Dynamic traffic scheduling optimization method for large-scale low-orbit satellite network

Abstract

The invention provides a dynamic traffic scheduling optimization method for a large-scale low-orbit satellite network, and relates to the technical field of network optimization, and the method comprises the following steps: predicting the load state of a satellite node in a next time period according to the load state of the satellite node in a previous time period; when the load state is a high-load state, executing a high-load flow scheduling optimization method; when the load state is a low load state, the queue of the satellite node output port in the next time period is configured to comprise a high priority queue and a low priority queue; in any time slot of the next time period, the service data in the high-priority queue are firstly sent, and after the service data in the high-priority queue are sent, the service data in the low-priority queue are sent, so that dynamic traffic scheduling optimization is completed; and when the next time slot starts, the high-priority queue and the low-priority queue are respectively converted into the low-priority queue and the high-priority queue.

Description

Technical Field

[0001] The present invention relates to the technical field of network optimization, and in particular to a dynamic traffic scheduling optimization method for large-scale low-Earth orbit satellite networks. Background Art

[0002] The spatial topology of large-scale low-Earth orbit satellite networks presents periodic time-variation, and new time-sensitive traffic flows are continuously injected into the network, and the loads of the out-ports of each satellite node in the network are constantly changing.

[0003] When the load is low, if the queue length of the out-port of the satellite node is too long, it will lead to too long queuing delay, and the end-to-end delay of the traffic flow exceeds the maximum service time, and the traffic cannot be scheduled. It is impossible to ensure that the time-sensitive service data packets in the two queues have the opportunity to be forwarded within one time slot.

[0004] When the load is high, if the queue length for transmitting time-sensitive traffic flows at the out-port of the satellite node is too short and the number of queues is too small, congestion will occur at the out-port, resulting in packet loss.

[0005] In the prior art, such as the Chinese invention patent with the publication number CN118381545A, a queue scheduling optimization method for large-scale low-Earth orbit constellations based on the particle swarm algorithm is disclosed. When the packet loss rate of time-sensitive services reaches the threshold, the queue length and the number of queues for transmitting time-sensitive traffic flows of each node are recalculated, so as to realize the dynamic optimization and adjustment of the load of each node according to the real-time state of the services in the large-scale low-Earth orbit constellation. However, the above comparative document works in a high-load state, and there is no dynamic traffic scheduling optimization method for low-load situations. Summary of the Invention

[0006] In order to solve the technical problems existing in the above prior art, the purpose of the present invention is to provide a dynamic traffic scheduling optimization method for large-scale low-Earth orbit satellite networks, which can reduce the delay at low load to ensure that time-sensitive service data packets are sent out within one time slot; at high load, reduce the underlying hardware resource overhead.

[0007] To achieve the above invention purpose, the present invention provides a dynamic traffic scheduling optimization method for large-scale low-Earth orbit satellite networks, and the steps of the method are as follows:

[0008] Predict the load status of the satellite node in the next time period according to the load status of the satellite node in the previous time period;

[0009] When the load status is a low-load status, execute the low-load traffic scheduling optimization method;

[0010] The low-load status is:

[0011] When the queue lengths and the number of queues of the satellite node's outgoing ports in the next time period are the same as those in the previous time period, in at least one time slot of the next time period, the state where the lowest-priority queue does not receive service data;

[0012] The low-load traffic scheduling optimization method is as follows:

[0013] Configure the queues of the satellite node's outgoing ports in the next time period to include a high-priority queue and a low-priority queue;

[0014] In any time slot of the next time period, first send the service data in the high-priority queue. When the service data in the high-priority queue is sent out, then send the service data in the low-priority queue to complete the dynamic traffic scheduling optimization; and when the next time slot starts, convert the high-priority queue and the low-priority queue in the previous time slot into a low-priority queue and a high-priority queue respectively.

[0015] According to one aspect of the present invention, it further includes:

[0016] When sending the service data in the high-priority queue, receive the service data transmitted by other satellite nodes into the low-priority queue.

