A software-defined tactical communication network system and communication method

Abstract

This invention discloses a software-defined tactical communication network architecture and communication method. The network architecture includes multiple network nodes running controller software within a tactical network. Several of these network nodes form a distributed network using an autonomous routing protocol, while the remaining network nodes form several local area networks (LANs), each controlled by a local controller. This architecture obtains a global resource view through centralized management, enabling flexible resource scheduling and efficient utilization. Distributed autonomous routing networking provides basic communication capabilities for the tactical communication network, enhancing its resilience. Based on this dual-plane network architecture, a lightweight southbound interface protocol, a link-state-based autonomous routing protocol, and a feature-based, efficient programmable forwarding implementation scheme are designed. These features reduce protocol overhead, provide more bandwidth and information resources for core control layer services, and adapt to diverse service requirements.

Description

Technical Field

[0001] This invention relates to the field of tactical communication technology, and in particular to a software-defined tactical communication network system and communication method. Background Technology

[0002] Tactical communication networks are characterized by low bandwidth, high mobility, and high survivability requirements. With the increasing demands of tactical communication users, higher requirements are being placed on network coverage, communication capacity, service quality, and real-time transmission. Traditional tactical communication network architectures are purely distributed, with tight coupling between control and forwarding, resulting in a closed architecture. This leads to significant problems such as slow deployment, difficult resource management, poor service quality, and weak adaptability, making it impossible to provide targeted and differentiated services.

[0003] Existing tactical communication networks are purely distributed network architectures. First, equipment, networks, and services are tightly integrated during design and construction, lacking flexibility and adaptability. This creates a huge workload for subsequent functional adjustments and development, leading to difficult network upgrades and long iteration cycles. Second, network node protocols, channel control protocols, and service transmission control are typically defined by a few manufacturers, lacking modular and open design. Users cannot quickly upgrade and optimize the network according to their own business needs, greatly hindering network service upgrades and application innovation. Third, the level of basic network service provisioning is insufficient. For example, when multiple services with different latency and packet loss quality of service requirements are concurrently running, high-demand services cannot receive priority service guarantees. In tactical network scenarios with limited bandwidth, users find it difficult to utilize network resources according to actual needs at any time. The network as a whole does not fully utilize existing bandwidth resources and lacks the ability to allocate resources on demand.

[0004] Therefore, there is an urgent need to propose a software-defined tactical communication network architecture and communication method. By optimizing the tactical communication network as a whole from the perspective of network architecture, and based on the optimized network architecture, the communication protocols and communication methods are designed in a targeted manner to adapt to the diverse service requirements in the tactical communication network. Summary of the Invention

[0005] In view of this, the present invention provides a software-defined tactical communication network system to solve the technical problems of insufficient flexibility and poor service capabilities in current tactical communication networks.

[0006] To achieve the above-mentioned technical objectives, the present invention adopts the following technical solution:

[0007] On one hand, the present invention provides a software-defined tactical communication network system, including multiple network nodes running controller software in a tactical network; wherein, a number of the network nodes are distributed and deployed through an autonomous routing protocol, and the remaining network nodes form a number of local area networks, each local area network being controlled by a local controller;

[0008] The network system includes a data plane, a local control plane, and a global control plane;

[0009] The data plane includes network nodes in each local area network (LAN) and distributed network nodes; the network nodes in the LAN interact with the local controller of their respective LAN via an improved TNCP protocol; the distributed network nodes perform autonomous routing and networking based on the TRP protocol and interact with the selected local controller via the improved TNCP protocol; all network nodes in the data plane perform matching and forwarding through multi-level flow tables based on feature recognition.

[0010] The local control plane includes multiple local controllers, which interact with each other via an east-west protocol;

[0011] The global control plane includes a top-level controller, which interacts with each local controller in the local control plane to perform global control of the network.

[0012] On the other hand, the present invention also provides a software-defined tactical communication network method, applied to the software-defined tactical communication network system described above, comprising:

[0013] The local controller in the local area network establishes a connection with the switch of the directly connected network node;

[0014] The local controller queries the routing table of the directly connected switch to obtain the IP addresses of each network node switch in the local area network, and interacts with each switch through the improved TNCP protocol based on the IP addresses.

[0015] When the network nodes are deployed in a distributed manner, all messages sent by the switches of the network nodes are forwarded by the switches directly connected to the local controller.

[0016] Both the switches in the local area network and the distributed switches use multi-level flow tables based on feature recognition for matching and forwarding.

[0017] Furthermore, the method also includes:

[0018] Switches in a local area network (LAN) use the link-state-based TRP protocol to establish a control channel with the local controller.

[0019] When the local controller is in the access state, the control link between the local controller and each switch in the local area network is completed.

[0020] Furthermore, during the connection establishment process between the local controller and the switch, the state of the local controller includes a disconnected state, a handshake successful state, and an access state.

