A method and system for implementing dynamic load balancing QoS guarantee in drone networks

Abstract

The present invention discloses a method and system for realizing dynamic load balancing QoS guarantee of unmanned aerial vehicle network. The QoS scheme is subdivided into four modules, and the network module and the control module are constructed using SDN technology. The characteristics of separation of data plane and control plane are utilized to realize network information management and control, especially traffic information. The queue module uses the idea of ​​differential service, configures queues with the help of TC tools to realize differentiated services, distinguishes link traffic through DSCP fields, allocates bandwidth according to different types of services, and provides sufficient link bandwidth for high-speed services. The routing module considers the impact of link bandwidth utilization on the QoS of unmanned aerial vehicle, combines bandwidth utilization with the shortest path algorithm, provides real-time traffic information through the control module, realizes dynamic load balancing routing calculation, and reduces the occurrence of high-bandwidth occupied links. Finally, an SDN unmanned aerial vehicle network is built for simulation implementation. The results show that the present invention has significant improvements in QoS indicators such as link bandwidth utilization, delay, packet loss rate and jitter.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a method and system for implementing dynamic load balancing QoS guarantee in an unmanned aerial vehicle network. Background Art

[0002] As a product of the information and automation era, drones have played a significant role in daily life in recent years and have broad development prospects, such as in freight logistics, post-disaster search and rescue, traffic law enforcement, film shooting and other fields. With the popularity of drones in all walks of life, it is a future development trend to expand from a single application to networked work. Drones will improve work efficiency in various fields through cluster operation. Compared with a single drone system, a multi-drone system can perform tasks more economically and effectively. At the same time, drone networks are also called flying ad-hoc networks (FANET), which is a subset of mobile ad-hoc networks (MANET).

[0003] The birth of software defined network (SDN) provides a new solution for drone self-organizing network. Software defined network decouples traditional network, separates control plane and data plane, and provides network developers with a flexible way to arrange the network. Among them, the control plane is responsible for coordinating the topology information of the whole network, planning the routing path, and sending control information to the switch to realize control forwarding. The data plane is composed of representative switches that support OpenFlow protocol, which is responsible for receiving the information sent by the controller and performing corresponding processing to complete the task of forwarding business data.

[0004] The purpose of load balancing is to distribute loads to optimize resource usage, maximize throughput, minimize response time, and avoid network congestion. In drone networks, load balancing can be used for drone deployment and task scheduling to alleviate the problem of tight computing resources for single drone nodes. It can also be used to deploy drone base stations and network link bandwidth resources.

[0005] The purpose of Quality of Service (QoS) is to provide end-to-end service quality assurance for various services according to their different needs. Due to the dynamic changes in network topology caused by the mobility of drones, the QoS of drone networks is easily affected by the link status, resulting in a decrease in indicators such as bandwidth and latency. At the same time, load balancing is required to maintain stable network operation when bandwidth resources are limited. Therefore, how to design a suitable QoS guarantee solution is an urgent problem to be solved in drone networks. Summary of the invention

[0006] Aiming at the problem that the QoS of drone network is easily affected by the link status, resulting in a decrease in service quality, the present invention utilizes the data plane and control plane separation characteristics of SDN to design a method and system for realizing dynamic load balancing QoS guarantee in drone network, which can more flexibly realize functions such as dynamic topology discovery and QoS guarantee in drone network.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] On the one hand, the present invention provides a method for implementing dynamic load balancing QoS guarantee in a drone network, which uses a network module, a queue module, a control module, a routing module and the following steps:

[0009] S1. The different types of business data flows generated by the network module are sent to the queue module, and the network module sends the acquired network traffic information to the control module;

[0010] S2, the queue module designs the speed limit queue into two types: interactive services for two-way communication between drones and streaming services for real-time audio and video data transmission of drones. The service flows are marked with DSCP according to the differential service model to achieve differentiation. At the same time, the control module monitors the real-time network traffic and provides the real-time traffic information of the network module to the routing module for dynamic load balancing shortest path calculation.

[0011] S3, the routing module, combines bandwidth utilization with the shortest path algorithm, provides real-time traffic information through the control module, and calculates the shortest path for dynamic load balancing.

[0012] Furthermore, both the network module and the control module are built using SDN, and the separation of the data plane and the control plane is used to achieve network information management and control.

[0013] Furthermore, the network module is a network that simulates the characteristics of drone nodes. The drone nodes are composed of a host and an OpenFlow switch. Business data is sent out through the host and forwarded through multiple hops via the OpenFlow switch to reach the destination.