[0017] According to one aspect of the present invention, the queue lengths of the high-priority queue and the first-priority queue are both l1;

[0018] l1 = MTU·u1

[0019] Wherein, MTU represents the maximum transmission unit for sending service data in a large-scale low-Earth orbit satellite network, with the unit of byte; u1 represents the number of MTUs included in the high-priority queue or the first-priority queue, and u1 = 1000.

[0020] According to one aspect of the present invention, the load state of the satellite node in the next time period is predicted through a long short-term memory network LSTM.

[0021] According to one aspect of the present invention, the number of neurons in the first layer of LSTM is 64, and the number of neurons in the second layer of LSTM is 128; the regularization coefficient is 0.7, the loss function is the mean square error MSE loss function, and the optimizer is the Adam optimizer; the batch size is 128, and the number of iterations is 150.

[0022] The present invention also provides a dynamic traffic scheduling optimization method for a large-scale low-Earth orbit satellite network, and the method steps are as follows:

[0023] Predict the load state of the satellite node in the next time period according to the load state of the satellite node in the previous time period;

[0024] When the load state is a high - load state, execute the high - load traffic scheduling optimization method;

[0025] The high - load state is:

[0026] In the case that the queue length and the number of queues at the satellite node's outgoing port in the next time period are the same as those in the previous time period, in at least one time slot in the next time period, the state where the service data received by the satellite node exceeds the sum of the storage spaces of all non - highest - priority queues;

[0027] The high - load traffic scheduling optimization method is:

[0028] Obtain the maximum load of the satellite node in the next time period according to the service data traffic received by the satellite node in the next time period, and calculate the queue length L b and the number of queues N b ;

[0029] Configure the queues at the satellite node's outgoing port in the next time period as queue length L b and the number of queues N b , and then send the service data for dynamic traffic scheduling optimization.

[0030] According to one aspect of the present invention, calculate the queue length L b and the number of queues N b that meet the maximum load demand of the satellite node in the next time period, specifically including:

[0031] Take the maximum load of the satellite node in the next time period as the optimization target, and use the Particle Swarm Optimization (PSO) algorithm to calculate the queue length L b and the number of queues N b .

[0032] According to one aspect of the present invention, predict the load state of the satellite node in the next time period through a Long Short - Term Memory (LSTM) network.

[0033] According to one aspect of the present invention, the number of neurons in the first layer of the LSTM is 64, and the number of neurons in the second layer of the LSTM is 128; the regularization coefficient is 0.7, the loss function is the Mean Squared Error (MSE) loss function, and the optimizer is the Adam optimizer; the batch size is 128, and the number of iterations is 150.

[0034] According to one aspect of the present invention, L b =MTU·U b , 1≤U b ≤1000; and 2≤N b ≤7.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The present invention proposes a dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks:

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The present invention proposes a dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks:

[0039] 1. In the case of low load, a time-sensitive traffic scheduling and forwarding mechanism with cyclic queue priority is adopted to cycle the priorities of two queues and forward service data, ensuring that the service data in both queues has the opportunity to be forwarded within one time slot, breaking through the bandwidth limitation.

[0040] 2. In the case of high load, the optimal queue length and the number of queues of the satellite node's out-port are calculated, which can minimize the occupation of underlying hardware resources on the premise of ensuring no packet loss due to congestion, and ensure that the implementation difficulty and cost will not be too high. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0042] Figure 1 Schematically showing the flowchart of a dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to an embodiment of the present invention;

[0043] Figure 2 Schematically showing the schematic diagram of the time-sensitive traffic scheduling and forwarding principle of cyclic queue priority in the low-load state in a dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to an embodiment of the present invention;

[0044] Figure 3 Schematically showing the flowchart of a queue optimization method based on the particle swarm algorithm in a dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] The description of the embodiments of this specification should be combined with the corresponding drawings, which should be part of the complete specification. In the drawings, the shape or thickness of the embodiments may be enlarged, and simplified or convenient markings may be used. Furthermore, the parts of each structure in the drawings will be described separately. It should be noted that the elements not shown or described in words in the drawings are in forms known to those of ordinary skill in the art.

