to process the trajectory points of the C most recently deployed VNFs to select a second satellite node for deploying the next VNF from the optional satellite node set. That is to say, this method can determine a more robust VNF deployment decision based on global context information and reduce the local optimum problem that may be caused by relying only on the current state. Description of the Drawings

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the accompanying drawings required for the description of the specific embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0018] Figure 1 It is a schematic diagram of the service function chain SFC orchestration architecture;

[0019] Figure 2 It is a flowchart of a service function chain orchestration method applicable to a low-earth orbit satellite network provided by an embodiment of the present invention;

[0020] Figure 3 It is a schematic diagram of equally dividing the spherical area covered by the communication of the first satellite node provided by an embodiment of the present invention;

[0021] Figure 4 It is a framework diagram of the ACDT algorithm provided by an embodiment of the present invention;

[0022] Figure 5 It is a comparison chart of the total reward value curves of four different algorithms under 13 low-earth orbit satellites;

[0023] Figure 6 It is a comparison chart of the total reward value curves of four different algorithms under 15 low-earth orbit satellites;

[0024] Figure 7 It is a comparison chart of the total reward value curves of four different algorithms under 17 low-earth orbit satellites;

[0025] Figure 8 It is a comparison chart of the SFC orchestration success rates of four different algorithms under 13 low-earth orbit satellites;

[0026] Figure 9 It is a comparison chart of the SFC orchestration success rates of four different algorithms under 15 low-earth orbit satellites;

[0027] Figure 10 It is a comparison chart of the SFC orchestration success rates of four different algorithms under 17 low-earth orbit satellites;

[0028] Figure 11 It is a functional module diagram of a service function chain orchestration device applicable to a low-earth orbit satellite network provided by an embodiment of the present invention;

[0029] Figure 12 It is a schematic diagram of an electronic device provided by an embodiment of the present invention. Specific Embodiments

[0030] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions 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. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Components of the embodiments of the present invention generally described and illustrated in the drawings herein can be arranged and designed in a variety of different configurations.

[0031] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0032] The following will describe in detail some embodiments of the present invention with reference to the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0033] Figure 1 is a schematic diagram of the service function chain SFC orchestration architecture. As Figure 1 shown, the SFC orchestration framework can be divided into four layers: the application layer, the service layer, the virtualization layer, and the physical layer. The application layer publishes different applications according to different scenarios, such as environmental monitoring, disaster relief, intelligent scheduling, etc. The service layer converts these applications into SFC requests, which contain various non-embedded virtual network function requests VNFs orchestrated in a specified order, and each SFC request has a service source point and a service end point. The virtualization layer is responsible for matching network resources with SFC resource requests. The low-earth orbit satellite network is abstracted as a resource pool, and virtual resources can be dynamically allocated in the resource pool to meet the resource requirements of the SFC. Finally, the physical layer corresponds to the low-earth orbit satellite network environment composed of multiple low-earth orbit satellite nodes and base stations. As edge nodes, the base stations drive the SFC request orchestration process. Some base stations, as source nodes, publish SFC requests to their neighboring low-earth orbit satellite nodes, while other base stations, as target nodes, receive SFC orchestration information from the last virtual network function embedding node. These four layers are interconnected to form a closed-loop optimization.

[0034] To meet the service function chain orchestration requirements of a large-scale low-earth orbit satellite network without a stable centralized controller, the embodiments of the present invention propose a distributed point-by-point service function chain orchestration scheme, which orchestrates service function chain requests by sequentially embedding (i.e., deploying) a series of virtual network functions VNFs onto a series of adjacent nodes. The method provided by the embodiments of the present invention will be introduced in detail below.

[0035] Embodiment 1

[0036] Figure 2The flowchart of a service function chain orchestration method applicable to a low-Earth orbit satellite network provided by an embodiment of the present invention is as follows Figure 2 As shown, the method specifically includes the following steps:

[0037] Step S102, receive a service function chain (SFC) request and reset the low-Earth orbit satellite network environment.

[0038] Among them, the SFC request includes: a source base station, a destination base station, and multiple virtual network functions (VNFs) orchestrated in a specified order.

[0039] Specifically, after receiving a new SFC request, the low-Earth orbit satellite network environment will be reset. Resetting can also be understood as initialization, and parameters such as the computing resources, storage space, and energy consumption of each low-Earth orbit satellite will be re-assigned. In the embodiment of the present invention, the SFC request can be expressed as , where represents the source base station, represents the destination base station, represents multiple virtual network functions (VNFs) orchestrated in a specified order, , represents the k-th VNF request, and each VNF request has corresponding resource requirements, includes the memory requirements of all VNF requests, includes the CPU requirements of all VNF requests, includes the bandwidth requirements of all VNF requests, includes the power requirements of all VNF requests, represents the maximum orchestration time, that is, a preset duration threshold. If the total orchestration time of the SFC exceeds , it is determined that the SFC request fails to be orchestrated due to orchestration timeout.

[0040] In the embodiment of the present invention, the SFC request orchestration process is modeled using a distributed point-by-point scheme. Starting from the source base station, iteratively search for adjacent low-Earth orbit satellite nodes to embed each VNF until the SFC request is embedded, and collect the orchestration information of the destination base station.

