Mobile self-organizing network service function chain arrangement method, device, system and storage medium

Abstract

The present invention relates to the field of wireless communications technology and discloses a method, device, system, and storage medium for orchestrating a service function chain in a mobile ad hoc network. The method comprises: obtaining an orchestration request and a starting node; wherein the orchestration request includes multiple entities; inserting the starting node into the header of the orchestration request to obtain a pre-processed orchestration request; and dividing the pre-processed orchestration request into multiple node-entity anchor pairs; capturing the topological dynamics in the mobile ad hoc network based on a time-constrained recursive topology discovery protocol; constructing a mesh forest graph using the topological dynamics; and, based on the mesh forest graph, determining the optimal segment corresponding to each node-entity anchor pair with the goal of minimizing the sum of communication cost and performance cost, and connecting the optimal segments to obtain an optimal service function chain. The present invention can be used for orchestrating the functions of functionally heterogeneous mobile ad hoc nodes in environments without infrastructure (e.g., outdoor areas without base stations).

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a method, device, system and storage medium for orchestrating a mobile self-organizing network service function chain. Background Art

[0002] Mobile Ad Hoc Networks (MANETs) are often used for rescue, environmental monitoring, and military operations in environments with poor communication infrastructure (such as outdoor areas without base stations). On-demand, on-site orchestration of mobile ad hoc nodes (such as drones and smart cars) plays a crucial role in executing these complex tasks. These tasks typically require multiple functions (such as camera capture, image recognition, data transmission, and target tracking) to be orchestrated into a unified service function chain (SFC), with limited human intervention and support from communication infrastructure. These components are invoked sequentially as needed to collaboratively complete the task.

[0003] The orchestration of the service function chain of a mobile self-organizing network includes several key aspects: 1. Functional aspect: The functions deployed on the nodes must be optimized and executed in the required order to meet the functional requirements of the task. 2. Communication aspect: Qualified nodes must be correctly selected to form the best communication path to provide good service quality (QoS) guarantees for data and command interaction. Therefore, the orchestration of the service function chain is more complex than pure routing (only involving communication aspects) that only includes nodes. At the same time, due to the high dynamics of mobile self-organizing networks in terms of topology, signal strength, node failures, and battery capacity, mobile orchestration in environments without infrastructure (such as outdoor areas without base stations) is more challenging, and there are still many issues that need to be further studied:

[0004] 1) How to quickly discover sufficient topology, function, communication, and performance information of mobile self-organizing nodes in an infrastructure-free environment and manage it in a unified manner as the data foundation for the orchestration of mobile service function chains.

[0005] 2) How to automatically orchestrate a high-quality mobile service function chain consisting of mobile self-organizing nodes based on topology discovery information without (or with little) human intervention or base station assistance, providing a series of ordered functions required for on-site tasks.

[0006] In the existing technology, the orchestration schemes of the service function chain involved include:

[0007] 1. Orchestration in site-based wireless networks

[0008] The paper "ZHANG P, WANG C, KUMAR N, et al. Space-Air-Ground Integrated Multi-Domain Network Resource Orchestration Based on Virtual Network Architecture: A DRL Method [J / OL]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(3): 2798-2808. DOI: 10.1109 / TITS.2021.3099477." focuses on network resource allocation in space-air-ground integrated networks, modeling it as a virtual network mapping (VNE) problem. A deep reinforcement learning (DRL) scheduling algorithm that considers node heterogeneity, temporal dynamics, and self-organization characteristics is proposed, and a five-layer DRL agent is constructed. The optimization goal is to maximize the service provider's profit while serving as many virtual network requests (VNRs) as possible. This paper focuses primarily on communication and does not delve into the orchestration of heterogeneous functions.

[0009] The paper "MA L, CHENG N, WANG X, et al. On-Demand Resource Management for 6GWireless Networks Using Knowledge-Assisted Dynamic Neural Networks[C / OL] / / ICC2022 - IEEE International Conference on Communications. 2022: 1-6. DOI:10.1109 / ICC45855.2022.9882279." The paper "KHAN AA, ABOLHASAN M, NI W, et al. An End-to-End (E2E) Network Slicing Framework for 5G Vehicular Ad-Hoc Networks[J / OL]. IEEE Transactions on Vehicular Technology, 2021, 70(7): 7103-7112. DOI:10.1109 / TVT.2021.3084735." proposes a model based on dynamic neural network to allocate user requested bandwidth resources in heterogeneous wireless networks. This model builds a knowledge base that comprehensively considers service characteristics and node computing capabilities, dynamically adjusting the depth of the neural network model to balance bandwidth allocation accuracy and overall processing latency, thereby achieving optimal solution and transmission overhead. However, this model does not consider the functional orchestration of heterogeneous nodes.

[0010] The paper "KHAN AA, ABOLHASAN M, NI W, et al. An End-to-End (E2E) Network Slicing Framework for 5G Vehicular Ad-Hoc Networks[J / OL]. IEEE Transactions on Vehicular Technology, 2021, 70(7): 7103-7112. DOI:10.1109 / TVT.2021.3084735." proposes an end-to-end 5G vehicular network slicing framework driven by SDN / NFV. The orchestration of heterogeneous resources is modeled as a multi-objective optimization problem, and a genetic algorithm is used to solve the service function chain to construct an end-to-end network slice. However, the mobility of nodes, the heterogeneity of functions, the self-organizing characteristics, and the interaction between mobile nodes and network infrastructure have not been effectively reflected in this framework.

[0011] The paper "THIRUVASAGAM PK, CHAKRABORTY A, MATHEW A, et al. Reliable Placement of Service Function Chains and Virtual Monitoring Functions WithMinimal Cost in Softwarized 5G Networks[J / OL]. IEEE Transactions on Networkand Service Management, 2021, 18(2): 1491-1507. DOI:10.1109 / TNSM.2021.3056917." proposes to reliably place service function chains and virtual monitoring functions with minimum cost in software-based 5G networks, and formulates the problem as an integer linear programming (ILP) to minimize the total deployment cost.

[0012] A common thread among the aforementioned studies is that service or resource orchestration is performed in managed wireless networks, where the infrastructure provides information useful for orchestration. However, these studies do not cover orchestration scenarios in infrastructure-free wireless environments. During the orchestration and collaboration process, self-organizing nodes may not receive assistance from base stations (because there is no base station or the base station is too far away), so their orchestration algorithms need to be redesigned.

[0013] 2. Orchestration in edge networks

[0014] The paper "LAROUI M, KHEDHER HI, MOUNGLA H, et al. Service function chains multi-resource orchestration in virtual mobile edge computing[J / OL]. Computer Networks, 2023, 224: 109582. DOI:https: / / doi.org / 10.1016 / j.comnet.2023.109582." comprehensively considers various system costs such as server power consumption and wireless transmission power consumption, combines wireless multicast to establish an optimization model, reduce system costs, and provide efficient services for mobile nodes (including IoT devices, user-end systems, etc.).

