Flow multi-path scheduling method for low-orbit satellite constellation

Abstract

The invention relates to the technical field of satellite communication, in particular to a multi-path traffic scheduling method for a low-orbit satellite constellation, which comprises the following steps: executing a multi-path routing algorithm based on ant colony optimization to obtain a global disjoint path set from a source satellite to a target satellite; based on the time delay, the bandwidth utilization rate and the energy residual rate of each path in the global disjoint path set, calculating the comprehensive overhead of each path and estimating the time delay sensitive type and time delay tolerant type traffic proportion; executing a multi-path traffic scheduling mechanism, determining the distribution proportion of the two types of traffic on each path in the global disjoint path set, and completing data forwarding; according to the method, a bidirectional interaction mechanism of the forward ant packet and the reverse ant packet is designed, low-overhead and high-reliability disjoint multipath detection and screening are realized in combination with link and path parameter constraints of the satellite network, and the detection and screening efficiency is improved in combination with the link and path parameter constraints of the satellite network. According to the invention, delay-sensitive and delay-tolerant traffic can be efficiently distributed on a plurality of paths.

Description

Technical Field

[0001] This invention relates to the field of satellite communication technology, and specifically to a traffic multipath scheduling method for low-Earth orbit satellite constellations. Background Technology

[0002] Low Earth Orbit (LEO) satellite networks have become the core infrastructure of next-generation integrated space-ground communication networks, widely carrying diverse services such as voice calls, industrial remote control, high-definition video-on-demand, and massive data backhaul for the Internet of Things. Traffic such as voice interaction, emergency command instructions, and industrial control signals have stringent requirements for end-to-end transmission latency; exceeding latency limits will directly lead to service interruption or control failure. In contrast, traffic such as high-definition video buffering, file downloads, and emails has a higher tolerance for latency fluctuations, accepting transmission latency within hundreds of milliseconds.

[0003] However, the highly dynamic nature of network topology leads to unstable path latency, making single-path transmission unsuitable for varying latency requirements. Low-Earth orbit satellites orbit the Earth at high speeds, with their relative positions changing in real time. This causes continuous fluctuations in the propagation and queuing latency of inter-satellite links. The end-to-end latency of the same path can surge from 30ms to 200ms, or even increase indefinitely due to link interruptions. Traditional single-path transmission schemes rely on a single fixed path to carry all types of traffic. If path latency suddenly increases, it will simultaneously affect the transmission efficiency of both types of traffic. Even if some schemes incorporate path switching mechanisms, the latency and signaling overhead of re-exploring the path during switching are still insufficient to meet the real-time transmission needs of latency-sensitive traffic.

[0004] Therefore, how to achieve precise scheduling of traffic with various latency requirements has become a core issue for low-orbit satellite networks to ensure service quality and improve resource utilization. Summary of the Invention

[0005] In view of this, this application discloses a traffic multipath scheduling method for low-Earth orbit satellite constellations to solve the above problems; including:

[0006] S1. In each routing update cycle, with the goal of discovering non-intersecting multipaths that meet bandwidth and energy constraints, execute a multipath routing algorithm based on ant colony optimization to obtain a global set of non-intersecting paths from the source satellite to the destination satellite.

[0007] S2. In each time slice of the routing update cycle, based on the real-time latency, bandwidth utilization and energy remaining rate of each path in the global disjoint path set, calculate the comprehensive cost of each path and estimate the proportion of latency-sensitive and latency-tolerant traffic.

[0008] S3. Based on the overall path cost and traffic ratio, with the goal of matching traffic latency requirements and balancing on-board resource utilization, a multi-path traffic scheduling mechanism is executed to determine the allocation ratio of the two types of traffic on each path in the global set of non-intersecting paths and complete data forwarding.

[0009] The beneficial effects of this application include:

[0010] By calculating the comprehensive cost of the path using multidimensional parameters, a bidirectional interaction mechanism between forward and reverse ant packets is designed. Combined with the link and path parameter constraints of the satellite network, low-cost and high-reliability non-intersecting multipath detection and filtering is achieved.