[0041] Step S104, calculate the unit vector of the orchestration direction of the SFC request according to the position coordinates of the source base station and the destination base station.

[0042] Considering that in a low-Earth orbit satellite network, satellite nodes often have high mobility. Therefore, the embodiment of the present invention proposes to specify the orchestration direction of the SFC request to provide guidance for satellite area selection when deploying adjacent VNFs. After determining the source base station and the destination base station of the SFC request, the unit vector of the orchestration direction of the SFC request can be calculated according to the position coordinates of the source base station and the destination base station 。

[0043] where represents the distance between the source base station and the destination base station, represents the location coordinates of the source base station, represents the location coordinates of the destination base station.

[0044] Step S106: Randomly select a satellite node within the communication range of the source base station, and deploy the first VNF on the satellite node when it is determined that the satellite node meets the deployment conditions of the first VNF in the SFC request.

[0045] When deploying the VNF, first randomly select a low-earth orbit satellite node within the communication range of the source base station and determine whether it meets the deployment conditions of the first VNF in the SFC request. In the embodiments of the present invention, the deployment conditions of the VNF are specifically the resource requirements of the VNF. If the remaining resources of the satellite node can meet the requirements of the first VNF, the first VNF can be deployed on the satellite node; otherwise, randomly select another satellite node within the communication range of the source base station. If no satellite node that can meet the deployment conditions of the first VNF can be found after traversing all satellite nodes within the communication range of the source base station, it is determined that the current SFC request orchestration fails.

[0046] Step S108: Determine a set of optional satellite nodes for deploying the next VNF based on the orchestration direction unit vector, the location coordinates of the first satellite node where the current last deployed VNF is located, and the location coordinates of the neighbor satellite nodes of the first satellite node.

[0047] As can be seen from the steps above, the orchestration direction unit vector of the SFC request can be calculated based on the locations of the source base station and the destination base station, that is, the deployment directions of a series of VNFs in the SFC request are determined. Therefore, after determining the satellite node for deploying the first VNF, the satellite node for deploying the second VNF should not deviate from the above deployment direction in principle. In addition, it should be noted that in the embodiments of the present invention, only one VNF request can be deployed on one satellite node, and one VNF request can only be deployed on one satellite node.

[0048] At the end of step S106, only the first VNF is successfully deployed. Therefore, when step S108 is initially executed, the first satellite node that deploys the last VNF currently is the satellite node that deploys the first VNF. In order to determine the satellite node for deploying the next VNF, the embodiment of the present invention models the SFC orchestration problem as a decentralized partially observable Markov decision process and proposes a communication area partitioning method to limit its deployable area. Specifically, the communication coverage area of the first satellite node is partitioned, and then an area with the highest similarity to the orchestration direction unit vector is selected from them. If there are neighbor satellite nodes of the first satellite node in this area, the satellite node for deploying the next VNF should be selected within this area, that is, the neighbor satellite nodes (of the first satellite node) located within this area form the set of candidate satellite nodes.

[0049] Step S110: Use the actor-critic network model of the integrated decision transformer to process the most recently deployed C VNF trajectory points to select the second satellite node for deploying the next VNF from the set of candidate satellite nodes.

[0050] Among them, each VNF trajectory point includes: the satellite node number of the deployed VNF, the resource status of the satellite node, and the reward obtained by deploying this VNF. The embodiment of the present invention does not specifically limit the calculation method of the reward, and the user can set it according to actual needs. For example, the reward is inversely proportional to the communication delay and resource consumption.

[0051] After the set of candidate satellite nodes is determined, the embodiment of the present invention uses the most recently deployed C VNF trajectory points as the input of the actor-critic network model of the integrated decision transformer. The action space of the agent in the model is the above set of candidate satellite nodes, and the output of the actor-critic network model is the second satellite node for deploying the next VNF. The embodiment of the present invention does not specifically limit the value of C. The larger the value of C, the greater the computational amount of the model, and the higher the decision-making quality of the satellite node; the smaller the value of C, the smaller the computational amount of the model, and the lower the decision-making quality of the satellite node. Therefore, the user can set the value of C according to actual needs. Optionally, C is set to 4. In the initial stage of SFC deployment, when the number of VNF trajectory points is less than C, they are filled with zero values. After the number of trajectory points increases, they are gradually replaced with real trajectory points, but only the most recently deployed C trajectory points are retained.

[0052] In the traditional model, the state of the agent is its local observation state. If only the observed values of satellite nodes (such as resource status) are used as the agent's state, then it is impossible to capture the changes in the LEO satellite network in the time dimension, nor can the past decision and reward information be used to correct the current decision. Compared with the traditional model, in the embodiment of the present invention, in the actor-critic network model, the state of the agent is the C VNF trajectory points deployed most recently. The trajectory (that is, the C VNF trajectory points) contains historical information of multiple time steps, which can help the system capture the change trend of node resources and topological status in the dynamic network. For example, if the bandwidth utilization rate of a satellite node has gradually increased in the past few steps, it may indicate that the node is about to run out of resources, and the trajectory can help the agent predict this trend and avoid selecting it.