[0015] The paper "LI H, KORDI M E. DSPPV: Dynamic service function chainsplacement with parallelized virtual network functions in mobile edgecomputing[J / OL]. Internet of Things, 2023, 22: 100733. DOI:https: / / doi.org / 10.1016 / j.iot.2023.100733." proposes a parallel dynamic service function chain deployment method to accelerate the computing process, minimize the length of the service function chain, and improve the ability to process future requests through parallel VNFs.

[0016] The paper "LI J, GUO D, XIE J, et al. Availability-aware provision of service function chains in mobile edge computing[J / OL]. ACM Trans. Sen. Netw., 2023,19(3). https: / / doi.org / 10.1145 / 3565483. DOI:10.1145 / 3565483." takes into account the reliability of hardware and software to maximize the reliability of service function chains in each service delivery.

[0017] The paper "NGUYEN TTL, PHAM TM, PHAM L M. Efficient redundancy allocation for reliable service function chains in edge computing[J / OL]. Journal of Network and Systems Management, 2022, 31(1)" proposes a service availability-aware deployment scheme that aims to improve the processing speed and reliability of user requests. This scheme mainly solves the problem of effectively mapping primary and backup virtual network functions while reducing user access latency.

[0018] The above studies follow a server-centric perspective, which orchestrates functions and resources to serve mobile nodes, which means that mobile nodes are service consumers rather than function providers or service orchestration nodes to complete complex ad hoc tasks, and do not meet the requirements of on-demand on-site orchestration of mobile self-organizing nodes.

[0019] 3. Routing in Self-Organizing Networks

[0020] The paper "CHEN X, SUN G, WU T, et al. RANCE: A Randomly Centralized and On-Demand Clustering Protocol for Mobile Ad Hoc Networks[J / OL]. IEEE Internet of Things Journal, 2022, 9(23): 23639-23658. DOI:10.1109 / JIOT.2022.3188679." proposes a random centralized on-demand clustering protocol for mobile ad hoc networks, aiming to extend the node's clustering time to support inter-node collaboration.

[0021] The paper "LAVANYA K, INDIRA R, VELMURUGAN AK, et al. Mobility-Based Optimized Multipath Routing Protocol on Optimal Link State Routing in MANET[C / OL] / / 2023 International Conference on Applied Intelligence and Sustainable Computing (ICAISC). 2023: 1-6[2024-02-16]. https: / / ieeexplore.ieee.org / document / 10199484. DOI:10.1109 / ICAISC58445.2023.10199484." proposes a mobility-based multipath routing protocol and a link state transmission algorithm for QoS-awareness in mobile ad hoc networks.

[0022] The paper "KUMBHAR FH, YOUNG SHIN S. CV-AODV: Compatibility BasedVehicular Ad-hoc On Demand Distance Vector Routing Protocol[C / OL] / / 2020International Conference on Information and Communication TechnologyConvergence (ICTC). 2020: 1004-1008[2024-02-16]." proposes compatibility-based on-demand distance vector routing for vehicle networks to improve vehicle connectivity.

[0023] The paper "CUI Y, ZHANG Q, FENG Z, et al. Topology-Aware Resilient Routing Protocol for FANETs: An Adaptive Q-Learning Approach[J / OL]. IEEE Internet of Things Journal, 2022, 9(19): 18632-18649. DOI:10.1109 / JIOT.2022.3162849." proposes a FANET topology-aware resilient routing protocol based on Q-learning method, which captures topology changes (especially in UAV cooperative environments) with low overhead and makes routing decisions in a distributed and autonomous manner.

[0024] The aforementioned research primarily focuses on the fundamental routing problem, specifically the communication aspect. The orchestration of heterogeneous functions deployed on mobile ad hoc nodes has been less well addressed. The requirement to traverse the different functions in the order requested along the optimal communication path across the mobile wireless nodes makes this problem more complex than pure routing. Summary of the Invention

[0025] To address the above-mentioned issues, the present invention provides a method, device, system, and storage medium for orchestrating mobile self-organizing network service function chains, enabling the rational orchestration of heterogeneous mobile self-organizing nodes in an infrastructure-free environment. An infrastructure-free environment means that no base station assistance is provided during the chain orchestration and collaboration process, and no clustering is pre-established through protocols such as ZigBee and LEACH. All nodes can operate equally and can freely participate in or leave orchestration-based collaboration. The present invention is well-suited for complex collaborative tasks, and is particularly suitable for remote areas with underdeveloped communication infrastructure.

[0026] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0027] According to a first solution of the present invention, a method for orchestrating a mobile ad hoc network service function chain is provided, the method comprising:

[0028] Obtaining an orchestration request and a starting node; wherein the orchestration request includes multiple entities, each of which is a node or a function;

[0029] inserting the start node into a header of the orchestration request to obtain a pre-processed orchestration request, and dividing the pre-processed orchestration request into a plurality of node-entity anchor point pairs;

[0030] Capturing topology dynamics in a mobile ad hoc network based on a time-constrained recursive topology discovery protocol; wherein the topology dynamics include topology information, functional information, communication information, and performance information;

[0031] The network forest graph is dynamically constructed using the topology; wherein the network forest graph is represented as , For nodes, is the communication link between nodes, For function, The functional link between the node and the function is weighted by the communication cost between the nodes, and the communication cost is determined according to the topology information and the communication information; the functional link between the node and the function is weighted by the performance cost, and the performance cost is determined according to the function information and the performance information;

[0032] Based on the mesh forest graph, with the goal of minimizing the sum of communication cost and performance cost, the optimal segment corresponding to each node-entity anchor point pair is determined, and the optimal service function chain is obtained by connecting the optimal segments.

[0033] Furthermore, a time-constrained recursive topology discovery protocol is proposed to capture the topology dynamics in mobile ad hoc networks, including:

[0034] A time-constrained recursive topology discovery protocol is proposed to capture topology dynamics in mobile ad hoc networks using on-demand mode, periodic mode, and / or hybrid mode.

[0035] In on-demand mode, each node initiates a recursive topology discovery process only when an orchestration request arrives to capture topology dynamics. The process is as follows:

[0036] (1) When any node When receiving an orchestration request, it becomes an orchestration node, starts the topology discovery process, generates and broadcasts a topology discovery request ,in Global topology discovery timer , the orchestration node is During this period, wait for other nodes to reply to the discovered topology. Request signal for topology discovery;

[0037] (2) When any node Receive a topology discovery request hour:

[0038] Immediately reply to the node that sent the topology discovery request with its own node information and the currently known topology dynamics ,in, , Indicates functional information, representing The functionality provided is represented by , Indicates whether the corresponding function can be provided; Indicates communication information, representing The various communication quality QoS indicators are expressed as , For the The value of a QoS indicator; Indicates performance information, representing The various performance indicators during function call are expressed as , For the The value of the performance indicator;

[0039] Set the local topology discovery timer , Used to characterize the time allowed for wireless round trip and processing, During this period, it waits for other nodes to reply to the discovered topology and broadcasts the forwarding topology discovery request. ;

[0040] (3) When any node Topology discovery timer time out:

[0041] If the node Is the arrangement node, in the current mesh forest graph Start the orchestration;

[0042] If the node Not an orchestration node, go up one node again Reply to newly found topology dynamics ;