[0014] Furthermore, in the queue module in step S2, the DSCP field of the communication interaction service is set to 0 field, and the DSCP field of the streaming media service is set to 0x04.

[0015] Furthermore, the queue module in step S2 uses a token bucket filter queue.

[0016] Furthermore, the queue module TC queue configuration structure in step S2 is: first, configure the root queue rule as the htb queue; then, based on the root htb queue, configure two categories: Category 1 is used for streaming media services, with a maximum bandwidth limit of 10 mbps, and Category 2 is used for communication interaction services, with a bandwidth limit of 1 mbps; finally, the filter distinguishes the service flow by matching the content of the data packet header to achieve differentiated services.

[0017] Furthermore, the filter uses u32 match to match the DSCP field in the data packet and sets an additional offset, where the value of the offset is the number of bytes from the start field of the packet header to the DSCP field.

[0018] Furthermore, the dynamic load balancing shortest path algorithm process is as follows: replace the distance between drone nodes with the link bandwidth utilization, obtain the link bandwidth and link service flow, calculate the bandwidth utilization of each link in the network graph, use the shortest path algorithm, calculate the path with the minimum bandwidth utilization, and update the link service flow of the entire network and store it.

[0019] Furthermore, for a certain UAV network link e i The bandwidth utilization μ can be expressed as

[0020]

[0021] The denominator represents the current link bandwidth w e , the numerator is the sum of the traffic passing through the current link, and n represents the total number of traffic.

[0022] On the other hand, the present invention also provides a UAV network dynamic load balancing QoS guarantee system, comprising the following modules to implement the method of any one of claims 1-9:

[0023] The network module is a network that simulates the characteristics of drone nodes. It can generate data streams of different business types and send them to the queue module to achieve QoS differentiation. At the same time, it sends the acquired network traffic information to the control module.

[0024] The queue module uses a token bucket filter queue to design the rate limit queue into two types: interactive services for two-way communication between drones and streaming services for real-time audio and video data transmission of drones. The DSCP mark service flow is used to distinguish according to the differentiated service model. The DSCP field of the communication interactive service is set to 0, and the DSCP field of the streaming service is set to 0x04.

[0025] The control module uses the SDN controller to monitor the real-time network traffic, obtain the switch link bandwidth information, and calculate the shortest path for dynamic load balancing through the routing module according to the traffic situation, and issue the flow table;

[0026] The routing module is used to combine bandwidth utilization and the shortest path algorithm, provide real-time traffic information through the control module, and calculate the shortest path for dynamic load balancing.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] The dynamic load balancing QoS guarantee system and method for unmanned aerial vehicle networks provided by the present invention subdivides the QoS scheme into four modules, and the network module and the control module are constructed using SDN technology, and the characteristics of the separation of the data plane and the control plane are used to realize network information management and control, especially traffic information. The queue module uses the idea of ​​differential service, configures the queue with the help of the TC tool to realize differentiated service, distinguishes link traffic through the DSCP field, allocates bandwidth according to different types of service types, and provides sufficient link bandwidth for high-speed services. The routing module considers the impact of link bandwidth utilization on the QoS of unmanned aerial vehicles, combines bandwidth utilization with the shortest path algorithm, and provides real-time traffic information through the control module to realize dynamic load balancing routing calculation and reduce the occurrence of high-bandwidth occupied links. Finally, an SDN unmanned aerial vehicle network is built for simulation implementation. The results show that the unmanned aerial vehicle network dynamic load balancing QoS guarantee implementation method and system proposed by the present invention have obvious improvements in QoS indicators such as link bandwidth utilization, delay, packet loss rate and jitter, and can effectively reduce the occurrence of high-bandwidth occupied links. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0030] Figure 1 A flow chart of a method for implementing dynamic load balancing QoS guarantee in a drone network provided by an embodiment of the present invention.

[0031] Figure 2 This is a diagram of the TC queue configuration structure provided by an embodiment of the present invention.

[0032] Figure 3 A flow chart of a dynamic load balancing routing algorithm provided by an embodiment of the present invention.

[0033] Figure 4 A drone network location coordinate map provided for an embodiment of the present invention.

[0034] Figure 5 A topological relationship diagram of a drone network provided by an embodiment of the present invention

[0035] Figure 6 A comparison diagram of TC queue speed limits provided in an embodiment of the present invention.

[0036] Figure 7 A comparison chart of link bandwidth utilization provided by an embodiment of the present invention.