[0046] Any reference to directions and orientations in the description of the embodiments herein is for convenience of description only and should not be construed as any limitation on the scope of protection of the present invention. The following description of the preferred embodiments involves combinations of features that may exist independently or in combination. The present invention is not particularly limited to the preferred embodiments. The scope of the present invention is defined by the claims.

[0047] As Figure 1 shown, a dynamic traffic scheduling optimization method for a large-scale low-Earth orbit satellite network of the present invention is as follows:

[0048] S11. Predict the load status of the satellite node in the next time period according to the load status of the satellite node in the previous time period;

[0049] S12. When the load status is a low-load status, execute the low-load traffic scheduling optimization method;

[0050] The low-load status is:

[0051] In the case that the queue length and the number of queues at the output port of the satellite node in the next time period are the same as those in the previous time period, in at least one time slot in the next time period, the state where no service data is received in the lowest-priority queue;

[0052] The low-load traffic scheduling optimization method is:

[0053] S111. Configure the queue at the output port of the satellite node in the next time period to include a high-priority queue and a low-priority queue;

[0054] S112. In any time slot in the next time period, first send the service data in the high-priority queue. When the service data in the high-priority queue is sent out, then send the service data in the low-priority queue to complete the dynamic traffic scheduling optimization; and when the next time slot starts, convert the high-priority queue and the low-priority queue in the previous time slot into a low-priority queue and a high-priority queue respectively.

[0055] In this embodiment, the service data sent (forwarded) by the satellite node is a time-sensitive traffic flow. The dynamic traffic scheduling optimization method is used to optimize the queue length and the number of queues at the output port of the satellite node in the large-scale low-Earth orbit satellite network for transmitting time-sensitive traffic.

[0056] The analysis and optimization are divided into two steps:

[0057] First step: Use a prediction model to predict the load time-sensitive service traffic of large-scale LEO satellite network nodes. By inputting the statistical data of the time-sensitive service traffic of satellite nodes in each time slot in a period before the current moment into the prediction model, predict the time-sensitive service traffic of satellite nodes in each time slot in the next period. Judge the load state of the time-sensitive service traffic of satellite nodes, and specifically judge whether the load is in a low-load state or a high-load state according to the conditions satisfied by the load state.

[0058] The above prediction model can select a long short-term memory network LSTM, etc.

[0059] Second step: Select a suitable traffic scheduling mechanism according to the load prediction result, and optimize the length and quantity of the output port queues of each satellite node.

[0060] In this embodiment, the optimization is mainly for the low-load state. The low-load state considers the queue length and quantity configuration of satellite nodes for forwarding time-sensitive service traffic in a period before the current moment. If this configuration is maintained in the next period at the current moment and there are no time-sensitive service data packets entering the lowest-priority queue in a certain time slot, the time-sensitive service load state of the satellite node is in a low-load state.

[0061] Such as Figure 2 As shown, when the time-sensitive service load state of the satellite node is in a low-load state, a cyclic priority scheduling and forwarding mechanism is adopted to transmit time-sensitive traffic.

[0062] Specifically, in any time slot, there is a high-priority queue and a low-priority queue in the queue for forwarding time-sensitive service traffic. The satellite node preferentially forwards the time-sensitive data packets in the high-priority queue, and the arriving time-sensitive data packets only enter the low-priority queue in the current time slot. When all the time-sensitive data packets in the high-priority queue are forwarded, the low-priority queue starts to forward the time-sensitive data packets until the end of the current time slot.

[0063] This embodiment adopts a cyclic queue priority time-sensitive traffic scheduling and forwarding mechanism, cycles the priorities of the two forwarding queues, configures the gating lists of the two queues, and forwards time-sensitive service data packets to ensure that the time-sensitive service data packets in the two queues have the opportunity to be forwarded within a time slot, breaking through the bandwidth limit of time-sensitive traffic.