[0053] Moreover, the trajectory can capture the long-term impact of past decisions and optimize the global policy. For example, if a certain satellite node has been selected as an embedding node multiple times in the past and caused high latency, the system can learn to avoid selecting similar nodes, thereby supporting long-term dependence modeling. Through trajectory modeling, the agent can make more robust decisions based on global context information, reducing the local optimality problem that may be caused by relying only on the current state.

[0054] Step S112, when it is determined that the second satellite node meets the deployment condition of the next VNF, calculate the SFC orchestration duration.

[0055] It is known that each VNF has corresponding deployment conditions. If the second satellite node selected by the actor-critic network model meets the deployment condition of the next VNF, it is necessary to further consider whether the SFC orchestration duration up to now meets the requirements. In the embodiment of the present invention, the SFC orchestration duration is the sum of the time consumption of historical round orchestration and the time consumption of the current round orchestration.

[0056] Step S114, when it is determined that the SFC orchestration duration is less than or equal to the preset duration threshold and all VNFs in the SFC request have not been deployed yet, return to step S108 until all VNFs in the SFC request are deployed.

[0057] After the second VNF is deployed, the first satellite node of the currently last deployed VNF is updated from the satellite node that deployed the first VNF to the satellite node that deployed the second VNF. Next, taking the satellite node that deployed the second VNF as the base point, determine the set of optional satellite nodes for deploying the next VNF. And so on, repeat the above steps S108 to S114 until all VNFs in the SFC request are deployed. Moreover, only when the satellite node that deploys the last VNF in the SFC request can communicate with the destination base station can it be determined that the SFC request deployment is successful.

[0058] An embodiment of the present invention provides a service function chain orchestration method applicable to a low-earth orbit satellite network. The essence of this method is a distributed point-by-point service function chain orchestration scheme, which orchestrates service function chain (SFC) requests by sequentially embedding a series of virtual network functions (VNFs) into a series of adjacent satellite nodes. Therefore, the computing power requirement for the central controller is relatively low, making it suitable for large-scale low-earth orbit satellite networks without a stable central controller. And during the distributed deployment of VNFs, first, a set of candidate satellite nodes for deploying the next VNF is determined, and then an actor-critic network model integrating decision transformers is used to process the trajectories of the C most recently deployed VNFs to select a second satellite node for deploying the next VNF from the set of candidate satellite nodes. That is, this method can determine a more robust VNF deployment decision based on global context information, reducing the local optimum problem that may be caused by relying only on the current state.

[0059] In an alternative embodiment, step S108, determining a set of candidate satellite nodes for deploying the next VNF based on the orchestration direction unit vector, the position coordinates of the first satellite node where the currently last deployed VNF is located, and the position coordinates of the neighbor satellite nodes of the first satellite node, specifically includes the following steps:

[0060] Step S1081, based on the equal division principle and the position coordinates of the first satellite node, divide the spherical area covered by the communication of the first satellite node into four quadrants to obtain four communication areas, and determine the center point coordinates of each communication area.

[0061] Step S1082, calculate the unit vector of each communication area based on the center point coordinates of each communication area and the position coordinates of the first satellite node.

[0062] Step S1083, calculate the dot product of the unit vector of each communication area and the orchestration direction unit vector, and sort the four communication areas in descending order based on the dot product result.

[0063] If there are neighbor satellite nodes in the target communication area ranked first based on the position coordinates of the neighbor satellite nodes of the first satellite node, execute the following step S1084; otherwise, execute the following step S1085.

[0064] Step S1084, construct a set of candidate satellite nodes based on all the neighbor satellite nodes in the target communication area.

[0065] Step S1085, sequentially select the communication areas ranked lower until a set of candidate satellite nodes is determined.

[0066] As described above, in the embodiments of the present invention, a communication area division method is used to simply calculate the service process direction of the SFC request, thereby restricting the deployable area of the next VNF to be deployed. Specifically, as Figure 3 shown, first, the spherical area (3D space) covered by the communication of the first satellite node is equally divided into four communication areas . The size of the spherical area covered by the communication of each satellite node is a known quantity. According to the position coordinates of the satellite node, the central point coordinates of each communication area can be determined by a mathematical method.

[0067] Next, according to the position coordinates of the first satellite node and the central point coordinates of each communication area, referring to the above calculation method of the arrangement direction unit vector, the unit vector of each communication area can be calculated respectively , that is, the position vector difference between the first satellite node and the central point of each area. Then, the deployable area sorting result is generated through vector dot product and descending order sorting. The target communication area ranked first is denoted as .

[0068] If has neighbor satellite nodes, then all neighbor satellite nodes in constitute the optional satellite node set; if has no neighbor satellite nodes, then the area ranked second is sequentially selected as the new target communication area . If the new has neighbor satellite nodes, then all neighbor satellite nodes in the new constitute the optional satellite node set; otherwise, the area ranked third is sequentially selected, and so on, until the optional satellite node set is determined.

[0069] In an optional embodiment, after selecting the second satellite node for deploying the next VNF in the optional satellite node set, the method of the present invention further includes the following steps:

[0070] Step S113, in the case where it is determined that the second satellite node does not meet the deployment conditions of the next VNF, release all the deployed VNFs in the low-earth orbit satellite network, return to step S106, and update the SFC orchestration duration.