[0037] Figure 8 Idle link delay comparison diagram provided for an embodiment of the present invention

[0038] Fig. 9 A comparison diagram of busy link delays provided by an embodiment of the present invention.

[0039] Fig.10 The UDP data graph generated by Iperf provided in the embodiment of the present invention.

[0040] Fig.11 This is a comparison chart of the average packet loss rate of services provided by an embodiment of the present invention.

[0041] Fig.12 This is a comparison chart of the average jitter of services provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0043] The present invention proposes a method for realizing dynamic load balancing QoS guarantee in UAV network. Figure 1 As shown, a network module, a queue module, a control module, a routing module and the following steps are used:

[0044] S1. The different types of business data flows generated by the network module are sent to the queue module, and the network module sends the acquired network traffic information to the control module;

[0045] S2, the queue module designs the speed limit queue into two types: interactive services for two-way communication between drones and streaming services for real-time audio and video data transmission of drones. The service flows are marked with DSCP according to the differential service model to achieve differentiation. At the same time, the control module monitors the real-time network traffic and provides the real-time traffic information of the network module to the routing module for dynamic load balancing shortest path calculation.

[0046] S3, the routing module, combines bandwidth utilization with the shortest path algorithm, provides real-time traffic information through the control module, and calculates the shortest path for dynamic load balancing.

[0047] The following is a detailed introduction to each module.

[0048] (1) Network module: The network module is a network that simulates the characteristics of drone nodes. It can generate data flows of different service types and mark the DSCP field in the IP to achieve QoS differentiation. The present invention uses SDN drone network simulation, and the drone node is composed of a host and an OpenFlow switch. The network module is abstracted as a Mesh network composed of OpenFlow switches. At the same time, the service data is sent through the host and forwarded through multiple hops of the OpenFlow switch to reach the destination.

[0049] (2) Queue module: Considering the service content of the drone network, the speed limit queue is designed into two types, and the service flow is distinguished by using DSCP marking according to the differentiated service model. As shown in Table 1, one is the interactive service for two-way communication between drones, and the other is the streaming service for real-time audio and video data transmission that drones may need.

[0050] The real-time communication interaction service between drones is used to transmit some basic information. Usually, due to the relatively small amount of data, the bandwidth requirement is not high. However, in order to ensure the real-time and symmetric transmission, the communication interaction service needs to ensure low latency during the transmission process. The present invention sets the default 0 field as the DSCP field of the communication interaction service.

[0051] Streaming media services are usually one-way transmission services. In most cases, they have high requirements for link bandwidth due to the large amount of data transmitted. Compared with communication interaction services, streaming media services have relatively loose requirements for latency, but jitter has a greater impact on their service quality. The present invention sets the 0x04 field as the DSCP field of streaming media services.

[0052] Table 1 Classification of UAV business types

[0053] Business Type DSCP Field Bandwidth requirements Latency requirements Communication and interaction 0x00 Low high Streaming Media 0x04 high normal

[0054] Queues refer to the logic of sorting messages in a cache. When the service transmission rate exceeds the port bandwidth or exceeds the bandwidth limit set for the service traffic, the service data will be temporarily stored in the cache in the form of queues. The time and order of service data leaving the queue and the relationship between the data leaving each queue are determined by the queue scheduling algorithm. The present invention uses a token bucket filter queue.

[0055] like Figure 2 The TC queue configured in the QoS guarantee solution of the present invention is shown in FIG. The specific structure is:

[0056] First, configure the root queue rule as a hierarchical token bucket (HTB) queue. The hierarchical token bucket expands on the token bucket algorithm and supports the configuration of multiple queues with different bandwidth limits. Then, based on the root HTB queue, configure two categories. Category 1 is used for the streaming media services designed by the present invention, with a maximum bandwidth limit of 10 Mbps, and Category 2 is used for the communication interaction services designed by the present invention, with a bandwidth limit of 1 Mbps. The rate control of the two services is achieved through the HTB queue, providing streaming media services with more sufficient bandwidth usage. Finally, there is the filter, which distinguishes the service flow by matching the content of the data packet header to achieve differential service.

[0057] Therefore, the present invention uses u32 match to match the DSCP field in the data packet. At the same time, due to the dual MAC transmission mechanism of the SDN drone network, an additional layer of data message will be added to the data packet before forwarding. Therefore, an additional offset should be set when matching the DSCP field. The specific value is the number of bytes from the start field of the packet header to the DSCP field. After configuring the TC settings, the data packet with a DSCP field of 0x04 in the service will be forwarded from the port according to category 1 restrictions, and the default DSCP field of 0 will be forwarded from the port according to category 2 restrictions.