[0043] (4) When any node Received from other nodes Topological dynamics of the reply When , it is merged with its own topology dynamics, and the communication cost and performance cost are calculated to weight the communication link and functional link respectively;

[0044] In periodic mode, any node sets a topology exchange timer , used for periodic exchange, the process is as follows:

[0045] (2) When any node Topology discovery timer Timeout, broadcast node Topology , and reset the topology discovery timer ;

[0046] (2) When any node Received from other nodes Topological dynamics of the reply When the node With node The topology is dynamically merged, and the communication cost and performance cost are calculated to weight the communication link and functional link respectively;

[0047] (3) When any node When receiving an orchestration request, it becomes an orchestration node and Start the orchestration;

[0048] In hybrid mode, both on-demand mode and periodic mode are integrated and work in a concurrent thread manner, including the following multiple concurrently running threads:

[0049] The thread that processes orchestration requests. This thread is shared by both on-demand and periodic modes and processes received orchestration requests in real time. In on-demand mode, it triggers on-demand topology discovery, while in periodic mode, it directly triggers service function chain orchestration.

[0050] The thread that processes the topology graph arriving is shared by both on-demand and periodic modes. It processes the topology graphs sent by other nodes in real time and calculates the communication cost and performance cost to weight the communication links and functional links respectively.

[0051] The periodic topology discovery thread is used only for periodic topology discovery and starts the regular topology map broadcast and exchange process in periodic mode.

[0052] Furthermore, the starting node is inserted into the header of the orchestration request to obtain a pre-processed orchestration request, and the pre-processed orchestration request is divided into a plurality of node-entity anchor pairs, including:

[0053] The starting node Insert into the header of the orchestration request to obtain the pre-processed orchestration request , the current pointer points to the first element of the arrangement request, take the pointer ,in Indicates the entities, Indicates the entities, is the total number of entities;

[0054] when When the starting node As the starting anchor point ;

[0055] when When , according to the previously found optimal segmentation ,by The second-to-last element of is used as the starting anchor point;

[0056] Determine the end anchor point ;

[0057] Forming a branch node-entity anchor pair, .

[0058] Furthermore, based on the mesh forest graph, with the goal of minimizing the sum of communication cost and performance cost, the optimal segmentation corresponding to each node-entity anchor point pair is determined, including:

[0059] Based on the mesh forest diagram, a mesh tree diagram is dynamically generated according to the functions included in the node-entity anchor point pairs; wherein, in the mesh tree diagram, only the functions included in the node-entity anchor point pairs and the functional links related to the functions, as well as all nodes and communication links in the mesh forest diagram are retained.

[0060] Furthermore, based on the mesh forest graph, with the goal of minimizing the sum of communication cost and performance cost, the optimal segment corresponding to each node-entity anchor point pair is determined, and the optimal service function chain is obtained by connecting the optimal segments, including:

[0061] use The algorithm determines the optimal segmentation between the starting anchor point and the ending anchor point in the current node-entity anchor point on the network tree diagram;

[0062] Connect the optimal segment between the starting anchor point and the ending anchor point in the current node-entity anchor point with the determined optimal segment, and extend the service function chain section by section;

[0063] The currently determined optimal segmentation is used as the starting point for the next segmentation until all node-entity anchor pairs are determined and connected in sequence to obtain the optimal service function chain.

[0064] Furthermore, the communication cost is calculated as follows:

[0065] Get Node No. QoS indicators The corresponding communication cost;

[0066] Different QoS indicators are converted using the following formula:

[0067] Delay cost: ;

[0068] Packet loss rate cost: ;

[0069] Jitter cost: ;

[0070] Bandwidth cost: ,in It is the bandwidth reference value, which can be set according to actual conditions;

[0071] Normalize the converted cost:

[0072] ;

[0073] Where, represents the normalized communication cost, Indicates the node number, Indicates the total number of nodes;

[0074] The nodes are calculated by the following formula Aggregate communication cost :

[0075] ;

[0076] Where, is the weight of each single communication cost, N is the total number of QoS indicators;

[0077] node To Node Communication Link The communication cost Empowerment is as follows: .

[0078] Furthermore, the performance cost is calculated as follows:

[0079] Get Node No. performance indicators The corresponding performance cost;

[0080] Different performance indicators are converted using the following formula:

[0081] CPU cost: ;

[0082] RAM cost: ;

[0083] Battery price: ,in It is the battery baseline value, which can be set according to actual conditions;

[0084] Reliability cost: ;

[0085] Performance cost after conversion Normalize:

[0086] ;

[0087] Where, represents the normalized performance cost;

[0088] The nodes are calculated by the following formula Aggregation performance cost:

[0089] ;

[0090] Where, is the weight of each single performance cost, is the total number of performance indicators;

[0091] node To function Functional Link Performance cost Empowerment is as follows: .

[0092] According to a second technical solution of the present invention, a mobile self-organizing network service function chain arrangement device is provided, the device comprising:

[0093] A request acquisition module is configured to acquire an orchestration request and a starting node; wherein the orchestration request includes multiple entities, each of which is a node or a function;

[0094] a request partitioning module configured to insert the start node into a header of the orchestration request to obtain a pre-processed orchestration request, and partition the pre-processed orchestration request into a plurality of node-entity anchor point pairs;

[0095] A topology discovery module configured to capture topology dynamics in a mobile ad hoc network based on a time-constrained recursive topology discovery protocol; wherein the topology dynamics include topology information, functional information, communication information, and performance information;

[0096] A forest graph construction module is configured to dynamically construct a network forest graph using the topology; wherein the network forest graph is represented as , For nodes, is the communication link between nodes, For function, The functional link between the node and the function is weighted by the communication cost between the nodes, and the communication cost is determined according to the topology information and the communication information; the functional link between the node and the function is weighted by the performance cost, and the performance cost is determined according to the function information and the performance information;

[0097] The orchestration module is configured to determine the optimal segment corresponding to each node-entity anchor point pair based on the mesh forest graph, with the goal of minimizing the sum of communication cost and performance cost, and connect the optimal segments to obtain an optimal service function chain.

[0098] According to a third technical solution of the present invention, a mobile self-organizing network service function chain orchestration system is provided, the system comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the method described above.

[0099] According to a fourth technical solution of the present invention, a non-transitory computer-readable storage medium storing instructions is provided. When the instructions are executed by a processor, the method described above is executed.

[0100] This invention can be used for functional orchestration of heterogeneous mobile self-organizing nodes in environments without infrastructure (such as outdoor areas without base stations). All nodes operate equally and can freely participate in or exit orchestration-based collaboration. The main technical effects are:

[0101] (1) A time-constrained recursive topology discovery mechanism is proposed to quickly discover partial topologies in mobile ad hoc networks.

[0102] (2) A mesh-forest graph is proposed, which is a graph-compatible data structure that can comprehensively accommodate topology, function, communication, and performance information. The mesh-tree graph can be dynamically customized according to each required function of the orchestration request to ensure the consistency of the orchestrated service function chain with the actual topology structure.