[0021] The local controller is initially in a disconnected state;

[0022] When not connected, a handshake request message is sent to the switch, and the local controller starts a timer to wait for the handshake response message;

[0023] If the local controller receives a handshake response message within the first preset time, it enters the handshake success state and sends a service registration request message to the switch; if it receives a registration success response message within the second preset time, it enters the access state.

[0024] In the access state, the local controller successfully registers with the switch and is in a connection-maintaining state. When the connection maintenance fails, the local controller reverts to the disconnected state.

[0025] Furthermore, the improved TNCP protocol includes a particle swarm optimization (PSO)-based routing method for each switch to select the optimal routing path within the tactical network; the routing method includes:

[0026] The fitness function for evaluating the performance of each switch is determined based on link connectivity parameters, link bandwidth parameters, and link congestion parameters.

[0027] Initialize multiple switch candidate states representing routing selection, each candidate state having a velocity vector;

[0028] The fitness value corresponding to each candidate state is calculated based on the fitness function, and the individual optimal position and global optimal position of each candidate state are obtained.

[0029] The velocity vector of each candidate state is iteratively updated based on the individual optimal position and the global optimal position, and each candidate state is updated according to the updated velocity vector;

[0030] The fitness value of each candidate state is calculated based on the updated candidate states, and the global optimal position is updated according to the updated candidate states.

[0031] When the number of iterations reaches the preset number or the fitness value reaches the preset fitness threshold, the update is completed, and the optimal route selection for each switch is determined based on the global optimal position at the time of the update.

[0032] Furthermore, the link connection parameters include link rate and link latency; the link bandwidth parameters include link bandwidth and bandwidth utilization; and the link congestion parameters include load, number of congested links, and number of unused links.

[0033] The fitness function for evaluating the performance of each switch is determined based on link connectivity parameters, link bandwidth parameters, and link congestion parameters, including:

[0034] The objectives of the fitness function are: to maximize link speed, minimize link latency, maximize link bandwidth, maximize bandwidth utilization, minimize load, minimize the number of congested links, and minimize the number of unused links.

[0035] Furthermore, the matching and forwarding via a multi-level flow table based on feature recognition includes:

[0036] The data stream is categorized based on network data and network functions, and the data is forwarded based on the flow table corresponding to the matching domain of the categorized data.

[0037] Matching and forwarding is performed based on the switch's cache status and service priority.

[0038] Furthermore, data flows are categorized based on network data and network functions, and data is forwarded based on the flow table corresponding to the matching domain of the categorized flow, including:

[0039] Calculate the compatibility between each pair of different matching fields contained in each matching category in the single-level flow table stored by the switch;

[0040] The matching domain set is divided according to the compatibility, so that the sum of the compatibility of the matching domains located in the same set is maximized;

[0041] The single-level flow table is converted into a multi-level flow table based on the partitioned set, and the classification and processing of different data flows are realized based on the multi-level flow table structure.

[0042] Furthermore, matching and forwarding are performed based on the switch's cache status and service priority level, including:

[0043] Different service types are pre-prioritized to obtain priority queues; each priority queue has a preset congestion threshold value, and the corresponding cache space is allocated according to the congestion threshold value;

[0044] Determine the service type of the real-time data stream, and cache the data stream in the cache space of the corresponding priority queue for data forwarding based on the service type of the data stream.

[0045] Furthermore, matching and forwarding based on the switch's cache status and service priority also includes:

[0046] The system state transition probability is determined based on the packet arrival rate, packet service rate, and congestion threshold of the priority queue.

[0047] Calculate the overall blocking probability of the system based on the system state transition probability;

[0048] The weights and congestion thresholds for each priority level are optimized based on the overall congestion probability.

[0049] Compared to existing technologies, the software-defined tactical communication network system and communication method provided by this invention simplifies the traditional distributed network architecture, retaining only autonomous routing to ensure that the distributed network provides basic communication services. On this basis, a software-defined control plane is superimposed for centralized control of the tactical communication network, thereby improving the quality of service. Based on the proposed architecture, a lightweight southbound interface protocol and autonomous routing protocol suitable for tactical communication system environments are designed, reducing protocol overhead and providing more bandwidth and information resources for core control layer services. This promotes the full realization of the advantages of software-defined tactical networks, solving several problems in flexibility, openness, resource allocation, and network upgrades in traditional tactical communication networks, and adapting to the diverse service needs of future tactical communication networks. Attached Figure Description

[0050] Figure 1 A schematic diagram of the structure of the software-defined tactical communication network system provided by the present invention;

[0051] Figure 2 A functional diagram of the controller and switch provided by the present invention;

[0052] Figure 3 A framework diagram of the improved TNCP protocol provided by the present invention;

[0053] Figure 4 A schematic diagram of the improved TNCP protocol interaction process provided by the present invention;

[0054] Figure 5 This invention provides a schematic diagram of the state transition process of the TNCP protocol.