[0011] By distributing the load through non-intersecting multi-paths, we can avoid single-path resource overload, flexibly allocate optimal path resources, improve resource utilization, and alleviate the contradiction of satellite resource constraints.

[0012] To ensure balanced resource utilization, a latency-differentiated service traffic scheduling strategy was designed. First, the allocation ratio of the two types of traffic on each path was determined to estimate the traffic proportion. Then, the path allocation weight was calculated, and finally, the specific allocation strategy was determined. Latency-sensitive traffic was preferentially allocated to the optimal path to control latency, while latency-tolerant traffic was allocated to non-optimal paths according to the overall path cost. This achieves efficient allocation of latency-sensitive and latency-tolerant traffic on multiple paths, avoids occupying critical resources, and accurately adapts to the needs of the two types of traffic. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of a multipath transmission scenario in a low-orbit satellite network, as illustrated in this application.

[0014] Figure 2 This is a schematic diagram of the forward ant packet message format in the embodiments of this application;

[0015] Figure 3 This is a schematic diagram of the backward ant packet message format in the embodiments of this application;

[0016] Figure 4 This is a pseudocode diagram of a multi-path routing algorithm based on ant colony optimization in the embodiments of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, features, and advantages of this application clearer and to enable those skilled in the art to better understand the technical solutions of this application, the following detailed description of this application is provided in conjunction with the accompanying drawings and embodiments.

[0018] Example 1:

[0019] This embodiment includes a traffic multipath scheduling method for low-Earth orbit satellite constellations, comprising:

[0020] S1. In each routing update cycle, with the goal of discovering non-intersecting multipaths that meet bandwidth and energy constraints, execute an ant colony-optimized multipath routing algorithm to obtain a global set of non-intersecting paths from the source satellite to the destination satellite.

[0021] S2. In each time slice of the routing update cycle, based on the real-time latency, bandwidth utilization, and energy reserve of each path in the global disjoint path set, calculate the comprehensive cost of each path and estimate the proportion of latency-sensitive and latency-tolerant traffic.

[0022] S3. Based on the overall path cost and traffic ratio, with the goal of matching traffic latency requirements and balancing on-board resource utilization, a multi-path traffic scheduling mechanism is executed to determine the allocation ratio of the two types of traffic on each path in the global set of non-intersecting paths and complete data forwarding.

[0023] Specifically, in this embodiment, the multipath transmission scenario of the low-orbit satellite network is as follows: Figure 1 As shown, the scenario includes a constellation of M×N low-Earth orbit satellites and ground terminals distributed in different areas. The satellites are interconnected via inter-satellite links, and the ground terminals are connected to the nearest satellite via satellite-to-ground links. The source ground terminal initiates a transmission request, which includes two types of traffic: latency-sensitive and latency-tolerant. Latency-sensitive traffic includes voice and commands, which accounts for a small proportion and requires low-latency transmission. Latency-tolerant traffic includes high-definition video and files, which accounts for a large proportion and can tolerate a certain level of latency. After receiving the request, the source satellite sends forward ant packets to explore paths, and finally selects 2-4 non-intersecting multi-paths without shared nodes. Then, the latency-sensitive traffic is allocated to the optimal path with the lowest latency, and the latency-tolerant traffic is allocated to other paths according to the overall path cost, so as to achieve efficient transmission of the two types of traffic and balanced utilization of on-board resources.

[0024] To accurately characterize the dynamic topology of low-Earth orbit satellite networks, time slots are defined. The network topology at time t is an undirected graph. ,in, Represents a set of satellite nodes. Indicates the number of orbital planes. Indicates the number of satellites in each orbit, satellite nodes Spatial coordinates are It is updated in real time according to the satellite's orbital movement; Indicates time slot The set of active inter-satellite links at any given time, if satellite With satellite The relationship must be satisfied with either "adjacent on the same track" or "corresponding location on adjacent tracks", and the link bandwidth utilization must be less than a preset threshold. The remaining energy of the satellites at both ends of the link is greater than a preset threshold. Define the source satellite. With the target satellite The set of available paths is ,in This represents the optimal path. Path From satellite node sequence and corresponding inter-satellite links Constitute, and any two paths It satisfies the "non-intersecting" constraint for the absence of shared satellite nodes.