[0103] (3) An efficient orchestration method is proposed to perform iterative segmented orchestration and splicing, thereby obtaining a service function chain that efficiently traverses the required functions. This method is also highly consistent with the actual topology structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0104] Figure 1 A network forest diagram according to an embodiment of the present invention is shown.

[0105] Figure 2 A schematic diagram of a false link problem of a network forest graph according to an embodiment of the present invention is shown.

[0106] Figure 3 A partial mesh forest diagram of the node itself is shown according to an embodiment of the present invention.

[0107] Figure 4 A partial mesh forest diagram of node merging neighbor discovery according to an embodiment of the present invention is shown.

[0108] Figure 5 A partial mesh forest diagram of node merging secondary neighbor discovery according to an embodiment of the present invention is shown.

[0109] Figure 6 FIG. 4 shows a network tree diagram of f2 according to an embodiment of the present invention.

[0110] Figure 7 A network tree diagram of f5 according to an embodiment of the present invention is shown.

[0111] Figure 8 A network tree diagram of f4 according to an embodiment of the present invention is shown.

[0112] Figure 9 A structural diagram of a mobile self-organizing network service function chain orchestration device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0113] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. The details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other unless they conflict.

[0114] The specific implementation of the present invention is further described in detail below with reference to the accompanying drawings and examples.

[0115] On-demand, on-site orchestration of mobile self-organizing network nodes plays an important role in environments lacking communication infrastructure, such as rescue and environmental monitoring. It requires arranging various functional combinations into service function chains (SFChains), whose constituent functions are called in the required order. SFChaining involves not only optimal communication path planning but also optimal function selection, making it a complex dual problem. It also faces issues such as the dynamically changing topology of mobile self-organizing networks. To address the above issues, an embodiment of the present invention provides a method for orchestrating service function chains in a mobile self-organizing network, which is used for orchestrating service function chains of functionally heterogeneous mobile self-organizing nodes in an environment without infrastructure (such as a field area without a base station). The implementation of this method relies on the time-constrained recursive topology discovery protocol constructed in this embodiment.

[0116] Service function chain orchestration relies heavily on the topology, function, communication, and performance information held by the orchestration node. If the orchestration node only knows the neighboring nodes within its wireless coverage area, but does not know the topological connectivity of more other nodes, orchestration will become extremely difficult. Therefore, topology discovery is necessary for successful orchestration. The "topology discovery" mentioned in this article will simultaneously discover topology, communication, function, and performance information in the process, the same below. In an environment without heterogeneous infrastructure and protocols, unlike topology discovery that serves routing purposes, discovering the complete topology is neither easy nor necessary for mobile self-organizing networks that require agile collaboration. Our goal is to orchestrate service function chains that meet functional and performance requirements in a timely manner. The process of quickly discovering partial topologies and efficient orchestration is more practical, especially in a large mobile self-organizing network. To this end, a time-constrained recursive topology discovery protocol (TIRE) is proposed. TIRE has the following characteristics:

[0117] First, TIRE gradually reduces the time it takes for each downstream node to discover topology, allowing the downstream nodes to feedback the results of topology discovery as early as possible, adapting to the time-varying topology characteristics of mobile ad hoc networks and ensuring the efficiency of topology discovery.

[0118] Second, TIRE can capture the topological dynamics in mobile ad hoc networks in multiple ways (on-demand mode, periodic mode, and hybrid mode).

[0119] Third, TIRE is designed as an application layer protocol that works according to its topological logic without considering the heterogeneity of network layer protocols.

[0120] Fourth, TIRE designs a two-layer graph data structure so that the discovered topology, functions, communications, performance information, etc. can be accommodated in a unified graph with quantized links, providing a numerical and data structure foundation for the service function chain in mobile self-organizing networks.

[0121] TIRE includes on-demand topology discovery, periodic topology discovery, and hybrid topology discovery.

[0122] In on-demand topology discovery, each node does not obtain network topology information in advance. Only when an orchestration request arrives does the recursive topology discovery process start to capture topology dynamics. The process is as follows:

[0123] (1) When any node When receiving an orchestration request, it becomes an orchestration node, starts the topology discovery process, generates and broadcasts a topology discovery request ,in Global topology discovery timer , the orchestration node is During this period, wait for other nodes to reply to the discovered topology. Request signal for topology discovery;

[0124] (2) When any node Receive a topology discovery request hour:

[0125] Immediately reply the node that sent the topology discovery request (called the "previous node") with its own node information and the currently mastered topology dynamics to improve the timeliness of topology discovery. , contains the key information for constructing the mesh-forest graph: Indicates functional information, representing What functions can be provided? , Indicates whether the corresponding functions can be provided, such as camera shooting, image recognition, target tracking, etc. Indicates communication information, representing The various communication quality QoS indicators are recorded as , For the The values ​​of QoS indicators, such as delay, packet loss rate, jitter, bandwidth, etc. Indicates performance information, representing The various performance indicators during function call are recorded as , For the The values ​​of various performance indicators, such as CPU, RAM, power, reliability, etc. All this information will be used to calculate the communication cost and performance cost, respectively weight the communication link and functional link, and be used for the orchestration of the mobile service function chain.

[0126] Set the local topology discovery timer , is a smaller than The amount of time used to characterize the time allowed for wireless round trips and processing, During this period, it waits for other nodes to reply to the discovered topology and broadcasts the forwarding topology discovery request. , delivered to more possible nodes that are beyond the signal coverage of the original orchestration node. This ensures that the downstream node always times out with the upstream node first, and feeds back the topology discovery results as early as possible, adapting to the time-varying topology characteristics of the mobile ad hoc network and ensuring the efficiency of topology discovery.

[0127] (3) When any node Topology discovery timer time out:

[0128] If the node It is an orchestration node that starts orchestration on the current topology;

[0129] If the node Not an orchestration node, go up one node again Reply to newly found topology dynamics To improve the richness of topology discovery;

[0130] (4) When any node Received from other nodes Topological dynamics of the reply When , it is merged with its own topology dynamics, and the communication cost and performance cost are calculated to weight the communication link and functional link respectively;

[0131] It should be noted that according to the above process, the receiving node Restore its topology at the beginning and end Figure 1 The first reply is to ensure the timeliness of topology discovery, and the second reply is to improve the richness of topology discovery. The first reply is optional.

[0132] In periodic topology discovery, any node sets a topology exchange timer regardless of whether it receives an orchestration request. , used for periodic exchange, the process is as follows:

[0133] (3) When any node Topology discovery timer Timeout, broadcast your own topology , and reset the topology discovery timer ;

[0134] (2) When any node Received from other nodes Topological dynamics of the reply When ,it is dynamically merged with its own topology and calculates the communication cost and the performance cost;

[0135] (3) When any node When an orchestration request is received, it becomes an orchestration node and starts orchestration on the current topology, as the topology has been discovered periodically in advance.