[0055] Figure 6 This is a schematic diagram of the multi-level flow table matching architecture provided by the present invention;

[0056] Figure 7 A schematic diagram of the state transition of the system provided by the present invention;

[0057] Figure 8 A schematic diagram showing the comparison of blocking probabilities of the system provided by the present invention under different thresholds;

[0058] Figure 9This is a schematic diagram comparing the blocking probabilities of the system provided by the present invention under different k values ​​and different weights.

[0059] Figure 10 A schematic diagram comparing the service link overhead and topology convergence overhead of the improved TNCP protocol and the OpenFlow protocol provided by this invention.

[0060] Figure 11 The improved TNCP protocol provided by this invention is illustrated in the protocol overhead diagram under different numbers of nodes;

[0061] Figure 12 This is a schematic diagram of the data forwarding performance test results provided by the present invention. Detailed Implementation

[0062] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0063] This invention provides a software-defined tactical communication network system and communication method, which will be described below.

[0064] Example 1

[0065] A specific embodiment of the present invention discloses a software-defined tactical communication network system, including multiple network nodes running controller software in a tactical network; wherein, a number of the network nodes are deployed in a distributed manner through an autonomous routing protocol, and the remaining network nodes form a number of local area networks, each local area network being controlled by a local controller;

[0066] The network system includes a data plane, a local control plane, and a global control plane;

[0067] The data plane includes network nodes in each local area network (LAN) and distributed network nodes; the network nodes in the LAN interact with the local controller of their respective LAN via an improved TNCP protocol; the distributed network nodes perform autonomous routing and networking based on the TRP protocol and interact with the selected local controller via the improved TNCP protocol; all network nodes in the data plane perform matching and forwarding through multi-level flow tables based on feature recognition.

[0068] The local control plane includes multiple local controllers, which interact with each other via an east-west protocol;

[0069] The global control plane includes a top-level controller, which interacts with each local controller in the local control plane to perform global control of the network.

[0070] The dual-plane control network architecture proposed in this embodiment adopts a logically centralized but physically distributed tactical communication network architecture. It obtains a global resource view through centralized management, thereby achieving flexible scheduling and efficient utilization of resources. This drives the tactical communication network towards greater intelligence and resilience. Distributed autonomous routing networking provides basic communication capabilities for the tactical communication network, enhancing its resilience. By dividing the network architecture into a data plane, a local control plane, and a global control plane, and implementing a modular design, flexible control and scheduling at different levels are allowed. For example, the local controller and the global controller interact through an east-west protocol, enabling independent management and adjustment of the local and global networks. This avoids the tight coupling of traditional architectures, greatly enhancing network flexibility and making it easier to adjust and optimize to meet changing needs in different scenarios.

[0071] As a specific example, such as Figure 1 As shown, Figure 1 This paper demonstrates a specific architecture for a software-defined tactical communication network. Figure 1 It consists of 11 network nodes, with nodes 1 through 11 all running the TRP (Tactical Routing Protocol). To enhance the controller's resilience, each node runs controller software.

[0072] During network operation, the controllers of nodes 4, 8, 9, 10, and 11 are in operation; at the same time, nodes 1, 2, and 3 are controlled by the controller on node 4, and nodes 5, 6, and 7 are controlled by the controller on node 8. Due to limited bandwidth resources, nodes 9, 10, and 11 are deployed in a distributed manner through an autonomous routing protocol.

[0073] In the network architecture described above, node 4, along with nodes 1, 2, and 3, forms the first local area network (LAN), and node 8, along with nodes 5, 6, and 7, forms the second LAN. Node 4 is the local controller in the first LAN, and node 8 is the local controller in the second LAN. The controllers on nodes 4 and 8 interact via an east-west protocol. The top-level controller controls nodes 4 and 8; in this scenario, the top-level controller runs on node 4. The controlled nodes 1, 2, and 3 in the first LAN, the controlled nodes 5, 6, and 7 in the second LAN, and the distributed nodes 9, 10, and 11 form the data plane; the local controller nodes 4 and 8 form the local control plane, while the top-level controller running on node 4 can be considered the global control plane.

[0074] Figure 1In this architecture, each node's switching equipment primarily runs the lightweight southbound interface protocol TNCP (Tactical Networking Control Protocol) and the lightweight autonomous routing protocol TRP (Tactical Routing Protocol). TRP enables autonomous routing networking between nodes and reports link resource information to the controller. Controlled network nodes forward data through multi-level flow table matching based on service characteristics. The local controller interacts with each controlled node via TNCP. Controllers interact with each other via eastbound protocols to form a top-level controller, thereby achieving overall control of the entire network.

[0075] By simplifying the traditional distributed network architecture and retaining only autonomous routing, the basic communication service functions of the distributed network are guaranteed. On this basis, a software-defined control plane is added to improve the service quality of the tactical communication network through centralized control.

[0076] like Figure 2 As shown, Figure 2 This demonstrates the main functions of switches and controllers in a network node. In a dual-control-plane network architecture, the controller is primarily responsible for ensuring basic communication services, including tactical protocol control, tactical routing control, and application management. The switches perform multi-level flow table matching and forwarding based on service characteristics and data traffic conditions.