[0064] S112 also includes:

[0065] When sending the service data in the high-priority queue, receive the service data transmitted by other satellite nodes into the low-priority queue.

[0066] In S111, the queue lengths of the high-priority queue and the first-priority queue are both l1;

[0067] l1 = MTU · u1

[0068] wherein, MTU represents the maximum transmission unit for transmitting service data in the large-scale low-earth orbit satellite network, with the unit of byte; u1 represents the number of MTUs included in the high-priority queue or the first-priority queue, and u1 = 1000.

[0069] In this embodiment, the maximum transmission unit MTU of the time-sensitive traffic flow transmitted in the large-scale low-earth orbit satellite network is 1500 bytes. To ensure that all time-sensitive data packets can enter the queue completely, the queue length is set to l1 = MTU · u1. The number of queues is n1, and 2 ≤ n1 ≤ 7.

[0070] Specifically, the number of queues for the satellite node out-port to forward the time-sensitive traffic flow can be set to n1 = 2, and the queue length can be set to u1 = 1000.

[0071] In S11, the load status of the satellite node in the next time period is predicted through a long short-term memory network LSTM.

[0072] The LSTM algorithm can be used to predict the time-sensitive traffic flow of the satellite node in each time slot in the next time period. And the specific structure of the LSTM in S11 is as follows: the number of neurons in the first layer of the LSTM is 64, and the number of neurons in the second layer of the LSTM is 128; the regularization coefficient is 0.7, the loss function is the mean square error MSE loss function, and the optimizer is the Adam optimizer; the batch size is 128, and the number of iterations is 150.

[0073] Specifically, by inputting the statistical data of the time-sensitive traffic flow of the satellite node in each time slot in a period before the current moment into the neural network model, the time-sensitive traffic flow of the satellite node in each time slot in the next time period is predicted. After completing the prediction of the time-sensitive traffic flow of the satellite node in all time slots in the next time period, the load status is then judged. The main parameters and set values of the model are as follows in the table.

[0074] Main parameters Set value Number of neurons in LSTM layer 1 64 Number of neurons in LSTM layer 2 128 Regularization coefficient 0.7 Loss function MSE loss Optimizer Adam Batch size 128 Number of epochs 150

[0075] A dynamic traffic scheduling optimization method for a large-scale low-earth orbit satellite network according to the present invention has the following method steps:

[0076] S21. Predict the load status of the satellite node in the next time period according to the load status of the satellite node in the previous time period;

[0077] S22. When the load status is a high-load status, execute the high-load traffic scheduling optimization method;

[0078] The high-load status is:

[0079] When the queue lengths and queue numbers of the satellite node's output ports in the next time period are the same as those in the previous time period, in at least one time slot in the next time period, the state where the service data received by the satellite node exceeds the total storage space of all non-highest-priority queues;

[0080] The high-load traffic scheduling optimization method is as follows:

[0081] S211. Obtain the maximum load of the satellite node in the next time period according to the service data traffic received by the satellite node in the next time period, and calculate the queue length and queue number that meet the maximum load demand of the satellite node in the next time period;

[0082] S212. Configure the queues at the output ports of the satellite node in the next time period to the queue length and queue number, and then send the service data to perform dynamic traffic scheduling optimization.

[0083] In this embodiment, the dynamic traffic scheduling optimization method is used to optimize the queue lengths and numbers of the queues at the output ports of the satellite nodes for transmitting time-sensitive service traffic in a large-scale low-Earth orbit satellite network.

[0084] The analysis and optimization are divided into two steps:

[0085] The first step: Use a prediction model to predict the load time-sensitive service traffic of the nodes in a large-scale low-Earth orbit satellite network. By inputting the statistical data of the time-sensitive service traffic of the satellite node in each time slot in a period of time before the current moment into the prediction model, predict the time-sensitive service traffic of the satellite node in each time slot in the next period of time. Judge the load state of the time-sensitive service traffic of the satellite node, and specifically judge whether the load is in a low-load state or a high-load state according to the conditions satisfied by the load state.