[0025] Furthermore, the goal is to discover disjoint multipaths that meet bandwidth and energy constraints. Since inter-satellite link delay directly determines path end-to-end delay and is crucial for ensuring the transmission quality of delay-sensitive traffic, and bandwidth and energy are scarce resources in low-Earth orbit (LEO) satellite networks, the bandwidth utilization and energy surplus of a path directly affect transmission stability and the satellite's on-orbit lifespan. Therefore, key parameters such as link and path delay, bandwidth, and energy are core indicators for evaluating the transmission quality of LEO satellite networks. To achieve a quantitative description of these parameters, mathematical models are performed for link delay, path bandwidth utilization, and path energy surplus.

[0026] Modeling link latency includes: inter-satellite links The total delay is decomposed into two parts: propagation delay and queuing delay. The formula for calculating the total delay is:

[0027]

[0028]

[0029]

[0030]

[0031] in, Indicates link During the time delay of time slot t, The propagation delay is determined by the spatial distance between satellites and the speed of electromagnetic wave propagation. Represents real-time physical distance, based on time slots. Inner satellite node and The spatial coordinates are obtained through calculation. This indicates queuing latency, which is determined by the backlog of data packets in the link buffer queue. Indicates time slot The total number of packets in the internal link buffer queue, This indicates that a single data packet has a fixed length. This represents the rated transmission bandwidth of the link. The total end-to-end delay of the path is obtained by summing the delays of all inter-satellite links included in the path, using the following formula:

[0032]

[0033] Modeling path bandwidth utilization includes: defining the path The bandwidth utilization rate is the maximum of the bandwidth utilization rates of all inter-satellite links within the path, and the formula is:

[0034]

[0035] in, Indicates link This parameter is used to determine whether there is a risk of bandwidth overload on the path, ensuring that the path meets the preset bandwidth constraints, given the bandwidth resources already occupied in time slot t.

[0036] Modeling the energy surplus rate includes: defining satellite nodes At the time slot The real-time remaining energy is calculated from the remaining energy of the previous time slot, the energy consumption in the current time slot, and the solar energy replenishment, using the following formula:

[0037]

[0038]

[0039] in, Indicates satellite In the time slot Battery energy, Indicates satellite Initial battery energy capacity, This indicates the solar energy acquired by the satellite. For time slots Time Satellite The total energy consumption. Considering satellite launch, of the total energy consumption, Indicates satellite Send to Number of data packets Indicates satellite node The set of neighboring nodes, Indicates satellite transmission power. Indicates satellite receiving power. Indicates the satellite's operating power. This indicates the total operating time of the satellite.

[0040] The energy stability of a path is determined by the minimum energy surplus rate of the satellite nodes it contains. (The path is defined as follows.) The energy surplus rate is the minimum energy surplus rate of all satellite nodes within the path, and its calculation formula is:

[0041]

[0042] This parameter is used to assess the energy support capability of the path and avoid transmission interruption of the entire path due to the depletion of energy of a single satellite node.

[0043] Furthermore, a path cost breakdown is constructed, which balances the impact of bandwidth utilization and energy reserve rate through weighting coefficients, and comprehensively evaluates the transmission cost of the path by combining the path hop count. The formula is as follows:

[0044]

[0045] in, Representing a path The total cost This is a weighting coefficient for bandwidth utilization, with a value range of [value range missing]. It can be dynamically adjusted according to the network's resource optimization goals. This is the weighting coefficient corresponding to the energy surplus rate. For path The number of inter-satellite links included is the hop count. Total path cost. The smaller the value, the better the transmission quality of the path. This parameter provides the core quantitative basis for the path selection and traffic allocation ratio calculation in subsequent multi-path traffic scheduling.

[0046] Ant colony optimization (ACO) algorithms possess the characteristics of distributed decision-making, self-organization, and dynamic convergence. They do not rely on global network information and are highly compatible with the distributed architecture of low-Earth orbit (LEO) satellite networks. An adapted ACO algorithm is designed to obtain high-quality, non-intersecting paths that meet bandwidth and energy constraints. Specifically, a bidirectional interaction mechanism between forward and reverse ant packets is designed, combined with the link and path parameter constraints of the satellite network, to achieve low-overhead, high-reliability non-intersecting multipath detection and selection.