[0077] Example 2

[0078] This invention also provides a software-defined tactical communication network method, applied to the software-defined tactical communication network system described above, comprising:

[0079] The local controller in the local area network establishes a connection with the switch of the directly connected controlled network node;

[0080] The local controller queries the routing table of the directly connected switch to obtain the IP addresses of each controlled network node switch in the local area network, and interacts with each controlled switch through the improved TNCP protocol based on the IP addresses.

[0081] When the network nodes are deployed in a distributed manner, all messages sent by the switches of the network nodes are forwarded by the switches directly connected to the local controller.

[0082] Both the controlled switches and the distributed switches in the local area network are matched and forwarded using multi-level flow tables based on feature recognition.

[0083] To address the characteristics of battlefield communication networks, such as narrow bandwidth, weak connectivity, and high latency, an improved TNCP protocol was developed as a specific implementation by optimizing the commercial southbound interface protocol OpenFlow. This improved protocol is suitable for tactical environments and serves as a "tactical networking control protocol." Specifically, the protocol was optimized in the following three aspects:

[0084] 1. By adopting a custom message format, the protocol content is simplified while ensuring minimum protocol scalability, reducing protocol overhead in channel transmission. At the same time, a reliable transmission mechanism is added to the protocol, enabling it to adapt to transmission in harsh channels in tactical communication networks.

[0085] The protocol integrates policy control, network management, resource awareness, and service registration to meet the flexible networking, resource utilization, and functional service requirements of SDTN. By proposing a security hardening and integration design concept, data compression and data transmission technologies are applied to the protocol design, effectively improving the applicability and inherent security of the protocol in narrowband tactical communication networks.

[0086] 2. Optimize the Packet_In message mechanism in OpenFlow. When a service packet fails to match in the flow table, first identify the message's characteristics. Subsequent packets adapted to the same service flow will not trigger the Packet_In message again. For such packets, the switch performs unified local routing processing, using the forwarding flow table generated by the routing protocol to achieve reachability forwarding for the service packet. This ensures that packets within the same service flow only complete the Packet_In message triggering process once, effectively reducing the impact of network latency on the control channel under the Packet_In guarantee mechanism.

[0087] 3. A bit-based formatted arrangement is adopted to minimize data gaps in the data structure, thereby achieving optimal design at the protocol body data structure level. By aggregating short messages, effective information is arranged and transmitted in a more compact manner, eliminating redundant information during message transmission, including duplicate message headers and effective data gaps. Furthermore, a dictionary encoding algorithm based on LZJH (Lempel-Ziv-Jeff-Heath) is used to further compress the protocol body data packets, further reducing protocol bandwidth overhead.

[0088] like Figure 3 As shown, Figure 3The improved TNCP protocol framework is demonstrated. After the controller connects to the network, it obtains the device IP address of the directly connected switch by querying the routing table of the switch. It then interacts with the switches in the network via TNCP using this IP address. In a distributed deployment, the controller directly interacts with the connected switches via TNCP, and TNCP messages from other switches in the network are forwarded through this switch.

[0089] Considering the needs of service-oriented application architecture design, channel device management in tactical communication networks, and network resource collection and management under centralized control logic, resource awareness and device management functions are added to the protocol based on the original flow rule control function. The link overhead is reduced through lightweight design, and the protocol is carried by UDP protocol so that it can better adapt to the tactical communication environment.

[0090] As a specific example, such as Figure 4 As shown, Figure 4 The TNCP protocol interaction process is demonstrated, which is mainly divided into two parts. The first part is the process of establishing a connection between the TNCP on the controller side and the switch side, and the second part is the process of the switch receiving management and control from the controller.

[0091] The switch uses the link-state-based TRP to build a control channel with the controller, realizing control links between nodes within the controller domain. The process includes service connection establishment, connection maintenance, and control message interaction. When the controller successfully connects to the switch, it uses the IP address of the directly connected switch to access services. It obtains resource information of the directly connected switch and the topology information of the entire network through resource query messages, manages the flow table matching rules of the directly connected switch through rule management messages, and manages the switch through device management messages.

[0092] The controller communicates with other switches based on the IP addresses of each switch in the network topology information obtained from the directly connected switches. When the network topology information of the directly connected switches is unavailable, the device IP addresses of each switch in the current network can be obtained from the switch routing table.

[0093] In a preferred embodiment, during the connection establishment process between the controller and the switch, the status of the local controller includes an unconnected state, a handshake successful state, and an access state.

[0094] The local controller is initially in a disconnected state;

[0095] When not connected, a handshake request message is sent to the switch, and the local controller starts a timer to wait for the handshake response message;

[0096] If the local controller receives a handshake response message within the first preset time, it enters the handshake success state and sends a service registration request message to the switch; if it receives a registration success response message within the second preset time, it enters the access state.

[0097] In the access state, the local controller successfully registers with the switch and is in a connection-maintaining state. When the connection maintenance fails, the local controller reverts to the disconnected state.