[0136] Hybrid topology discovery combines on-demand topology discovery and periodic topology discovery to overcome the shortcomings of running the two modes separately. On-demand topology discovery reflects the real-time status of the current wireless network because the information is obtained in real time. However, it will cause more orchestration delays because the topology discovery is initiated after receiving the orchestration request. On the other hand, periodic topology discovery can shorten the delay before orchestration because the topology discovery is always in progress periodically. However, due to the frequent periodic sending of the topology map, it may incur additional bandwidth overhead. In addition, it may provide non-real-time or outdated node and link status. Therefore, in addition to on-demand and periodic, TIRE can also be configured as a hybrid, which is also the main loop of TIRE. The TIRE main loop works in a concurrent thread manner and contains multiple concurrently running threads:

[0137] Routine on_orchestration_request( ): This is the thread that processes orchestration requests. It is code shared by on-demand mode and periodic mode, and processes received orchestration requests in real time. In on-demand mode, the on-demand topology discovery process (including discovery initialization, timer setting, etc.) is triggered, while in periodic mode, the service function chain orchestration is directly triggered.

[0138] Routine on_topo_graph_arrival( , ): This is the thread that processes the arrival of the topology graph. It is the code shared by the on-demand mode and the periodic mode. It processes the topology graphs transmitted by other nodes in real time, performs the above-mentioned dynamic merging of topologies, and calculates the communication cost and performance cost to respectively weight the communication links and functional links.

[0139] Routine on_topo_exchange_timer( ): This is a thread used only for periodic topology discovery, which starts the regular topology map broadcast exchange process in the above periodic mode.

[0140] Node communication information collected based on TIRE , calculate the communication cost between nodes, the steps are as follows:

[0141] Step 101: Representation node No. QoS indicators Corresponding communication costs. Different conversion methods are used for different QoS indicators, and their goal is to convert QoS indicators into additive costs (i.e., costs, where larger values ​​represent worse performance), as shown below:

[0142] Delay cost: ;

[0143] Packet loss cost: ;

[0144] Jitter cost: ;

[0145] Bandwidth cost: .

[0146] In order to compare different QoS indicators fairly, Normalize, that is , to eliminate the impact of different scales and units used by different QoS indicators.

[0147] Step 102: According to the conversion step of step 101, the node is obtained by weighted summation. Aggregate communication cost ,in is the weighted value.

[0148] ;

[0149] Step 103: Aggregate communication cost Assigned to a node All egress wireless links are therefore connected through the node The data communication will experience its communication cost .

[0150] For slave nodes To Node egress wireless link The communication cost is allocated in the following way, .

[0151] Node performance information collected based on TIRE and functional information ,The performance cost of computing nodes and function keys is ,followed by the following steps.

[0152] Step 201, Representation node No. performance indicators Corresponding performance cost. Different conversion methods are used for different performance indicators, and their goal is to convert the performance indicator into an additive cost (i.e., cost, where a larger value represents worse performance), as shown below:

[0153] CPU cost: ;

[0154] RAM cost: ;

[0155] Battery cost: ;

[0156] Availability cost: .

[0157] In order to compare different performance indicators fairly, Normalize, that is , to eliminate the effects of different scales and units used for different performance indicators.

[0158] Step 202: According to the previous conversion step, the node is obtained by weighted summation. Aggregation performance cost ,in is the weighted value.

[0159] ;

[0160] Step 203, The performance cost is assigned to all its functional links, i.e. , so through the node Call a function , will experience its performance cost .

[0161] The difference between TIRE and other topology discovery protocols is that it constructs a graph data structure that contains topology, function, communication, and performance information. Figure 1 The two-layer mesh forest diagram (mesh-forest) shown is denoted as , where the upper white vertices are functions, forming a function tier (in this case, there are 5 functions), recorded as , the black vertices at the bottom are mobile self-organizing nodes, forming a network tier (in this case, it contains 6 nodes), which is recorded as Function links (solid lines) between functions and nodes use performance costs Empower, denoted as , while the communication links (dashed lines) between mobile ad hoc nodes use the communication cost Empower, denoted as . Figure 1 Contains enough information to perform service function chaining. Figure 1 As can be seen in the figure, several nodes providing the same functionality form a two-layer tree, with the functionality being the root (white vertices) and the nodes being the leaves (black vertices). Clearly, multiple such trees can be found in a mobile ad hoc network, forming a forest. Simultaneously, the nodes themselves form a flat mesh. Therefore, the discovered topology is a mesh-forest graph, where the forest provides functionality and performance information, and the mesh provides topology and communication information.

[0162] The network forest graph contains rich topology, function, communication, and performance information, which is suitable for the orchestration of service function chains. Intuitively, for an orchestration request , the two required functions are shown in the previous and next paragraphs Using the two starting and ending anchor points as the starting and ending anchor points, we find the optimal segments with the lowest communication and performance costs, and then connect these segments together to form the optimal service function chain. This idea is called "function-function segmentation", but there are loopholes in the algorithm logic.

[0163] Since the network forest graph contains communication links , functional link If the network forest graph is not properly processed for the two types of links, functional links may be mistakenly identified as communication links, resulting in the problem of false communication links. Figure 2 As shown, due to the long distance, and Physically unreachable. However, using the “function-function segmentation” logic and When finding the optimal segment between and pass reachable, that is, the functional link is mistaken for a communication link, and then find the two segments separately 、 , and finally spliced ​​into an incorrect service function chain ,because and It is not physically reachable at all. Therefore, “function-function segmentation” cannot be used to orchestrate the mobile service function chain.

[0164] Logically, since functions are provided by nodes and communication is done between nodes, the service function must provide orchestration requests. A function in , you should find Topological segmentation, that is, a number of consecutive nodes Provides a communication link, where the last node provides a functional link to the It can be seen that the starting anchor point of such a segment is a node and the ending anchor point is a function, so it is called "node-function segmentation". This requires that each time the shortest path search of "node-function segmentation" is performed, the graph should only contain one function, and temporarily delete other irrelevant functions and their functional links to avoid the above-mentioned communication false link problem caused by irrelevant functions and their functional links. After processing, such a graph is called a mesh-tree graph, denoted as , the function layer contains only one function and its related functional links , used to find node-function segmentation , so that the service function chain provides functions . And so on, when you need to find When , another network tree diagram is generated , and is used to find topological segments .

[0165] Based on the above description of the basic principles of the method for orchestrating a service function chain in a mobile ad hoc network, this embodiment below will introduce specific steps for orchestrating a service function chain in response to different orchestration requests.

[0166] In one embodiment, when an orchestration request is received When , the following topological configuration of service function chain is found (where each line represents a node-function segment):

[0167] ;

[0168] Minimize the total cost of the service function chain, that is, .

[0169] When the received orchestration request only includes functions, the specific steps of the mobile ad hoc network service function chain orchestration method include:

[0170] Step 301: Input parameter preprocessing:

[0171] Step 3011, obtain input parameters: The algorithm takes two parameters as input, a) the arrangement request , contains multiple functions that must be traversed ; b) Starting node of the service function chain .

[0172] Step 3012, input preprocessing: In order to facilitate node-function segmentation, Insert into orchestration request The head, that is , and the current pointer points to the first element of the arrangement request, that is, take the pointer .

[0173] Step 302: Determine segment anchor points:

[0174] Step 3021, determine the starting anchor point.