[0047] The routing source node discovers multiple paths by sending ant packets into the network, thus obtaining effective path analysis without the need for global data collection.

[0048] Ant packets are divided into forward ant packets (FAP) and backward ant packets (BAP). In each time slot, multiple disjoint paths that meet the constraints of latency, bandwidth and energy are discovered, thereby satisfying the transmission requirements of services with different latency requirements.

[0049] At the start of each time slot, the source node A batch of FAPs is generated and sent to the network; each node forwards the FAPs according to its local pheromone concentration; when the FAPs reach the destination node... At that time, a path is discovered, and the first discovered path's FAP has the minimum latency, i.e., the optimal path; the set of node addresses is stored, and if the hop count and latency exceed preset constraints, the FAP is destroyed at the current node; target satellite. Upon receiving a FAP, a BAP is generated. The BAP is used for path information feedback and pheromone concentration updates. The BAP packet is returned to the source node along the path stored in the FAP. At the end, the set of globally disjoint paths from the source satellite to the destination satellite is obtained.

[0050] The message format of the Forward Ant Packet (FAP) designed in this application is as follows: Figure 2 As shown, it includes: a type field, a delay field, a hop count field, a reserved field, a source satellite IP address, a destination satellite IP address, the set of optimal path addresses from the previous time slot, and the set of satellite node addresses already traversed. The type field distinguishes between forward and backward ant packets, the delay field records the total cumulative delay of the FAP from the source satellite through all links, the hop count field records the number of satellite nodes traversed by the FAP, and the reserved field is used for expansion. Node In the time slot Forward FAP to the node The probability is:

[0051]

[0052] in, Represents a node In the time slot Send to node The probability, Indicates link In the time slot pheromone concentration, Indicates the heuristic factor. They represent the weighting factors, Represents a node The set of neighboring nodes. To satisfy bandwidth and energy consumption constraints, each node is limited to containing only those that satisfy... , , Time slot The bandwidth resources that are currently occupied on the current link. This indicates the preset link bandwidth utilization threshold. Indicates time slot Satellite energy remaining rate at any time This indicates the preset satellite energy remaining rate threshold; otherwise, the FAP will be discarded.

[0053] The message format of a backward ant packet is as follows: Figure 3 As shown, it includes: type field, latency field, energy field, bandwidth field, source satellite IP address, fixed satellite IP address, and set of addresses of nodes already passed; the energy field records the minimum remaining energy of all satellite nodes on the FAP detection path, and the bandwidth field records the maximum bandwidth utilization of all links on the FAP detection path; BAP records the links passed through. The following formula is used to update pheromones:

[0054]

[0055] in, Indicates time slot After internal pheromone updates, Indicates the first The path latency of each record. Indicates the pheromone decay rate. Indicates the process The total number of BAPs.

[0056] To adapt to dynamic changes in satellite topology, the source node continuously transmits FAP (Fast-Ahead Prediction), but the pheromone concentration increases over time along the optimal path, thus weakening the FAP's ability to discover disjoint paths. To address this issue, the source node employs the following interaction mechanism to ensure that FAP discovers disjoint multiple paths:

[0057] 1. The source node can communicate with a maximum of four neighboring satellites and send FAP to the neighboring satellites in proportion.

[0058] 2. The source node records the optimal path as follows: If the next time slot to If the minimum hop count constraint remains unchanged, meaning the service latency requirement remains the same, then the source node will... Add it to the traversed node record of the FAP sent to non-optimal neighboring nodes; if the minimum hop count constraint changes, do not modify the node record.

[0059] To ensure balanced resource utilization, this application designs a latency-differentiated service traffic scheduling strategy to efficiently allocate latency-sensitive and latency-tolerant traffic across multiple paths. The strategy first determines the allocation ratio of the two types of traffic on each path, estimates the traffic proportion, calculates the path allocation weight, and finally determines the specific allocation strategy.