[0086] The second step: Select a suitable traffic scheduling mechanism according to the load prediction result, and optimize the lengths and numbers of the queues at the output ports of each satellite node.

[0087] In this embodiment, the optimization is mainly for the high-load state. The high-load state considers the queue lengths and number configurations of the satellite nodes for forwarding time-sensitive service traffic in a period of time before the current moment. If this configuration is maintained in the next period of time at the current moment, and there are time-sensitive service data packets arriving in a certain time slot that exceed the total storage space of all non-highest-priority queues, then the time-sensitive service load state of the satellite node in the next period of time at the current moment is a high-load state.

[0088] The above prediction model can be selected as a long short-term memory network LSTM, etc.

[0089] According to the traffic prediction results of the prediction model, record the maximum load of time-sensitive service traffic within a period of time, and use an optimization algorithm to calculate the queue length that meets the requirements and the number of queues for transmitting time-sensitive service traffic. Then configure the gating control list GCL of the satellite node's output port with the above queue length and number of queues.

[0090] The above optimization algorithm can select the particle swarm optimization (PSO) algorithm, etc.

[0091] In this embodiment, when the time-sensitive service load state of the satellite node is in a high-load state, according to the traffic prediction results of the LSTM algorithm, record the maximum load L of each satellite node within each time slot. peak Taking this as the optimization goal, use the particle swarm optimization algorithm, that is, the PSO algorithm, to calculate the optimal solutions \(l\), \(m\) of the queue length \(l\) and the number of queues \(n\) for forwarding time-sensitive service traffic at the output port of each satellite node. b 、m b Set the maximum transmission unit MTU of the time-sensitive service flow transmitted in the large-scale low-earth orbit satellite network to 1500 bytes. To ensure that all data packets can enter the queue completely, the queue length \(l\) is set to \(l = MTU\cdot u\), where \(100\leq u\leq1000\). The number of queues is set to \(n\), then \(2\leq n\leq7\).

[0092] Round up the optimal solutions \(u\) b 、n b of the queue length and number obtained by the PSO algorithm to get Judge whether \(U\) b 、N b satisfies \(MTU\cdot U\) b \(\cdot(N\) b - 1)\(\geq L\) peak . If it is satisfied, set the queue length for transmitting time-sensitive service traffic at the output port of the satellite node to \(L\) b \(= MTU\cdot U\) B , and set the number of queues to \(N\) B . Otherwise, it is determined that the optimization fails, and re-optimize the length and number of queues for transmitting time-sensitive service traffic at the output port of the satellite node.

[0093] In S211, calculate the queue length and number of queues that meet the maximum load requirements of the satellite node in the next time period, specifically including:

[0094] Take the maximum load of the satellite node in the next time period as the optimization goal, and use the particle swarm optimization (PSO) algorithm to calculate the queue length and number of queues that meet the maximum load requirements of the satellite node in the next time period.

[0095] In this embodiment, the particle swarm optimization algorithm, i.e., the PSO algorithm, is used to calculate the queue length that meets the requirements and the number of queues for transmitting time-sensitive service traffic. And the gating control list (GCL) of the satellite node's output port is configured accordingly.

[0096] Specifically as follows:

[0097] Step 1: Initialize the parameters of the PSO algorithm. The initialization of the parameters of the PSO algorithm is as follows: the number of iterations T = 10 3 , the number of swarm particles NP = 20, the learning factor c1 = 1.5, the learning factor c2 = 0.5, the inertia weight ω = 0.9, the maximum value of the velocity of each particle is V max =(2, 2), and the minimum value of the velocity of each particle is V min =(0, 0).

[0098] Step 2: Randomly initialize the initial positions of the particles in the swarm as:

[0099] X i =(u 1i , n 1i , u 2i , n 2i , u 3i , n 3i , u 4i , n 4i ), 1 ≤ i ≤ 20. Randomly initialize the initial velocities V i of the particles in the swarm, and calculate the fitness function value p i =[MTU · u i ·(n i - 1) - L peak 2 at this time, and record the historical best position pbest i of each particle and the global best position gbest of the swarm.