[0175] when When , the starting anchor point is determined for the first optimal segment to be found, that is, , start choreographing the service function chain from scratch.

[0176] when When , it indicates that the first optimal segment has been found, and it is necessary to find the subsequent optimal segments and connect them one by one to form a service function chain. To do this, it is necessary to find the optimal segment based on the previous one. , to determine the starting anchor point for the optimal segment to be found later, and to search for the optimal segment. According to the rule of node-function segmentation, The last element is a function (it does not have the ability to forward data in the network), and the second to last element is a node (because only nodes can provide functions, so the second to last element must be a node, and it has the ability to forward data in the network and can connect different segments). Therefore, to connect the previous segment , the starting anchor point of this segment is The second to last element of , obviously is a node.

[0177] Determine the ending anchor point. ,according to The structure, obviously is a function.

[0178] Form segment anchor pairs, .

[0179] Step 303: construct a mesh-tree graph:

[0180] For the current function Dynamically generate a network tree diagram ,Right now , to avoid other functions and functional links from forming false communication links and other problems, and to ensure the consistency of the orchestration results with the actual topology. The operation will be the current function Keep in In the functional layer, other functions are all intersected. Temporarily delete so that , but also makes The irrelevant functional links in the are temporarily deleted.

[0181] Step 304: Find the optimal segment and connect it:

[0182] Step 3041, using Dijkstra algorithm in the network tree diagram Found on and The optimal segmentation between .

[0183] Step 3042, splice with the previously found segments to iteratively form a service function chain, i.e. .

[0184] Step 3043: Use the currently found segment as the starting point of the next segmentation, i.e. .

[0185] Step 3044, start the next iteration, i.e. ,exist In the case of , jump to step 302 to find the next segment and splice it.

[0186] In the above steps, in step 301, the input arrangement request It is a list of functions, and the function list is divided into node-function anchor pairs through node-function segmentation, that is, arranged by function constraints. However, anchors are not limited to ordered node-function pairs, nodes can also be mixed with functions as input to the algorithm In each iteration of step 302, Segmentation is performed to generate "node-entity" anchors, where "entity" can be a function or a node. This is referred to as "node-entity segmentation," and node-entity segmentation has multiple variations to meet the diverse requirements of service function chaining for nodes and functions.

[0187] In some embodiments, the orchestration is performed with node constraints, and the input orchestration request is , where each Is a node, only the required nodes are specified. This makes the anchor points continuous node-node pairs. The intersection operation in step 303 above No functionality is retained at all, i.e. in the tree diagram , without any functions and function links. Therefore, each iteration of the orchestration process is limited to traversing Each consecutive node in the segment that serves as an anchor point at the end of the segment , just adjust the corresponding code in step 3021 to , because the last entity of the previous segment It is a node itself and can be used as the starting anchor point for a new segment This variant is suitable for pure routing that does not require traversing any functions, which reflects the applicability of the method proposed in this invention and can be used for both service function chaining and traditional routing.

[0188] In some embodiments, the link is constrained by both function and node for orchestration, and the input orchestration request is , where each Can be a node or a function. is a node, it forms a node-node pair; when the currently traversed When is a function, it forms a node-function pair. This restricts both node and function traversal, making it suitable for complex tasks requiring simultaneous traversal of certain functions and nodes. Because this algorithm variant requires traversal of certain nodes, it is suitable for tasks that report status to one or more specific nodes (e.g., a master control node) for fine-grained control, demonstrating the polymorphic nature of the proposed method.

[0189] In some embodiments, considering that in some application scenarios, nodes may need to be avoided during the orchestration process of the service function chain due to security vulnerabilities, poor performance, heavy load, etc., it is necessary to mark the nodes to be avoided in the orchestration request. Node avoidance has different granularities. (1) Chain-level avoidance: Nodes are Marked as , so that it can be dynamically removed from the mesh forest during the orchestration process of the entire service function chain, that is, permanently avoided, and restored after the orchestration is completed. (2) Segment-level avoidance: Nodes in Marked as , so that it can be dynamically deleted when searching for the current optimal segment, i.e., temporarily avoided. Since the segment-level avoidance granularity is relatively fine, if some nodes no longer need to be avoided, they can be restored during the iterative search for the next optimal segment.

[0190] The implementation of the above variants benefits from the flexible and fast generation of the network forest graph and the corresponding network tree graph proposed in the present invention.

[0191] The feasibility and advancement of the present invention will be fully illustrated below with reference to specific examples.

[0192] Figures 3 to 5 An example of the topology discovery process of on-demand TIRE is shown. The node that initiates topology discovery. Figure 3 yes The original part of the topology discovered by itself is small due to its wireless coverage, which is not conducive to the arrangement of service function chains. Set a timer , multicast a topology discovery request. Neighbor and After receiving the request for topology discovery, recursively perform topology discovery and first send the existing topology map to , ensure timeliness and set timers for each , forward the topology discovery request and try to discover more topology structures. Timeout, if a new topology is discovered, and Send your topology map to , And the newly discovered topology forms a larger topological map, Figure 4 is the partial topology graph after merging. And so on, Figure 5 It is a partial topology obtained by further merging the topology discovered recursively by the secondary neighbors. Timer If the timeout is exceeded, the recursive topology discovery ends. Figure 5 This is the discovered mesh forest diagram, which will be used to orchestrate the service function chain later.

[0193] Example 1: Functional Constraint Orchestration

[0194] For the orchestration node, the orchestration request received is , according to the found network forest diagram (such as Figure 5 as shown).

[0195] Step 401: Input parameter preprocessing:

[0196] In order to facilitate node-function segmentation, Insert into orchestration request The head, that is , and the current pointer points to the first element of the arrangement request, that is, .

[0197] Step 402: Determine segment anchor points:

[0198] Step 4021, determine the starting anchor point.

[0199] when When , the starting anchor point is determined for the first optimal segment to be found, that is, , start choreographing the service function chain from scratch.

[0200] Step 4022, determine the end anchor point,

[0201] Step 4023, forming segment anchor pairs, .

[0202] Step 403: construct a mesh-tree graph:

[0203] For the current function Dynamically generate a network tree diagram ,Right now , to avoid other functions and functional links from forming false links and other problems, to ensure the consistency of the orchestration results with the actual topology, and also to make The irrelevant functional links in are temporarily deleted, and the resulting network tree diagram is as follows Figure 6 shown.

[0204] Step 404: Find the optimal segment and connect it:

[0205] Step 4041, using Dijkstra algorithm in the network tree diagram Found on and The optimal segment between ,like Figure 6 Indicated by the dotted line.

[0206] Step 4042, combine with the previously found segments to iteratively form a service function chain, i.e. .

[0207] Step 4043: Use the currently found segment as the starting point of the next segmentation, i.e. .

[0208] Step 4044, start the next iteration, i.e. ,exist In the case of , jump to step 4042, find the next segment and splice it. 、 The corresponding network tree diagram (such as Figures 7 and 8 As shown), find the segment 、 , and finally spliced ​​into a service function chain: .

[0209] Example 2: Node Constraint Orchestration

[0210] For the orchestration node, the orchestration request received is , according to the found network forest diagram (such as Figure 5 The basic principle and operation process are similar to those in Example 1 and will not be repeated here. The final result is the service function chain: .