[0058] (3) Control module: The main function of the control module is to monitor the real-time network traffic and provide the traffic information in the network module to the routing module for dynamic load balancing calculation. The present invention uses the characteristics of the separation of the SDN control plane and the data plane to use the SDN controller to perform traffic supervision, obtain the switch link bandwidth information, calculate the path through the routing module according to the traffic situation, and issue the flow table.

[0059] (4) Routing module: Considering the UAV network structure, the multi-hop routing problem in the network can be abstracted as the pathfinding problem in graph theory.

[0060] If we consider drone nodes as vertices in the graph and links between drone nodes as edges in the graph, we can get the entire network topology as G = (V, E), where V is the vertex set of graph G, representing drone nodes in the drone network, and E is the edge set of graph G, representing links between drone nodes in the network. Each link is represented by e, then e∈E. Each drone link has its own bandwidth w e ,The bandwidth is affected by the signal strength between two nodes on the link.,When the UAV network transmits services, real-time traffic f will appear on the link.

[0061] The present invention distinguishes the real-time traffic according to the different source and destination addresses of the drone nodes. Therefore, there may be real-time traffic of multiple services on one link in the drone network. iThe bandwidth utilization μ can be expressed as:

[0062]

[0063] The denominator represents the current link bandwidth w e , the numerator is the sum of the traffic passing through the current link, and n represents the total number of traffic.

[0064] The present invention considers combining the UAV network link bandwidth utilization μ with the shortest path algorithm, and replacing the node-to-node distance with the link bandwidth utilization. At this time, through weighting, the shortest path algorithm can select the shortest path transmission service with less bandwidth occupancy according to the current link situation, avoiding high delays and packet loss rates caused by continuing to transmit services when the link is congested. By maintaining the full network link service flow list F and the link real-time bandwidth list W, the algorithm can derive the bandwidth utilization of each link and use it to calculate the shortest path. After calculating the shortest path, the full network link service flow list data is updated to achieve real-time dynamic load balancing shortest path calculation.

[0065] like Figure 3 The flowchart of the algorithm is shown. Compared with the traditional shortest path algorithm, this algorithm takes into account the influence of link bandwidth utilization and can avoid using high-load links to transmit services as much as possible during actual route selection.

[0066] Table 2 is the pseudo code of the algorithm. Assume that the entire network topology is G = (V, E), the link bandwidth list is W, and the link service flow list is F. Now a set of service data M is generated, where each item m i Each has its own starting address s i and the destination address d i , and assume that the default service rate is the same, which is v.

[0067] Table 2 Pseudo code of dynamic load balancing routing algorithm

[0068]

[0069] The simulation is implemented by building an SDN drone network. The method and results are as follows.

[0070] Considering the use of drone networks in real scenarios, the specific network deployment is as follows: Figure 4 As shown in the figure, the drone nodes transmit image data through streaming services and communicate in real time through communication interaction services.

[0071] like Figure 4This is the drone network location map created in Mininet-WiFi, with a total of nine nodes. Each node consists of a host and an OpenFlow switch. The controllers are located at nodes 2 and 6 to achieve multi-controller networking. The specific address settings of the network devices are shown in Table 3.

[0072] Table 3 UAV network simulation parameters

[0073] Simulation area 300x300 Drone altitude 10m Antenna gain 5dBi Transmit power 14dBm frequency 2.412GHz Coverage 74m

[0074] In order to more clearly display the network topology, the present invention uses NetworkX to draw a topology diagram of the nodes, such as Figure 5 shown.

[0075] In order to test the rate limit of different services for htb queues configured using TC tools, the present invention uses Iperf tool to generate TCP data stream, because TCP uses the maximum link bandwidth as best effort by default. Data stream is sent from node 1 to node 2, DSCP field is not set, each transmission is 10s, the average rate during the transmission time is counted, and a total of ten groups of simulations are performed.

[0076] like Figure 6 As shown in the figure, the service rate under the QoS solution of the present invention is obviously limited and maintained at about 1 mbps. The average service rate without adding TC queue speed limit fluctuates between 5 mbps and 6 mbps. The image shows that the token bucket queue realizes the rate limit for the matching service flow in and out of the switch port.