[0100] Step 3: According to the historical best positions of each particle and the global historical best position of the swarm, update the velocities V i ' of each particle using the velocity update formula: V i ' = ω · V i + c1 · rand() · (pbest i - X i ) + c2 · rand() · (gbest i - X i ), where rand() is a random number uniformly distributed in [0, 1]. Further, update the positions X i ' of each particle: X i ' = X

[0101] Step 4: Further, calculate the fitness function value p of each particle at this time i ′, and update the individual historical best position pbest of each particle i and the global best position gbest of the population.

[0102] Step 5: Repeat Step 4 iteratively to calculate the individual historical best position pbest of each particle i and the global best position gbest of the population until the number of iterations reaches the upper limit T = 10 3 . At this time, the global best position gbest of the population = (u b , n b ). That is, the optimal solutions for the queue length and the number of queues of the out - port for transmitting time - sensitive service traffic under high - load conditions of each satellite node in the large - scale low - Earth - orbit satellite network in the embodiment of the present invention obtained by using the PSO algorithm.

[0103] Step 6: Further, round up u b , n b to obtain and judge whether U b , N b satisfy MTU·U b ·(N b - 1)≥L peak . If satisfied, then L b = MTU·U b is the queue length of the out - port for transmitting time - sensitive service traffic under high - load conditions of the satellite node in the large - scale low - Earth - orbit satellite network in the embodiment of the present invention, and N b is the number of queues of the out - port for transmitting time - sensitive service traffic in the satellite node in the large - scale low - Earth - orbit satellite network in the embodiment of the present invention. Otherwise, it is determined that the calculation fails, and return to Step 1.

[0104] Step 7: Further, use L b and N b to configure the gating list of each queue for transmitting time - sensitive service traffic at the out - port under high - load conditions of the satellite node in the large - scale low - Earth - orbit satellite network in the embodiment of the present invention, so as to realize the queue analysis and optimization of the large - scale low - Earth - orbit satellite network. The queue optimization mechanism based on the particle swarm algorithm in the large - scale low - Earth - orbit satellite network is as Figure 3 shown.

[0105] In this embodiment, the particle swarm algorithm is used to calculate the optimal queue length and the number of queues for transmitting time - sensitive service traffic at the out - port of each satellite node, which occupies as little underlying hardware resources as possible on the premise of not causing packet loss due to congestion, ensuring that the implementation difficulty and cost will not be too high.

[0106] In S21, the load status of satellite nodes in the next time period is predicted through a long short-term memory network (LSTM).

[0107] The LSTM algorithm can be used to predict the time-sensitive service traffic of satellite nodes in each time slot in the next period. The LSTM structure in the above S21 is as follows: the number of neurons in the first layer of the LSTM is 64, and the number of neurons in the second layer of the LSTM is 128; the regularization coefficient is 0.7, the loss function is the mean square error (MSE) loss function, and the optimizer is the Adam optimizer; the batch size is 128, and the number of iterations is 150.

[0108] Specifically, by inputting the statistical data of the time-sensitive service traffic of satellite nodes in each time slot in a period before the current moment into the neural network model, the time-sensitive service traffic of satellite nodes in each time slot in the next period is predicted. After completing the prediction of the time-sensitive service traffic of satellite nodes in all time slots in the next period, the load status is then judged.

[0109] In S211, the queue length that meets the maximum load demand of satellite nodes in the next time period is L b and the number of queues is N b , and L b = MTU·U b , 1 ≤ U b ≤ 1000; 2 ≤ N b ≤ 7.