[0098] Figure 5 The diagram illustrates the connection state transitions of the controller. When the handshake response message exceeds the first preset time, the controller's TNCP module resets the timer and restarts the handshake process; if the handshake is successful but no registration success response message is received within the second preset time, or a registration failure response message is received before the timeout, the controller enters a disconnected state.

[0099] In the access state, the controller side has successfully registered with the switch and is in a connection maintenance state. When the connection maintenance fails, the module returns to the disconnected state. Only in the access state will the controller side forward data from the upper-layer service function module to the TNCP module.

[0100] Example 3

[0101] To better adapt to the dual-control-plane architecture in tactical scenarios, we optimized the traditional link-state routing protocol. Furthermore, for different tactical communication scenarios, and to optimize the utilization of resource information such as link rate, link latency, bandwidth utilization, load, and the number of congested links, we introduced a particle swarm optimization algorithm based on a biological genetic algorithm. The specific method is as follows:

[0102] Step S1: Define the fitness function

[0103] First, define a fitness function to evaluate the performance of each particle (switch), including the following factors: link rate. Link latency Link bandwidth Bandwidth utilization ,load Number of congested links and number of unused links The fitness function aims to maximize link speed, minimize link latency, maximize link bandwidth, maximize bandwidth utilization, minimize load, minimize the number of congested links, and minimize the number of unused links, in order to optimize the performance of the tactical network. It can be defined as:

[0104] (1)

[0105] in, Represents particles fitness value, Represents particles Link rate metrics Represents particles Link latency metrics Represents particles Link bandwidth metrics, Represents particles Bandwidth utilization metrics Represents particles The load index, Represents particles The number of congested links. Represents particles The number of unused links, , , , , , and These are the weights of the corresponding indicators, which can be adjusted according to actual needs to reflect the importance of different indicators in the fitness function.

[0106] Step S2: Define the state space

[0107] The routing strategy of the switch is abstracted into a state space, where each state represents a possible routing option.

[0108] Step S3: Initialize the particle swarm

[0109] Each particle represents a candidate solution (i.e., a route selection). The swarm consists of multiple particles. Each particle is randomly initialized with a set of states, which correspond to different route selections. Each particle has a velocity vector used to update its position.

[0110] Step S4: Update particle position and velocity

[0111] The fitness value of each particle is calculated according to the fitness function defined in formula (1), and compared with its individual best position Pbest and global best position Gbest.

[0112] The velocity and position of each particle are updated, and the velocity update formula is as follows:

[0113] (2)

[0114] The position update formula is as follows:

[0115] (3)

[0116] in, It is the first Particles in time The speed of time, It is a constraint factor on speed. It is the first Particles in time Location, It is the first The best position found so far for each particle This is the best location the entire group has found so far. It is inertial weight. and It is a learning factor used to control the weights of individual and global experiences. It is a random number between [0, 1].

[0117] Step S5: Update the global optimum

[0118] Update the global optimum based on the fitness value of each particle. .

[0119] Step S6: Repeat the iteration

[0120] Check whether the predetermined number of iterations or the desired fitness threshold has been met. If the termination condition has not been met, repeat steps S4 to S6.

[0121] Step S7: Optimal Route Selection

[0122] The optimal route selection for each switch is determined based on the final global optimum.

[0123] By employing the particle swarm optimization algorithm and flexibly adjusting the weights of the fitness function, the algorithm can make reasonable optimization decisions for different requirements (such as link speed, latency, bandwidth, etc.), ultimately improving the stability, efficiency, and reliability of the entire network. It can provide multi-objective optimized routing in complex and ever-changing tactical network environments, ensuring optimal network performance.

[0124] Example 4

[0125] Tactical network communication services are diverse and require more flexible forwarding mechanisms. We designed a multi-service programmable forwarding system based on SOPC (System on a Programmable Chip) switch hardware platform to support flexible programmable forwarding of services. We adopted TCAM (Ternary Content Addressable Memory) and SRAM (Static Random-Access Memory) to implement multi-level flow table matching forwarding based on service categories.

[0126] In a preferred embodiment, the matching and forwarding via a multi-level flow table based on feature recognition includes:

[0127] The data stream is categorized based on network data and network functions, and the data is forwarded based on the flow table corresponding to the matching domain of the categorized data.

[0128] Matching and forwarding is performed based on the switch's cache status and service priority.

[0129] The following is a detailed explanation of the two matching and forwarding processes:

[0130] (1) Performing category matching on data flows based on network data and network functions, and forwarding data based on the flow table corresponding to the matching domain of the category, including:

[0131] Calculate the compatibility between each pair of different matching fields contained in each matching category in the single-level flow table stored by the switch;

[0132] The matching domain set is divided according to the compatibility, so that the sum of the compatibility of the matching domains located in the same set is maximized;

[0133] The single-level flow table is converted into a multi-level flow table based on the partitioned set, and the classification processing of different data streams is realized based on the multi-level flow table structure.