[0071] Specifically, in the embodiments of the present invention, the SFC request can be rearranged in multiple rounds. If the second satellite node selected by the actor-critic network model does not meet the deployment conditions of the next VNF, the SFC orchestration of the current round fails, the current round ends, and a new round begins. At the start of the new round, all the embedded VNFs on each LEO satellite node in the previous round need to be released, and then the above step 106 is returned, that is, a satellite node for deploying the first VNF is reselected within the communication range of the source base station, and the subsequent method process is executed. Moreover, the time consumed by the service process that fails to be orchestrated in the current round needs to be accumulated into the SFC orchestration duration.

[0072] In an alternative embodiment, in the above step S112, calculating the SFC orchestration duration specifically includes the following content:

[0073] Using the formula to calculate the SFC orchestration duration; where represents the SFC orchestration duration of the SFC request , represents the time consumption of the current round of orchestration, represents the time consumption of the historical round of orchestration.

[0074] , L represents the set of physical network links, represents the set of virtual network links, represents the one-hop transmission delay from satellite node u to satellite node v at time slot t, represents the VNF virtual link deployed on the physical network link uv, represents the VNF virtual link not deployed on the physical network link uv, represents the processing delay of satellite node i, represents that the kth VNF is deployed at satellite node i, represents that the kth VNF is not deployed at satellite node i, represents the total number of satellites in the LEO satellite network, represents the set of all VNFs in the SFC request.

[0075] Based on the above formula, for the current round, the time consumption mainly includes transmission delay and processing delay caused by the multi-hop orchestration process. For any link embedding any adjacent VNF requests, the transmission delay can be expressed as , where is the size of the data packet transmitted on the physical network link uv, represents the bandwidth of the physical network link uv.

[0076] In an embodiment of the present invention, after calculating the SFC scheduling duration, the method of the present invention further includes the following steps: Step S115, when it is determined that the SFC scheduling duration is greater than a preset duration threshold, it is determined that the SFC request scheduling fails. That is to say, if , then it is determined that the SFC request scheduling fails.

[0077] To generate a stable SFC scheduling strategy in a low-earth orbit satellite network, an embodiment of the present invention proposes an ACDT (Actor-Critical Decision Transformer) algorithm, which is a distributed multi-agent reinforcement learning algorithm that integrates a decision transformer network into an actor-critic network, thereby obtaining an actor-critic network model integrated with a decision transformer. Figure 4 It is a framework diagram of the ACDT algorithm provided by an embodiment of the present invention.

[0078] The decision transformer is a popular neural network model that can solve sequential decision problems through context trajectory modeling. An embodiment of the present invention proposes a simplified decision transformer model for stable SFC scheduling in a low-earth orbit satellite network. It consists of a position encoding model and an attention-based model. The position encoding model encodes the trajectory sequences at different positions using a position encoding layer and an embedding layer. It consists of two linear transformations connected by a ReLU activation function: , where and represent weights, and represent biases.

[0079] The attention-based model can capture the context time-related features of the input trajectory. It maps the trajectory to a vector combination . For each trajectory point in the trajectory , the array point attention model converts it into a query vector , a key vector and a value vector through three different linear transformations: .

[0080] , , and d represent the dimensions of the query vector , the key vector , the value vector and the trajectory point respectively. Then, the array point attention model combines all the token-based vectors along the time dimension to generate a complete vector combination : ; wherein, represents the dimension of the trajectory . On this basis, the attention is calculated through a masking mechanism: ; wherein, represents preventing the softmax function from falling into the minimum gradient solution, and M is a mask matrix.

[0081] When the above neural network model is applied to the SFC orchestration in the low-earth orbit satellite network, for the SFC distributed point-by-point orchestration process, the decision transformer network extracts the C VNF trajectory points deployed most recently according to the chronological order. Each VNF trajectory point includes: the satellite node number where the VNF is deployed, the resource status of the satellite node, and the reward obtained by deploying the VNF, thereby forming a context trajectory , t represents the current time slot. In the embodiment of the present invention, , t - u + 1 = C, that is, C VNF trajectory points are used, and the overall context length of the trajectory is C - 1. Among them, , represents that the VNF is deployed on satellite node i at time slot u, represents the resource status of satellite node i at time slot u, represents the reward obtained by deploying the VNF on satellite node i at time slot u. At the initial stage of the model application, the initial trajectory needs to be filled with zero values , and the trajectory is gradually filled with real trajectory points as the VNF is deployed.

[0082] In the actor-critic network, each agent shares the same Actor-Critic network to train together distributively. At the same time, the trajectory contains the temporary local network state features from the sequentially embedded satellite nodes to promote the algorithm to converge to the global optimum, that is, the algorithm can form a stable SFC orchestration strategy. In the actor-critic network, the actor network adaptively generates an action strategy according to the trajectory of the C VNF trajectory points deployed most recently, while the critic evaluates through the output value and guides the action to converge to the global optimum. The action output by the actor network, wherein, represents the network parameters of the actor network, represents the probability distribution on the action space , and the action space is also all the neighboring satellite nodes in Z(i), that is, the set of optional satellite nodes.