[0211] Example 3: Function and Node Constraint Links

[0212] For the orchestration node, the orchestration request received is , according to the found network forest diagram (such as Figure 5 The basic principle and operation process are similar to those in Example 1 and will not be repeated here. The final result is the service function chain: .

[0213] Example 4:

[0214] For the orchestration node, the orchestration request received is , according to the found network forest diagram (such as Figure 5 The basic principle and operation process are similar to those in Example 1 and will not be repeated here. The final result is the service function chain: .

[0215] Figure 9 The present invention also provides a mobile self-organizing network service function chain arrangement device, such as Figure 9 As shown, the device includes:

[0216] The request acquisition module 901 is configured to acquire an orchestration request and a starting node; wherein the orchestration request includes multiple entities, each of which is a node or a function;

[0217] a request partitioning module 902 configured to insert the start node into the header of the orchestration request to obtain a pre-processed orchestration request, and partition the pre-processed orchestration request into a plurality of node-entity anchor pairs;

[0218] A topology discovery module 903 is configured to capture topology dynamics in a mobile ad hoc network based on a time-constrained recursive topology discovery protocol; wherein the topology dynamics include topology information, functional information, communication information, and performance information;

[0219] The forest graph construction module 904 is configured to dynamically construct a network forest graph using the topology; wherein the network forest graph is represented as , For nodes, is the communication link between nodes, For function, The functional link between the node and the function is weighted by the communication cost between the nodes, and the communication cost is determined according to the topology information and the communication information; the functional link between the node and the function is weighted by the performance cost, and the performance cost is determined according to the function information and the performance information;

[0220] The orchestration module 905 is configured to determine the optimal segment corresponding to each node-entity anchor point pair based on the mesh forest graph, with the goal of minimizing the sum of communication cost and performance cost, and connect the optimal segments to obtain the optimal service function chain.

[0221] It should be noted that the device described in this embodiment and the method described previously belong to the same technical concept, have the same technical principles, and can achieve the same beneficial effects, so they will not be described in detail here.

[0222] An embodiment of the present invention further provides a mobile self-organizing network service function chain orchestration system, the system comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the method described in any of the above embodiments.

[0223] An embodiment of the present invention further provides a non-transitory computer-readable storage medium storing instructions, and when the instructions are executed by a processor, the method described in any of the above embodiments is executed.

[0224] The above embodiments are only used to illustrate the present invention, and are not intended to limit the present invention. Ordinary technicians in the relevant technical field can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, all equivalent technical solutions also fall within the scope of the present invention.

Claims

1. A method for arranging a mobile self-organizing network service function chain, characterized in that: The method comprises: Obtaining an orchestration request and a starting node; wherein the orchestration request includes multiple entities, each of which is a node or a function; inserting the start node into a header of the orchestration request to obtain a pre-processed orchestration request, and dividing the pre-processed orchestration request into a plurality of node-entity anchor point pairs; Capturing topology dynamics in a mobile ad hoc network based on a time-constrained recursive topology discovery protocol; wherein the topology dynamics include topology information, functional information, communication information, and performance information; The network forest graph is dynamically constructed using the topology; wherein the network forest graph is represented as , For nodes, is the communication link between nodes, For function, The functional link between the node and the function is weighted by the communication cost between the nodes, and the communication cost is determined according to the topology information and the communication information; the functional link between the node and the function is weighted by the performance cost, and the performance cost is determined according to the function information and the performance information; Based on the mesh forest graph, with the goal of minimizing the sum of communication cost and performance cost, determining the optimal segment corresponding to each node-entity anchor point pair, and connecting the optimal segments to obtain an optimal service function chain; A time-constrained recursive topology discovery protocol is used to capture topology dynamics in mobile ad hoc networks, including: A time-constrained recursive topology discovery protocol is proposed to capture topology dynamics in mobile ad hoc networks using on-demand mode, periodic mode, and / or hybrid mode. In on-demand mode, each node initiates a recursive topology discovery process only when an orchestration request arrives to capture topology dynamics. The process is as follows: (1) When any node When receiving an orchestration request, it becomes an orchestration node, starts the topology discovery process, generates and broadcasts a topology discovery request ,in Global topology discovery timer , the orchestration node is During this period, wait for other nodes to reply to the discovered topology. Request signal for topology discovery; (2) When any node Receive a topology discovery request hour: Immediately reply to the node that sent the topology discovery request with its own node information and the currently known topology dynamics ,in, , Indicates functional information, representing The functionality provided is represented by , Indicates whether the corresponding function can be provided; Indicates communication information, representing The various communication quality QoS indicators are expressed as , For the The value of a QoS indicator; Indicates performance information, representing The various performance indicators during function call are expressed as , For the The value of the performance indicator; Set the local topology discovery timer , Used to characterize the time allowed for wireless round trip and processing, During this period, it waits for other nodes to reply to the discovered topology and broadcasts the forwarding topology discovery request. ; (3) When any node Topology discovery timer time out: If the node Is the arrangement node, in the current mesh forest graph Start the orchestration; If the node Not an orchestration node, go up one node again Reply to newly found topology dynamics ; (4) When any node Received from other nodes Topological dynamics of the reply When , it is merged with its own topology dynamics, and the communication cost and performance cost are calculated to weight the communication link and functional link respectively; In periodic mode, any node sets a topology exchange timer , used for periodic exchange, the process is as follows: (1) When any node Topology discovery timer Timeout, broadcast node Topology , and reset the topology discovery timer ; (2) When any node Received from other nodes Topological dynamics of the reply When the node With node The topology is dynamically merged, and the communication cost and performance cost are calculated to weight the communication link and functional link respectively; (3) When any node When receiving an orchestration request, it becomes an orchestration node and Start the orchestration; In hybrid mode, both on-demand mode and periodic mode are integrated and work in a concurrent thread manner, including the following multiple concurrently running threads: The thread that processes orchestration requests. This thread is shared by both on-demand and periodic modes and processes received orchestration requests in real time. In on-demand mode, it triggers on-demand topology discovery, while in periodic mode, it directly triggers service function chain orchestration. The thread that processes the topology graph arriving is shared by both on-demand and periodic modes. It processes the topology graphs sent by other nodes in real time and calculates the communication cost and performance cost to weight the communication links and functional links respectively. The periodic topology discovery thread is used only for periodic topology discovery and starts the regular topology map broadcast and exchange process in periodic mode.

2. The method for arranging a mobile self-organizing network service function chain according to claim 1, wherein: Inserting the start node into the header of the orchestration request to obtain a pre-processed orchestration request, and dividing the pre-processed orchestration request into a plurality of node-entity anchor point pairs, including: The starting node Insert into the header of the orchestration request to obtain the pre-processed orchestration request , the current pointer points to the first element of the arrangement request, take the pointer ,in Indicates the entities, Indicates the entities, is the total number of entities; when When the starting node As the starting anchor point ; when When , according to the previously found optimal segmentation ,by The second-to-last element of is used as the starting anchor point; Determine the end anchor point ; Forming a branch node-entity anchor pair, .