[0077] In order to facilitate the statistical test of bandwidth utilization, the present invention uses Iperf to simulate and generate five groups of streaming media service data flows, uses UDP transmission, sets the transmission speed to 500kbps and modifies the DSCP field to 0x04. The source and destination addresses of the service flows are respectively node 1 to node 4, node 1 to node 3, node 1 to node 8, node 6 to node 2, and node 7 to node 5, and are generated in sequence.

[0078] like Figure 7The figure shows a statistical bar graph of link bandwidth utilization when the above-mentioned business flow is transmitted. The blue solid line box in the figure represents the link bandwidth utilization without adding a QoS solution, and the red dotted line box represents the link bandwidth utilization using the dynamic load balancing QoS guarantee solution of the present invention. It can be seen from the figure that the link bandwidth utilization with QoS guarantee is distributed more evenly, without obviously larger or smaller values. However, links 1 to 3 without adding a QoS solution have a significantly larger bandwidth utilization, with a value of about 25%. This is because the default shortest path algorithm does not have the ability to calculate the path based on the initial traffic. When a link is used by multiple business flows, it is easy for the controller to continue to send down the flow table passing through the path, which will easily cause link congestion, and finally affect the overall business data transmission, and the network QoS quality will decrease. Therefore, the QoS solution of the present invention has obvious effects in reducing link bandwidth utilization and reducing the occurrence of congested links.

[0079] In order to test the latency in the drone network, the present invention uses the ping command to simulate the real-time communication interaction between drone nodes. The latency from node 1 to the remaining nodes is tested respectively. Each group of ping sends 10 ICMP packets, the default DSCP field is 0, and the average latency is recorded. Figure 8 It is the idle link delay when the link does not transmit streaming media services. At this time, since there is no high-bandwidth service occupation, the dynamic load balancing algorithm does not distinguish paths according to bandwidth utilization, and the average delay of the scheme without adding QoS and adding QoS is not much different.

[0080] In order to compare the impact of two different schemes on link delay when the link is busy, the present invention uses Iperf to send TCP data stream from node 1 to node 2, marking the DSCP field as 0x04. Continue to measure the delay from node 1 to the remaining nodes, send 10 ICMP packets each time, and take the average delay. Fig. 9 The figure shows the delay comparison of the two schemes under a busy link. It can be seen from the figure that there is no obvious difference in the delay from node 1 to its adjacent nodes 2 and 4, but as the number of hops increases, the delay from node 1 to the remaining nodes with the QoS guarantee scheme of the present invention is significantly smaller than that without the QoS guarantee scheme. This is due to the dual effects of the dynamic load balancing algorithm and the queue. The load balancing algorithm can avoid transmission on links with high bandwidth occupancy and reduce the impact of congested links on data packet delay. At the same time, the queue can reduce the mutual influence of the two services. Finally, the QoS guarantee scheme of the present invention has a lower delay when the link is congested and can provide a relatively stable connection.

[0081] The present invention uses Iperf software to send UDP data streams to count the service data packet delay and packet loss rate. The data stream still uses the five groups of data streams described above, each sending 10s, respectively, from node 1 to node 4, node 1 to node 3, node 1 to node 8, node 6 to node 2, node 7 to node 5. Fig.10 This is an example of using Iperf to measure. This simulation is performed for five groups, and the average values ​​of the five groups of jitter and packet loss rate for each service flow are calculated.

[0082] like Fig.11 and Fig.12 The figure is a comparison chart of the average packet loss rate and jitter of five groups of experiments. Due to the existence of direct links, there is no obvious difference in the packet loss rate and jitter of the two schemes for the data flows from node 1 to node 4, node 1 to node 3, and node 7 to node 5. However, for the two multi-hop forwarding service flows from node 1 to node 8 and node 6 to node 2, the packet loss rate and jitter of the scheme without adding QoS guarantee are significantly increased. This is because the dynamic load balancing routing algorithm can give priority to links with low bandwidth utilization for service transmission, reducing the impact of congested links on service QoS. When there are relatively busy links in the multi-hop links, the packet loss rate and jitter of the entire service will increase. The image shows that the QoS guarantee scheme used in the present invention is effective in reducing service packet loss rate and jitter.