[0083] In the embodiment of the present invention, after selecting the second satellite node for deploying the next VNF from the set of optional satellite nodes, the following steps are further included:

[0084] Step S111, based on the formula update the network parameters of the actor network in the actor-critic network model; based on the formula update the network parameters of the critic network.

[0085] Among them, represents the network parameters of the actor network, represents the resource status of satellite node i at the current time slot t, represents that the VNF is deployed on satellite node i at the current time slot t, represents the old policy of the actor network , , represents the important weight of the relative difference between the old and new policies of the actor network, represents the clipping hyperparameter, represents the new policy of the actor network, , represents the advantage function calculated by the TD error, represents the discount factor, represents the reward obtained when the VNF is deployed on satellite node m at time slot x, and t - u + 1 = C, represents the value function of the critic network in the actor-critic network model, represents the network parameters of the critic network. The critic network updates its value function and returns the value to the actor network, guiding the update process of the actor network towards the global optimum.

[0086] In the actor-critic network, the policy refers to the rule or distribution by which the agent selects actions at each time step. The old policy is the existing policy of the current network, which represents the reference rule when the previous network generated actions and is used to measure the improvement of the new policy to ensure that the network update does not deviate from the optimization goal. The new policy is the optimized policy, which represents a better rule generated based on the current context trajectory (state) and reward signal. The new policy can better capture the context information and improve the effect of action selection.

[0087] In summary, the embodiment of the present invention proposes a distributed point-by-point service function chain orchestration scheme, which orchestrates service function chain requests by sequentially embedding a series of virtual network functions into a series of adjacent satellite nodes to reduce the computing power requirements for the centralized controller; and, during the VNF deployment process, the embodiment of the present invention specifically uses the actor-critic network model integrating decision transformers to generate service function chain orchestration strategies and globally optimize them. Because the decision transformer can extract historical local network state features into the context trajectory, the model can use this trajectory to adaptively generate more robust service function chain orchestration strategies, thus effectively avoiding the problem of local optimum.

[0088] To evaluate the performance of the method provided by the embodiments of the present invention. Figures 5 to 7 The total reward value curves of different algorithms applied to different numbers of low-earth orbit satellites are respectively shown. By Figures 5 to 7 It can be seen that as the training progresses, two artificial intelligence-based SFC orchestration algorithms (deep Q-learning algorithm and ACDT algorithm) can obtain higher rewards. In contrast, the rewards of other algorithms fluctuate within a certain range. This indicates that reinforcement learning can improve the SFC orchestration performance by interacting with the low-earth orbit satellite networking environment. Moreover, the results show that the optimal solution algorithm ranks first. This is because it greedily selects a series of optimal nodes to sequentially embed VNFs. The ACDT-based SFC orchestration algorithm proposed by the embodiments of the present invention is better than the other two baseline algorithms and can even approach the theoretical optimal solution. The first-fit algorithm performs the worst due to its randomness. It is also noted that as the number of satellites increases, the average reward values obtained by the four algorithms also increase. This is because more satellites mean that the network can provide more resources to support SFC orchestration. Moreover, the embedding of VNF requests is more dispersed, thus avoiding the single-point bottleneck problem. In addition, for the ACDT-based algorithm proposed by the embodiments of the present invention, its convergence time is also shortened. In a 13-low-earth orbit satellite network, it converges after about 4000 episodes. In a 15-low-earth orbit satellite network, the time is reduced to about 2000 episodes. While in a 17-low-earth orbit satellite network, only about 1200 episodes are needed. The deep Q-learning algorithm always converges after about 3000 episodes. This shows that benefiting from the context correlation ability, the algorithm proposed by the embodiments of the present invention can make full use of the local information in the historical context trajectory to accelerate the convergence speed.

[0089] Figures 8 to 10 The SFC orchestration success rates of different algorithms applied to different numbers of low-earth orbit satellites are respectively shown. By Figures 8 to 10 It can be seen that similar to the convergence performance of the total reward value, the orchestration success rate also increases with the increase of the satellite scale. It can also be found that the gap between the optimal solution algorithm and the ACDT-based algorithm proposed by the present invention gradually narrows. Specifically, in a 17-low-earth orbit satellite network, the difference in success rate performance between the optimal solution algorithm and the ACDT-based algorithm proposed by the present invention is only about 10%. This further proves that the algorithm proposed by the decision transformer structure can capture the temporal correlation of resources and local topology during the orchestration process and use them to adaptively select stable nodes with sufficient resources to embed VNF requests.

[0090] Embodiment 2

[0091] An embodiment of the present invention also provides a service function chain orchestration device applicable to a low-earth orbit satellite network. This device is mainly used to execute the service function chain orchestration method applicable to a low-earth orbit satellite network provided in the first embodiment above. The following is a specific introduction to the service function chain orchestration device applicable to a low-earth orbit satellite network provided in the embodiment of the present invention.