[0110] In summary, the dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks of the present invention comprehensively considers the periodic time-variation of the spatial topology structure in large-scale low-earth orbit satellite networks, the injection of new time-sensitive service flows into the network, etc. Based on the LSTM algorithm, the time-sensitive service traffic of large-scale low-earth orbit satellite network nodes is predicted, and a time-sensitive traffic scheduling and forwarding mechanism with cyclic queue priority is adopted. The PSO algorithm is used to select and optimize the queue length and number of satellite node output ports for transmitting time-sensitive service traffic. The present invention realizes the real-time maintenance and guarantee of low delay and low packet loss rate for the transmission of time-sensitive service traffic.

[0111] The present invention uses the LSTM algorithm to predict the time-sensitive service traffic arriving at satellite nodes in each time slot in the next period. In the case of low load, a time-sensitive traffic scheduling and forwarding mechanism with cyclic queue priority is adopted to cycle the priorities of two forwarding queues, configure the gating lists of the two queues, and forward time-sensitive service data packets to ensure that the time-sensitive service data packets in the two queues have the opportunity to be forwarded within one time slot, breaking through the bandwidth limitation of time-sensitive traffic.

[0112] Under the condition of high load, the present invention uses the particle swarm optimization algorithm to calculate the optimal queue length and the number of queues for the satellite nodes to transmit time-sensitive service traffic at the outgoing ports, occupying as little underlying hardware resources as possible on the premise of ensuring no packet loss due to congestion, and ensuring that the implementation difficulty and cost will not be too high.

[0113] According to one aspect of the present invention, there is provided an electronic device, including: one or more processors, one or more memories, and one or more computer programs; wherein, the processor is connected to the memory, and the above-mentioned one or more computer programs are stored in the memory. When the electronic device runs, the processor executes the one or more computer programs stored in the memory, so that the electronic device executes the dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks as described in any one of the above technical solutions.

[0114] According to one aspect of the present invention, there is provided a computer-readable storage medium for storing computer instructions. When the computer instructions are executed by a processor, the dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks as described in any one of the above technical solutions is implemented.

[0115] The computer-readable storage medium may include any medium capable of storing or transmitting information. Examples of computer-readable storage media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical discs, hard disks, fiber optic media, radio frequency (RF) links, and so on. Code segments can be downloaded via computer networks such as the Internet, intranet, etc.

[0116] The dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks of the present invention includes: predicting the load status of satellite nodes in the next time period according to the load status of satellite nodes in the previous time period; when the load status is a high-load status, executing the high-load traffic scheduling optimization method; when the load status is a low-load status, configuring the queues at the outgoing ports of satellite nodes in the next time period to include a high-priority queue and a low-priority queue; in any time slot of the next time period, first send the service data in the high-priority queue, and when the service data in the high-priority queue is sent out, then send the service data in the low-priority queue to complete the dynamic traffic scheduling optimization; and when the next time slot starts, the high-priority queue and the low-priority queue are respectively converted into a low-priority queue and a high-priority queue.

[0117] In addition, it should be noted that the present invention can be provided as a method, an apparatus, or a computer program product. Therefore, embodiments of the present invention can take the form of an all-hardware embodiment, an all-software embodiment, or an embodiment combining software and hardware aspects. Moreover, embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media that contain computer-usable program code.

[0118] Embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, an embedded processor, or other programmable data processing terminal devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing terminal devices generate a device for realizing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0119] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal devices to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device realizes the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks. These computer program instructions can also be loaded onto a computer or other programmable data processing terminal devices, such that a series of operation steps are executed on the computer or other programmable terminal devices to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable terminal devices provide steps for realizing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0120] It should also be noted that in this document, the terms "comprising", "including", or any other variation thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device including a series of elements not only includes those elements but also includes other elements not explicitly listed, or further includes elements inherent to such a process, method, article, or terminal device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article, or terminal device including the said element.