[0134] Specifically, by default, switches store all routing and flow tables in SRAM. Based on data traffic conditions, the switch distributes the forwarding tables from SRAM to the TCAM in the form of flow tables for flow forwarding. Due to the diverse and highly dynamic nature of tactical network functions, different matching rules are required for different data flows. For example, Layer 2 routing matches MAC values, while Layer 3 forwarding matches IP values. Therefore, different network data and network functions require different types of matching for data packets.

[0135] Assuming for the flow table This includes a series of matching categories, flow tables. The set of included matching categories express, There are a total of Each table item category, and each category It contains different matching fields.

[0136] Considering that different matching fields may exist in different categories of entries, storing matching fields belonging to the same entry category in different levels will increase the flow table forwarding processing latency. For two different matching fields... The compatibility between the two is defined as:

[0137] (4)

[0138] for and Coexisting entries This is the total list of entries.

[0139] For the same matching domain, compatibility is defined as:

[0140] (5)

[0141] To contain only the matching field For each partitioning method of a single-level flow table, the compatibility between pairwise matching fields is calculated. The goal is to maximize the sum of compatibility between matching fields located in the same set, i.e., to partition matching fields of the same entry type into the same set as much as possible. The optimization function is:

[0142] (6)

[0143] in, , , , , ,

[0144] Divide the K matching domains of a single-level flow table into m subsets. As the basis for dividing each flow table, to simplify the search space, we group matching domains of the same category into the same set, maximizing overall compatibility.

[0145] By dividing a single-level flow table into a multi-level flow table structure, it is possible to classify and process different data flows to a certain extent.

[0146] As shown in Figure 6 Figure 6 The single-level flow table is divided into two levels. The first-level flow table mainly stores general rules and performs routine matching to achieve fast forwarding of general services. The second-level flow table mainly stores flow tables for specific service categories. It is divided according to different service categories. After the service category matching is completed, the flow table matching and forwarding for specific service domains are performed.

[0147] (2) Match and forward according to the switch's buffer status and service priority level.

[0148] Tactical environments are characterized by diverse service requirements, with typical data including general packets and voice data. This necessitates prioritizing different services within the tactical environment. Matching and forwarding should be performed based on the switch's buffer status and service priority levels, including:

[0149] Different service types are pre-prioritized to obtain priority queues; each priority queue has a preset congestion threshold value, and the corresponding cache space is allocated according to the congestion threshold value;

[0150] Determine the service type of the real-time data stream, and cache the data stream in the cache space of the corresponding priority queue for data forwarding based on the service type of the data stream.

[0151] In the switch's buffer queue, assuming that high-priority and low-priority packets share a storage space of B, the arrival rate of packets in priority queue i is... The Poisson process, the service time follows the parameter is The exponential distribution, with each priority queue using weights. The importance of the queue is indicated by the length of each priority queue. Let the load of the i-th priority queue be denoted as . The total load of this output port is denoted as Let k represent the ratio of the load on the low-priority queue to the load on the high-priority queue, i.e. The voice service priority queue is The priority queue for ordinary message services is ,but .

[0152] By default, switches prioritize the transmission of high-priority packets. If the switch does not allocate its buffer space appropriately, data of different priorities will not receive optimal service during transmission. To solve this problem, appropriate threshold values ​​T need to be set for different priority queues.

[0153] In some embodiments, matching and forwarding based on the switch's cache state and service priority level further includes:

[0154] The system state transition probability is determined based on the packet arrival rate, packet service rate, and congestion threshold of the priority queue.

[0155] Calculate the overall blocking probability of the system based on the system state transition probability;

[0156] The weights and congestion thresholds for each priority level are optimized based on the overall congestion probability.

[0157] like Figure 7 As shown, if Let i be the group arrival rate of priority queue i. Let i be the packet service rate of priority queue i. Then the state transition diagram of the system is as follows: Figure 7 As shown. According to Figure 7 The transition probabilities of the system from state n to state n+1 and from state n to state n-1 are respectively:

[0158] (7)

[0159] (8)

[0160] According to the global equilibrium equation, we can obtain:

[0161] (9)

[0162] definition , Solving the global equilibrium equations yields:

[0163] (10)

[0164] because Then it can be calculated for:

[0165] (11)

[0166] use To indicate the probability of high-priority blocking, use Let B represent the low-priority blocking probability. Then, the high-priority blocking probability is the probability that the queue length reaches B, and the low-priority blocking probability is the probability that the length reaches T. Combining formulas (10) and (11), we can obtain:

[0167] (12)

[0168] (13)

[0169] According to formulas (12) and (13), the overall system blocking rate X can be obtained as:

[0170] (14)

[0171] The maximum buffer space in the switch system is 20, i.e., B=20. The total system load is set here. =0.9, high priority weight =2, low priority weight The relationship between the threshold value and the blocking probability when k = 1, k = 2, 1.25, 0.8, and 0.5 is as follows: Figure 8 As shown.