[0092] Figure 11 It is a functional module diagram of a service function chain orchestration device applicable to a low-earth orbit satellite network provided in an embodiment of the present invention. As Figure 11 shown, the device mainly includes: a receiving and resetting module 11, a first calculation module 12, a deployment module 13, a determination module 14, a selection module 15, a second calculation module 16, and a return module 17, where:

[0093] The receiving and resetting module 11 is used to receive a service function chain (SFC) request and reset the low-earth orbit satellite network environment; wherein, the SFC request includes: a source base station, a destination base station, and a plurality of virtual network functions (VNFs) arranged in a specified order.

[0094] The first calculation module 12 is used to calculate the orchestration direction unit vector of the SFC request according to the position coordinates of the source base station and the position coordinates of the destination base station.

[0095] The deployment module 13 is used to randomly select a satellite node within the communication range of the source base station, and deploy the first VNF on the satellite node when it is determined that the satellite node meets the deployment conditions of the first VNF in the SFC request.

[0096] The determination module 14 is used to determine a set of optional satellite nodes for deploying the next VNF based on the orchestration direction unit vector, the position coordinates of the first satellite node of the currently last deployed VNF, and the position coordinates of the neighbor satellite nodes of the first satellite node.

[0097] The selection module 15 is used to process the trajectory points of the recently deployed C VNFs by using the actor-critic network model of the integrated decision transformer to select a second satellite node for deploying the next VNF in the set of optional satellite nodes; wherein, each VNF trajectory point includes: the satellite node number where the VNF is deployed, the resource status of the satellite node, and the reward obtained by deploying this VNF.

[0098] The second calculation module 16 is used to calculate the SFC orchestration duration when it is determined that the second satellite node meets the deployment conditions of the next VNF.

[0099] The return module 17 is used to return and call the determination module until all VNFs in the SFC request are deployed when it is determined that the SFC orchestration duration is less than or equal to a preset duration threshold and all VNFs in the SFC request have not been deployed.

[0100] An embodiment of the present invention provides a service function chain orchestration device applicable to a low-earth orbit satellite network. This device essentially implements a distributed point-by-point service function chain orchestration scheme. It orchestrates service function chain (SFC) requests by sequentially embedding a series of virtual network functions (VNFs) into a series of adjacent satellite nodes. Therefore, the computing power requirement for the central controller is relatively low, making it suitable for large-scale low-earth orbit satellite networks without a stable central controller. And during the distributed deployment of VNFs, first, a set of candidate satellite nodes for deploying the next VNF is determined. Then, an actor-critic network model integrating a decision transformer is used to process the trajectories of the C most recently deployed VNFs to select a second satellite node for deploying the next VNF from the set of candidate satellite nodes. That is, this device can determine a more robust VNF deployment decision based on global context information, reducing the local optimality problem that may arise from relying solely on the current state.

[0101] Optionally, the determination module 14 is specifically configured to:

[0102] Based on the equal division principle and the position coordinates of the first satellite node, divide the spherical region covered by the communication of the first satellite node into four quadrants to obtain four communication regions, and determine the center point coordinates of each communication region.

[0103] Based on the center point coordinates of each communication region and the position coordinates of the first satellite node, calculate the unit vector of each communication region.

[0104] Calculate the dot product of the unit vector of each communication region and the unit vector of the orchestration direction, and sort the four communication regions in descending order based on the dot product result.

[0105] If there are neighbor satellite nodes in the target communication region ranked first in the sorting based on the position coordinates of the neighbor satellite nodes of the first satellite node, construct a set of candidate satellite nodes based on all the neighbor satellite nodes in the target communication region.

[0106] Otherwise, sequentially select the communication regions ranked lower until a set of candidate satellite nodes is determined.

[0107] Optionally, the second calculation module 16 is specifically configured to:

[0108] Use the formula to calculate the SFC orchestration duration; where represents the SFC orchestration duration of the SFC request , represents the time consumption of the current round of orchestration, represents the time consumption of the historical round of orchestration; , L represents the set of physical network links, Represents a set of virtual network links, Represents the one-hop transmission delay from satellite node u to satellite node v at time slot t, Represents the VNF virtual link Deployed on the physical network link uv, Represents the VNF virtual link Not deployed on the physical network link uv, Represents the processing delay of satellite node i, Represents that the k-th VNF is deployed at satellite node i, Represents that the k-th VNF is not deployed at satellite node i, Represents the total number of satellites in the low-earth orbit satellite network, Represents the set of all VNFs in the SFC request.

[0109] Optionally, after selecting the second satellite node for deploying the next VNF from the set of optional satellite nodes, the device is further configured to:

[0110] Based on the formula Update the network parameters of the actor network in the actor-critic network model; based on the formula Update the network parameters of the critic network; where, Represents the network parameters of the actor network, Represents the resource status of satellite node i at the current time slot t, Represents that the VNF is deployed at satellite node i at the current time slot t, Represents the old policy of the actor network , , Represents the important weight of the relative difference between the old and new policies of the actor network, Represents the clipping hyperparameter, Represents the new policy of the actor network, , Represents the advantage function calculated by the TD error, Represents the discount factor, Represents the reward obtained when the VNF is deployed at satellite node m at time slot x, and t - u + 1 = C, Represents the value function of the critic network in the actor-critic network model, Represents the network parameters of the critic network.