[0083] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0084] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, electronic device embodiment, computer-readable storage medium embodiment, and computer program product embodiment, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0085] The above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the aforementioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the aforementioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical specialties therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for implementing dynamic load balancing QoS guarantee in a drone network, characterized in that: Using network module, queue module, control module, routing module and the following steps: S1. The different types of business data flows generated by the network module are sent to the queue module, and the network module sends the acquired network traffic information to the control module; S2, the queue module designs the speed limit queue into two types: interactive services for two-way communication between drones and streaming services for real-time audio and video data transmission of drones. The service flows are marked with DSCP according to the differential service model to achieve differentiation. At the same time, the control module monitors the real-time network traffic and provides the real-time traffic information of the network module to the routing module for dynamic load balancing shortest path calculation. S3, the routing module, combines bandwidth utilization and the shortest path algorithm, provides real-time traffic information through the control module, and calculates the shortest path for dynamic load balancing; The specific process includes: The drone nodes are regarded as vertices in the graph, and the links between drone nodes are regarded as edges in the graph. The entire network topology is G = (V, E), where V is the vertex set of graph G, representing the drone nodes in the drone network, and E is the edge set of graph G, representing the links between drone nodes in the network. Each link is represented by e, then e∈E, and each drone link has its own bandwidth w e ; Replace the distance between drone nodes with link bandwidth utilization, obtain link bandwidth and link business flow, calculate the broadband utilization of each link in the network diagram, use the shortest path algorithm to calculate the path with the minimum bandwidth utilization, update the link business flow of the entire network and store it.

2. The method for implementing dynamic load balancing QoS guarantee in a drone network according to claim 1, characterized in that: Both the network module and the control module are built using SDN, and the separation of the data plane and the control plane is used to achieve network information management and control.

3. The method for implementing dynamic load balancing QoS guarantee in a drone network according to claim 1, characterized in that: The network module is a network that simulates the characteristics of drone nodes. The drone nodes are composed of hosts and OpenFlow switches. Business data is sent out through the host and forwarded through multiple hops of the OpenFlow switch to reach the destination.

4. The method for implementing dynamic load balancing QoS guarantee in a drone network according to claim 1, characterized in that: In the queue module in step S2, the DSCP field of the communication interaction service is set to 0 field, and the DSCP field of the streaming media service is set to 0x04.

5. The method for implementing dynamic load balancing QoS guarantee in a drone network according to claim 1, characterized in that: The queue module in step S2 uses a token bucket filter queue.

6. The method for implementing dynamic load balancing QoS guarantee in a drone network according to claim 5 is characterized in that: The queue module TC queue configuration structure in step S2 is: first configure the root queue rule as the htb queue; then, based on the root htb queue, configure two categories: Category 1 is used for streaming media services, with a maximum bandwidth limit of 10 mbps, and Category 2 is used for communication interaction services, with a bandwidth limit of 1 mbps; finally, the filter distinguishes the service flow by matching the content of the data packet header to achieve differentiated services.

7. The method for implementing dynamic load balancing QoS guarantee in a drone network according to claim 6 is characterized in that: The filter uses u32 match to match the DSCP field in the packet and sets an additional offset, the value of the offset is the number of bytes from the start field of the packet header to the DSCP field.

8. The method for implementing dynamic load balancing QoS guarantee in a drone network according to claim 1, characterized in that: The dynamic load balancing shortest path algorithm process is as follows: replace the distance between drone nodes with link bandwidth utilization, obtain link bandwidth and link business traffic, calculate the bandwidth utilization of each link in the network diagram, use the shortest path algorithm, calculate the path with the minimum bandwidth utilization, update the link business traffic of the entire network and store it.

9. The method for implementing dynamic load balancing QoS guarantee in a drone network according to claim 8, characterized in that: For a drone network link e i Bandwidth utilization μ i Expressed as The denominator represents the current link bandwidth w e , the numerator is the sum of the traffic passing through the current link, and n represents the total number of traffic.

10. A UAV network dynamic load balancing QoS guarantee system, characterized in that: The method comprises the following modules to implement any one of claims 1 to 9: The network module is a network that simulates the characteristics of drone nodes. It can generate data streams of different business types and send them to the queue module to achieve QoS differentiation. At the same time, it sends the acquired network traffic information to the control module. The queue module uses a token bucket filter queue to design the rate limit queue into two types: interactive services for two-way communication between drones and streaming services for real-time audio and video data transmission of drones. The DSCP mark service flow is used to distinguish according to the differentiated service model. The DSCP field of the communication interactive service is set to 0, and the DSCP field of the streaming service is set to 0x04. The control module uses the SDN controller to monitor the real-time network traffic, obtain the switch link bandwidth information, and calculate the shortest path for dynamic load balancing through the routing module according to the traffic situation, and issue the flow table; The routing module is used to combine bandwidth utilization and the shortest path algorithm, provide real-time traffic information through the control module, and calculate the shortest path for dynamic load balancing.

Images