[0172] from Figure 8 It can be seen that under different load conditions, the blocking probability is lowest when the threshold value is controlled at around 17. Priority weights are set for different traffic volumes. =1, When K = 2~5, the blocking status of the system queue when K is set to 2 / 1.25 / 0.8 / 0.5 respectively is as follows: Figure 9 As shown.

[0173] from Figure 9 It can be seen that regardless of the changes in system traffic or the weights of high and low priorities, when the system cache size is 20, the system performance is highest when the optimal threshold value for the low priority queue is between 16 and 18. In the test system, we took the middle value of 17 as the threshold value for the low priority of the switch.

[0174] To verify the effectiveness of this invention, we tested the aforementioned dual-control plane architecture using an OPNET tactical simulation platform and a wireless network consisting of 12 physical nodes. The controller, physical nodes, and simulation communicated via the TNCP protocol, while the physical nodes communicated via the TRP protocol. Based on the test environment, we verified the performance of the improved TNCP and TRP protocols. The comparison results between the improved TNCP protocol and the OpenFlow protocol are shown below. Figure 10 As shown. From Figure 10 It can be seen that as the number of nodes in the tactical communication network increases, the service connection overhead and topology convergence overhead of our designed TNCP protocol are much lower than those of the OpenFlow protocol.

[0175] Figure 11 The graph shows the total overhead of the tactical communication system from startup to topology convergence. As can be seen, while both TNCP and OpenFlow protocols exhibit exponential overhead growth with the number of nodes, TNCP's overhead growth is significantly lower than OpenFlow's. Specifically, with 100 nodes, TNCP's overhead is 198KB, while OpenFlow's overhead reaches 1.7MB. Considering the bandwidth characteristics of the tactical environment, OpenFlow can no longer meet the networking requirements of tactical communication networks with 100 nodes, while TNCP performs better in meeting these requirements.

[0176] Based on a dual control plane architecture, we verified the performance of the proposed TRP protocol and compared it with the traditional link-state protocols OSPF and ISIS. For tactical communication scenarios with wireless networking requirements, TRP is optimized according to the network topology. During the dynamic convergence process of the network, the overhead of the TRP protocol is about 400 Bps, the overhead of the OSPF protocol is about 900 Bps, and the overhead of the ISIS protocol is about 2500 Bps. The TRP protocol has better performance advantages.

[0177] Under the dual control plane architecture, we tested the performance of the proposed multi-level flow table matching and multi-priority queue data forwarding. The results are as follows: Figure 12 As shown.

[0178] This application proposes a software-defined tactical communication network system and communication method. Tactical communication networks are characterized by high dynamism, low bandwidth, and high survivability requirements. By designing a software-defined tactical network architecture with dual control planes, when the tactical network communication conditions are good, the network can be networked using the software-defined tactical communication network. Based on a global view, comprehensive resource scheduling is achieved, providing users with a more flexible networking method. When the tactical network communication conditions are poor (low bandwidth, weak connection, and other tactical terminal wireless scenarios), the network automatically switches to autonomous routing networking. Since the switching equipment uniformly converts the flow table and routing table into a category-based flow table and distributes it to the chip, the switching process does not affect the use of the network, making the network more intelligent and highly adaptable.

[0179] Based on the proposed dual-control-plane software-defined tactical network architecture, we designed an improved lightweight southbound interface protocol, TNCP. Compared to the traditional OpenFlow protocol, test results show an 87% performance improvement, effectively alleviating the problem of high bandwidth consumption under centralized control. Simultaneously, we designed a link-state-based autonomous routing protocol, TRP, which collects latency, bandwidth, and topology data from each tactical communication device in the system and uploads it to the controller, providing more comprehensive information for path calculation and intelligent scheduling. The solution employs a particle swarm optimization algorithm based on genetic algorithms and the LZJH compression algorithm, reducing the protocol overhead of TRP. Compared to traditional OSPF and IS-IS protocols, it reduces bandwidth consumption by nearly 40%, making it more suitable for tactical communication network scenarios. The solution expands the system's forwarding methods by designing category-based multi-level flow table matching and multi-priority queue-based data forwarding, resulting in lower packet loss rate and lower forwarding latency, suitable for multi-priority services in tactical communication scenarios. This invention comprehensively considers the potential problems of applying SDN to tactical scenarios, proposes corresponding solutions based on the proposed architecture, and verifies the effectiveness of the solutions through physical implementation.

[0180] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A software-defined tactical communication network system, characterized in that, It includes multiple network nodes running controller software in a tactical network; wherein, a number of the network nodes are deployed in a distributed manner through an autonomous routing protocol, and the remaining network nodes form several local area networks, each of which is controlled by a local controller; The network system includes a data plane, a local control plane, and a global control plane; The data plane includes network nodes in each local area network (LAN) and distributed network nodes; the network nodes in the LAN interact with the local controller of their respective LAN via an improved TNCP protocol; the distributed network nodes perform autonomous routing and networking based on the TRP protocol and interact with the selected local controller via the improved TNCP protocol; all network nodes in the data plane perform matching and forwarding through multi-level flow tables based on feature recognition. The local control plane includes multiple local controllers, which interact with each other via an east-west protocol; The global control plane includes a top-level controller, which interacts with each local controller in the local control plane to perform global control of the network.