[0111] Optionally, after selecting the second satellite node for deploying the next VNF from the set of optional satellite nodes, the device is further configured to:

[0112] In case it is determined that the second satellite node does not meet the deployment conditions for the next VNF, release all the deployed VNFs in the LEO satellite network, return to step S106, and update the SFC orchestration duration.

[0113] Optionally, after calculating the SFC orchestration duration, the apparatus is further configured to:

[0114] In case it is determined that the SFC orchestration duration is greater than a preset duration threshold, determine that the SFC request orchestration fails.

[0115] Embodiment III

[0116] Refer to Figure 12 , an embodiment of the present invention provides an electronic device, which includes: a processor 60, a memory 61, a bus 62, and a communication interface 63. The processor 60, the communication interface 63, and the memory 61 are connected through the bus 62. The processor 60 is configured to execute an executable module stored in the memory 61, such as a computer program.

[0117] Among them, the memory 61 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 63 (which may be wired or wireless), a communication connection is established between the system network element and at least one other network element, and the Internet, wide area network, local area network, metropolitan area network, etc. can be used.

[0118] The bus 62 may be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, Figure 12 only a bidirectional arrow is used in

[0119] Among them, the memory 61 is used to store a program. After receiving an execution instruction, the processor 60 executes the program. The method executed by the apparatus defined by any of the foregoing embodiments of the present invention can be applied to the processor 60 or implemented by the processor 60.

[0120] The processor 60 may be an integrated circuit chip with the ability to process signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 60 or the instructions in the form of software. The above-mentioned processor 60 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present invention can be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 61, and the processor 60 reads the information in the memory 61 and combines its hardware to complete the steps of the above method.

[0121] A computer program product of a service function chain orchestration method and device applicable to a low-earth orbit satellite network provided by an embodiment of the present invention includes a computer-readable storage medium storing non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the method in the foregoing method embodiments. For specific implementation, reference can be made to the method embodiments and will not be elaborated here.

[0122] In addition, in each embodiment of the present invention, each functional unit may be integrated in a processing unit, may also exist physically separately for each unit, or two or more units may be integrated in one unit.

[0123] If a function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium executable by a processor. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0124] It should be noted that: similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0125] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of this invention is usually placed during use. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation to the present invention. In addition, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0126] In addition, the terms "horizontal", "vertical", "hanging", etc. do not mean that the components are required to be absolutely horizontal or hanging, but can be slightly inclined. For example, "horizontal" only means that its direction is more horizontal relative to "vertical", and does not mean that this structure must be completely horizontal, but can be slightly inclined.

[0127] In the description of the present invention, it should also be noted that unless otherwise clearly specified and limited, the terms "set", "install", "connect", "couple" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A service function chain arrangement method applicable to a low-orbit satellite network, characterized in that: include: Step S102, receiving a service function chain SFC request and resetting the low-orbit satellite network environment; wherein the SFC request includes: a source base station, a destination base station, and a plurality of virtual network functions VNFs arranged in a specified order; Step S104, calculating the arrangement direction unit vector of the SFC request according to the location coordinates of the source base station and the location coordinates of the destination base station; Step S106, randomly selecting a satellite node within the communication range of the source base station, and deploying the first VNF ​​on the satellite node if it is determined that the satellite node meets the deployment condition of the first VNF ​​in the SFC request; Step S108, determining a set of optional satellite nodes for deploying the next VNF ​​based on the orchestration direction unit vector, the position coordinates of the first satellite node that currently deploys the last VNF, and the position coordinates of the neighboring satellite nodes of the first satellite node; Step S110, using the actor-critic network model with integrated decision transformer to process the most recently deployed C VNF trajectory points, so as to select a second satellite node for deploying the next VNF ​​from the optional satellite node set; wherein each of the VNF trajectory points includes: the satellite node number of the deployed VNF, the resource status of the satellite node, and the reward obtained by deploying the VNF; Step S112, when it is determined that the second satellite node meets the deployment condition of the next VNF, calculate the SFC orchestration duration; Step S114: if it is determined that the SFC orchestration duration is less than or equal to the preset duration threshold and the deployment of all VNFs in the SFC request has not been completed, return to step S108 until the deployment of all VNFs in the SFC request is completed; Wherein, based on the orchestration direction unit vector, the position coordinates of the first satellite node that currently deploys the last VNF, and the position coordinates of the neighboring satellite nodes of the first satellite node, determining a set of optional satellite nodes for deploying the next VNF ​​includes: Based on the equal division principle and the position coordinates of the first satellite node, the spherical area covered by the communication of the first satellite node is divided into four quadrants to obtain four communication areas, and the coordinates of the center point of each of the communication areas are determined; Calculate the unit vector of each communication area based on the center point coordinates of each communication area and the position coordinates of the first satellite node; Calculating the dot product of the unit vector of each communication area and the arrangement direction unit vector, and sorting the four communication areas in descending order based on the dot product result; In a case where it is determined based on the position coordinates of the neighbor satellite nodes of the first satellite node that there is a neighbor satellite node in the first-ranked target communication area, constructing the selectable satellite node set based on all the neighbor satellite nodes in the target communication area; Otherwise, the communication areas ranked later are selected in order until the selectable satellite node set is determined.