3. The method for arranging a mobile self-organizing network service function chain according to claim 2, wherein: Based on the mesh forest graph, with the goal of minimizing the sum of communication cost and performance cost, the optimal segmentation corresponding to each node-entity anchor point pair is determined, including: Based on the mesh forest diagram, a mesh tree diagram is dynamically generated according to the functions included in the node-entity anchor point pairs; wherein, in the mesh tree diagram, only the functions included in the node-entity anchor point pairs and the functional links related to the functions, as well as all nodes and communication links in the mesh forest diagram are retained.

4. The method for arranging a mobile self-organizing network service function chain according to claim 3, wherein: Based on the mesh forest graph, with the goal of minimizing the sum of communication cost and performance cost, the optimal segment corresponding to each node-entity anchor point pair is determined, and the optimal service function chain is obtained by connecting the optimal segments, including: use The algorithm determines the optimal segmentation between the starting anchor point and the ending anchor point in the current node-entity anchor point on the network tree diagram; Connect the optimal segment between the starting anchor point and the ending anchor point in the current node-entity anchor point with the determined optimal segment, and extend the service function chain section by section; The currently determined optimal segmentation is used as the starting point for the next segmentation until all node-entity anchor pairs are determined and connected in sequence to obtain the optimal service function chain.

5. The method for arranging a mobile self-organizing network service function chain according to claim 1, wherein: The communication cost is calculated as follows: Get Node No. QoS indicators The corresponding communication cost; Different QoS indicators are converted using the following formula: Delay cost: ; Packet loss rate cost: ; Jitter cost: ; Bandwidth cost: ,in It is the bandwidth reference value, which can be set according to actual conditions; Normalize the converted cost: ； Where, represents the normalized communication cost, Indicates the node number, Indicates the total number of nodes; The nodes are calculated by the following formula Aggregate communication cost : ； Where, is the weight of each single communication cost, N is the total number of QoS indicators; node To Node Communication Link The communication cost Empowerment is as follows: .

6. The method for arranging a mobile self-organizing network service function chain according to claim 1, wherein: The performance cost is calculated as follows: Get Node No. performance indicators The corresponding performance cost; Different performance indicators are converted using the following formula: CPU cost: ; RAM cost: ; Battery price: ,in It is the battery baseline value, which can be set according to actual conditions; Reliability cost: ; Performance cost after conversion Perform normalization: ； Where, represents the normalized performance cost; The nodes are calculated by the following formula Aggregation performance cost: ； Where, is the weight of each single performance cost, is the total number of performance indicators; node To function Functional Link Performance cost Empowerment is as follows: .

7. A mobile self-organizing network service function chain arrangement device, characterized in that: The device: A request acquisition module is configured to acquire an orchestration request and a starting node; wherein the orchestration request includes multiple entities, each of which is a node or a function; a request partitioning module configured to insert the start node into a header of the orchestration request to obtain a pre-processed orchestration request, and partition the pre-processed orchestration request into a plurality of node-entity anchor point pairs; A topology discovery module configured to capture topology dynamics in a mobile ad hoc network based on a time-constrained recursive topology discovery protocol; wherein the topology dynamics include topology information, functional information, communication information, and performance information; A forest graph construction module is configured to dynamically construct a network forest graph using the topology; wherein the network forest graph is represented as , For nodes, is the communication link between nodes, For function, The functional link between the function and the node is weighted by the communication cost between the nodes, and the communication cost is determined according to the topology information and the communication information; the functional link between the node and the function is weighted by the performance cost, and the performance cost is determined according to the function information and the performance information; an orchestration module configured to determine, based on the mesh forest graph and with the goal of minimizing the sum of communication cost and performance cost, the optimal segments corresponding to each node-entity anchor point pair, and connect the optimal segments to obtain an optimal service function chain; Based on the mesh forest graph, with the goal of minimizing the sum of communication cost and performance cost, determining the optimal segment corresponding to each node-entity anchor point pair, and connecting the optimal segments to obtain an optimal service function chain; A time-constrained recursive topology discovery protocol is used to capture topology dynamics in mobile ad hoc networks, including: A time-constrained recursive topology discovery protocol is proposed to capture topology dynamics in mobile ad hoc networks using on-demand mode, periodic mode, and / or hybrid mode. In on-demand mode, each node initiates a recursive topology discovery process only when an orchestration request arrives to capture topology dynamics. The process is as follows: (1) When any node When receiving an orchestration request, it becomes an orchestration node, starts the topology discovery process, generates and broadcasts a topology discovery request ,in Global topology discovery timer , the orchestration node is During this period, wait for other nodes to reply to the discovered topology. Request signal for topology discovery; (2) When any node Receive a topology discovery request hour: Immediately reply to the node that sent the topology discovery request with its own node information and the currently known topology dynamics ,in, , Indicates functional information, representing The functionality provided is represented by , Indicates whether the corresponding function can be provided; Indicates communication information, representing The various communication quality QoS indicators are expressed as , For the The value of a QoS indicator; Indicates performance information, representing The various performance indicators during function call are expressed as , For the The value of the performance indicator; Set the local topology discovery timer , Used to characterize the time allowed for wireless round trip and processing, During this period, it waits for other nodes to reply to the discovered topology and broadcasts the forwarding topology discovery request. ; (3) When any node Topology discovery timer time out: If the node Is the arrangement node, in the current mesh forest graph Start the orchestration; If the node Not an orchestration node, go up one node again Reply to newly found topology dynamics ; (4) When any node Received from other nodes Topological dynamics of the reply When , it is merged with its own topology dynamics, and the communication cost and performance cost are calculated to weight the communication link and functional link respectively; In periodic mode, any node sets a topology exchange timer , used for periodic exchange, the process is as follows: (1) When any node Topology discovery timer Timeout, broadcast node Topology , and reset the topology discovery timer ; (2) When any node Received from other nodes Topological dynamics of the reply When the node With node The topology is dynamically merged, and the communication cost and performance cost are calculated to weight the communication link and functional link respectively; (3) When any node When receiving an orchestration request, it becomes an orchestration node and Start the orchestration; In hybrid mode, both on-demand mode and periodic mode are integrated and work in a concurrent thread manner, including the following multiple concurrently running threads: The thread that processes orchestration requests. This thread is shared by both on-demand and periodic modes and processes received orchestration requests in real time. In on-demand mode, it triggers on-demand topology discovery, while in periodic mode, it directly triggers service function chain orchestration. The thread that processes the topology graph arriving is shared by both on-demand and periodic modes. It processes the topology graphs sent by other nodes in real time and calculates the communication cost and performance cost to weight the communication links and functional links respectively. The periodic topology discovery thread is used only for periodic topology discovery and starts the regular topology map broadcast and exchange process in periodic mode.

8. A mobile self-organizing network service function chain orchestration system, characterized by: The system comprises: Memory for storing computer programs; A processor, configured to execute the computer program to implement the method according to any one of claims 1 to 6. 9 . A non-transitory computer-readable storage medium storing instructions, which, when executed by a processor, executes the method according to claim 1 .

Images