2. A software-defined tactical communication network method, applied to the software-defined tactical communication network system of claim 1, characterized in that, include: The local controller in the local area network establishes a connection with the switch of the directly connected network node; The local controller queries the routing table of the switch to obtain the IP addresses of each network node switch in the local area network, and interacts with each switch through the improved TNCP protocol based on the IP addresses. When the network nodes are deployed in a distributed manner, all messages sent by the switches of the network nodes are forwarded by the switches directly connected to the local controller. Both the switches in the local area network and the distributed switches use multi-level flow tables based on feature recognition for matching and forwarding.

3. The software-defined tactical communication network communication method according to claim 2, characterized in that, The method further includes: In a local area network, switches use the link-state-based TRP protocol to establish a control channel with the local controller; When the local controller is in the access state, the control link between the local controller and each switch in the local area network is completed.

4. The software-defined tactical communication network communication method according to claim 3, characterized in that, During the connection establishment process between the local controller and the switch, the state of the local controller includes a disconnected state, a handshake successful state, and an access state. The local controller is initially in a disconnected state; When not connected, a handshake request message is sent to the switch, and the local controller starts a timer to wait for the handshake response message; If the local controller receives a handshake response message within the first preset time, it enters the handshake success state and sends a service registration request message to the switch. If a successful registration response is received within the second preset time, the system will enter the access state. In the access state, the local controller successfully registers with the switch and is in a connection-maintaining state. When the connection maintenance fails, the local controller reverts to the disconnected state.

5. The software-defined tactical communication network communication method according to claim 2, characterized in that, The improved TNCP protocol includes a particle swarm optimization (PSO)-based routing method for each switch to select the optimal routing path in the tactical network. The routing method includes: The fitness function for evaluating the performance of each switch is determined based on link connectivity parameters, link bandwidth parameters, and link congestion parameters. Initialize multiple switch candidate states representing routing selection, each candidate state having a velocity vector; The fitness value corresponding to each candidate state is calculated based on the fitness function, and the individual optimal position and global optimal position of each candidate state are obtained. The velocity vector of each candidate state is iteratively updated based on the individual optimal position and the global optimal position, and each candidate state is updated according to the updated velocity vector; The fitness value of each candidate state is calculated based on the updated candidate states, and the global optimal position is updated according to the updated candidate states. The update is completed when the number of iterations reaches the preset number or the fitness value reaches the preset fitness threshold. The optimal route selection for each switch is determined based on the global optimal position at the time of the update.

6. The software-defined tactical communication network communication method according to claim 5, characterized in that, The link connection parameters include link rate and link latency; the link bandwidth parameters include link bandwidth and bandwidth utilization; the link congestion parameters include load, number of congested links, and number of unused links. The fitness function for evaluating the performance of each switch is determined based on link connectivity parameters, link bandwidth parameters, and link congestion parameters, including: The objectives of the fitness function are: to maximize link speed, minimize link latency, maximize link bandwidth, maximize bandwidth utilization, minimize load, minimize the number of congested links, and minimize the number of unused links.

7. The software-defined tactical communication network communication method according to claim 2, characterized in that, The matching and forwarding via a multi-level flow table based on feature recognition includes: The data stream is categorized based on network data and network functions, and the data is forwarded based on the flow table corresponding to the matching domain of the categorized data. Matching and forwarding is performed based on the switch's cache status and service priority.

8. The software-defined tactical communication network communication method according to claim 7, characterized in that, Data streams are categorized based on network data and network functions, and data is forwarded based on the flow table corresponding to the categorized matching domain, including: Calculate the compatibility between each pair of different matching fields contained in each matching category in the single-level flow table stored by the switch; The matching domain set is divided according to the compatibility, so that the sum of the compatibility of the matching domains located in the same set is maximized; The single-level flow table is converted into a multi-level flow table based on the partitioned set, and the classification processing of different data streams is realized based on the multi-level flow table structure.

9. The software-defined tactical communication network communication method according to claim 7, characterized in that, Matching and forwarding based on the switch's cache status and service priority level, including: Different service types are pre-prioritized to obtain priority queues; each priority queue has a preset congestion threshold value, and the corresponding cache space is allocated according to the congestion threshold value; Determine the service type of the real-time data stream, and cache the data stream into the cache space of the corresponding priority queue for data forwarding based on the service type of the data stream.

10. The software-defined tactical communication network communication method according to claim 9, characterized in that, Matching and forwarding based on the switch's cache status and service priority also includes: The system state transition probability is determined based on the packet arrival rate, packet service rate, and congestion threshold of the priority queue. Calculate the overall blocking probability of the system based on the system state transition probability; The weights and congestion thresholds for each priority level are optimized based on the overall congestion probability.

Images