2. The service function chain arrangement method applicable to a low-orbit satellite network according to claim 1, characterized in that: Calculate the SFC scheduling time, including: Using the formula Calculate the SFC scheduling duration; wherein, Indicates SFC request The SFC arrangement time, Indicates the time consumption of the current round of scheduling. Indicates the time consumption of historical round scheduling; , L represents the set of physical network links, Represents a collection of virtual network links. represents the single-hop transmission delay from satellite node u to satellite node v in time slot t, Indicates VNF virtual link Deployed on the physical network link uv, Indicates VNF virtual link Not deployed on the physical network link uv, represents the processing delay of satellite node i, indicates that the kth VNF is deployed at satellite node i, indicates that the kth VNF is not deployed at satellite node i, represents the total number of satellites in the low-orbit satellite network, Represents the set of all VNFs in the SFC request.

3. The service function chain arrangement method applicable to a low-orbit satellite network according to claim 1, characterized in that: After selecting a second satellite node for deploying the next VNF ​​from the set of optional satellite nodes, the method further includes: Step S111, based on the formula Update the network parameters of the actor network in the actor-critic network model; based on the formula Updating network parameters of the critic network; in, represents the network parameters of the actor network, represents the resource status of satellite node i in the current time slot t, Indicates that the VNF is deployed on satellite node i at the current time slot t, Represents the old policy of the actor network , , The importance weights representing the relative differences between the new and old strategies of the actor network, represents the clipping hyperparameter, A new strategy for representing actor networks, , represents the advantage function calculated by TD error, represents the discount factor, represents the reward obtained by deploying VNF on satellite node m in time slot x, and , represents the value function of the critic network in the actor-critic network model, represents the network parameters of the critic network.

4. The service function chain arrangement method applicable to a low-orbit satellite network according to claim 1, characterized in that: After selecting a second satellite node for deploying the next VNF ​​from the set of optional satellite nodes, the method further includes: Step S113, when it is determined that the second satellite node does not meet the deployment conditions of the next VNF, release all deployed VNFs in the low-orbit satellite network, return to step S106, and update the SFC orchestration duration.

5. The service function chain arrangement method applicable to a low-orbit satellite network according to claim 1, characterized in that: After calculating the SFC scheduling duration, it also includes: Step S115: When it is determined that the SFC scheduling duration is greater than the preset duration threshold, it is determined that the SFC scheduling request fails.

6. A service function chain arrangement device suitable for a low-orbit satellite network, characterized in that: include: A receiving and resetting module, configured to receive a service function chain (SFC) request and reset a low-orbit satellite network environment; wherein the SFC request includes: a source base station, a destination base station, and a plurality of virtual network functions (VNFs) arranged in a specified order; A first calculation module, configured to calculate an arrangement direction unit vector of the SFC request according to the location coordinates of the source base station and the location coordinates of the destination base station; A deployment module, configured to randomly select a satellite node within the communication range of the source base station, and deploy the first VNF ​​on the satellite node if it is determined that the satellite node meets the deployment condition of the first VNF ​​in the SFC request; A determination module, configured to determine a set of optional satellite nodes for deploying a next VNF ​​based on the orchestration direction unit vector, the position coordinates of the first satellite node that currently deploys the last VNF, and the position coordinates of the neighboring satellite nodes of the first satellite node; A selection module is used to process the C most recently deployed VNF trajectory points using an actor-critic network model with an integrated decision transformer to select a second satellite node for deploying the next VNF ​​from the set of optional satellite nodes; wherein each of the VNF trajectory points includes: a satellite node number for deploying the VNF, a resource status of the satellite node, and a reward obtained by deploying the VNF; A second calculation module is used to calculate the SFC orchestration duration when it is determined that the second satellite node meets the deployment condition of the next VNF; A return module, configured to return to the calling determination module when it is determined that the SFC orchestration duration is less than or equal to a preset duration threshold and the deployment of all VNFs in the SFC request has not been completed, until the deployment of all VNFs in the SFC request is completed; Wherein, the determination module is specifically used for: Based on the equal division principle and the position coordinates of the first satellite node, the spherical area covered by the communication of the first satellite node is divided into four quadrants to obtain four communication areas, and the coordinates of the center point of each of the communication areas are determined; Calculate the unit vector of each communication area based on the center point coordinates of each communication area and the position coordinates of the first satellite node; Calculating the dot product of the unit vector of each communication area and the arrangement direction unit vector, and sorting the four communication areas in descending order based on the dot product result; In a case where it is determined based on the position coordinates of the neighbor satellite nodes of the first satellite node that there is a neighbor satellite node in the first-ranked target communication area, constructing the selectable satellite node set based on all the neighbor satellite nodes in the target communication area; Otherwise, the communication areas ranked later are selected in order until the selectable satellite node set is determined.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the computer program, the service function chain orchestration method applicable to a low-orbit satellite network described in any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by the processor, the service function chain orchestration method applicable to a low-orbit satellite network as described in any one of claims 1 to 5 is implemented.

Images