[0121] Finally, it should be noted that the above description is the preferred embodiment of the present invention. It should be pointed out that although the preferred embodiments of the present invention have been described, for those skilled in the art of this technology, once they know the basic creative concept of the present invention, without departing from the principle described in the present invention, several improvements and refinements can still be made, and these improvements and refinements should also be regarded as the protection scope of the present invention. Therefore, the appended claims are intended to be interpreted to include the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A dynamic traffic scheduling optimization method for large-scale low-Earth orbit satellite networks, characterized in that The method steps are as follows: Predict the load status of the satellite node in the next time period according to the load status of the satellite node in the previous time period; When the load status is a low-load status, execute the low-load traffic scheduling optimization method; The low-load status is: In the case that the queue length and the number of queues at the satellite node output port in the next time period are the same as those in the previous time period, in at least one time slot in the next time period, the state where the lowest-priority queue receives no service data; The low-load traffic scheduling optimization method is: Configure the queue at the satellite node output port in the next time period to include a high-priority queue and a low-priority queue; In any time slot in the next time period, first send the service data in the high-priority queue. When the service data in the high-priority queue is sent completely, then send the service data in the low-priority queue to complete the dynamic traffic scheduling optimization; and when the next time slot starts, convert the high-priority queue and the low-priority queue in the previous time slot into a low-priority queue and a high-priority queue respectively.

2. The dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to claim 1, wherein It also includes: When sending the service data in the high-priority queue, receive the service data transmitted by other satellite nodes into the low-priority queue.

3. The dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to claim 1, wherein The queue lengths of the high-priority queue and the low-priority queue are both l1; l1 = MTU·u1 Wherein, MTU represents the maximum transmission unit for sending service data in the large-scale low-earth orbit satellite network, with the unit of byte; u1 represents the number of MTUs included in the high-priority queue or the low-priority queue, and u1 = 1000.

4. The dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to any one of claims 1 to 3, characterized in that Predict the load status of the satellite node in the next time period through the long short-term memory network LSTM.

5. The dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to claim 4, characterized in that The number of neurons in the first layer of the LSTM is 64, and the number of neurons in the second layer of the LSTM is 128; the regularization coefficient is 0.7, the loss function is the mean square error MSE loss function, and the optimizer is the Adam optimizer; the batch size is 128, and the number of iterations is 150.

6. A dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks, characterized in that The method steps are as follows: Predict the load status of the satellite node in the next time period according to the load status of the satellite node in the previous time period; When the load status is a high-load status, execute the high-load traffic scheduling optimization method; The high-load status is: In the case that the queue length and the number of queues at the satellite node output port in the next time period are the same as those in the previous time period, in at least one time slot in the next time period, the state where the service data received by the satellite node exceeds the total storage space sum of all non-highest-priority queues; The high-load traffic scheduling optimization method is: Obtain the maximum load of the satellite node in the next time period based on the service data traffic received by the satellite node in the next time period, and calculate the queue length L that meets the maximum load requirement of the satellite node in the next time period b and the number of queues N b ; Configure the queue at the satellite node's outgoing port in the next time period to the queue length L b and the number of queues N b , and then send service data for dynamic traffic scheduling optimization.

7. The dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to claim 6, characterized in that Calculate the queue length L that meets the maximum load demand of the satellite nodes in the next time period b and the number of queues N b , specifically including: Taking the maximum load of satellite nodes in the next time period as the optimization goal, the particle swarm optimization (PSO) algorithm is used to calculate the queue length L that meets the maximum load demand of satellite nodes in the next time period b and the number of queues N b .

8. The dynamic traffic scheduling optimization method for large-scale low-Earth orbit satellite networks according to claim 6 or 7, characterized in that Predict the load status of the satellite node in the next time period through the long short-term memory network LSTM.

9. The dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to claim 8, characterized in that The number of neurons in the first layer of the LSTM is 64, and the number of neurons in the second layer of the LSTM is 128; the regularization coefficient is 0.7, the loss function is the mean square error MSE loss function, and the optimizer is the Adam optimizer; the batch size is 128, and the number of iterations is 150.

10. The dynamic traffic scheduling optimization method for large-scale low-earth orbit satellite networks according to claim 7 or 9, characterized in that L b = MTU·U b , 1 ≤ U b ≤ 1000; and 2 ≤ N b ≤ 7.

Images