[0060] Specifically, due to the inability to obtain the current time slot in real time The exact number of data packets for both types of traffic is estimated using a weighted method based on historical data to calculate the traffic share of the current time slot. The formula is:

[0061]

[0062]

[0063] in, This indicates the current percentage of traffic in a time slot. This indicates the number of delay-sensitive traffic packets in the previous time slot. This indicates the number of delay-tolerant traffic packets in the previous time slot. This represents the weighting coefficient, used to integrate historical estimates. Compared with the actual flow rate of the previous time slot .

[0064] Calculate the traffic allocation weight for a single path The path weight is determined by the proportion of total traffic that a path should bear. The design principle for traffic allocation weight is that the lower the path cost, the higher the allocation weight. In this embodiment, the path weight formula is:

[0065]

[0066] in, Represents a set of paths The total number of paths in Indicates path cost.

[0067] Based on traffic share With path weight A differentiated allocation rule is adopted, prioritizing latency-sensitive traffic and then allocating latency-tolerant traffic reasonably. In this embodiment, the allocation rule includes:

[0068] For latency-sensitive traffic, it is assigned to the optimal path regardless of the path weight.

[0069] For delay-tolerant traffic, the following two scenarios apply:

[0070] like If the optimal path has sufficient remaining resources, the flow is allocated according to the following formula:

[0071]

[0072] in, Indicates the flow rate at the th Weights assigned to each path (in percentage form). Represents path weight, This indicates the current percentage of traffic in a time slot. This represents the path weight for delay-tolerant traffic.

[0073] like In other words, the optimal path resources can only satisfy latency-sensitive traffic, and all latency-sensitive traffic is allocated to the optimal path. Latency-tolerant traffic is allocated according to the following formula: .

[0074] Example 2:

[0075] This embodiment includes a traffic multipath scheduling method for low-Earth orbit satellite constellations. The difference from Embodiment 1 is that, in this embodiment, the pseudocode for the ant colony-optimized multipath routing algorithm is as follows: Figure 4 As shown, it includes:

[0076] Step 1: Initialize network parameters and probe packets.

[0077] In the satellite network, the pheromone concentration of each node is set to the same initial value, and the optimal path recorded in the initial FAP is... Empty, in time slot Source satellite In equal proportions, a batch of generated FAPs will be sent to their neighboring node sets. .

[0078] Step 2: Forward ant packet constraint forwarding and invalid packet destruction.

[0079] If FAP reaches satellite nodes adjacent to the source node ,and The set of neighboring nodes Non-empty ( Only includes those that meet bandwidth and energy constraints and are not (Neighbors of the nodes in the satellite node) Parse FAP, if satellite nodes If it is not the destination node of this FAP, then send it with probability. Send FAP to neighboring nodes And update the relevant FAP fields; if If the value is empty (no neighbor matching the constraints), discard the FAP and process other FAPs.

[0080] Step 3: Reverse ant packet generation and dynamic pheromone update.

[0081] If satellite node This is the destination node and destination satellite of the FAP. A corresponding reverse ant packet (BAP) is generated, carrying the complete path information recorded in the original FAP. The FAP is then destroyed. The BAP returns to the source satellite along the original path recorded in the FAP. The pheromone is updated on the network nodes along the path according to the pheromone update formula.

[0082] Step 4: Filtering non-intersecting paths and determining the optimal path.

[0083] The source node parses the BAP upon receiving it. If this BAP is the first one to arrive... BAP: Directly add its corresponding path to the set of disjoint paths. (The initial optimal path for the current time slot) ), and calculate the source satellite to the target satellite "Minimum number of hops" If it is not the first arrival, check if the path corresponding to this BAP is the same as... All existing paths are "non-intersecting" (i.e., have no shared satellite nodes). If so, the new path is added. If not (path overlap), then discard the path.

[0084] Step 5: Adjust the path set output to match the next time slot probe packet.

[0085] Time slot After all FAP processing is completed, the set of non-intersecting paths for the current time slot is output. Next time slot At the start of the calculation, compare the optimal path with the previous time slot. The number of hops and the minimum number of hops under the current flow constraint. If the two are equal, the FAP sent to the "non-optimal neighbor node" carries... Complete path information; if the two are not equal, there is no need to include it in the FAP. information.

[0086] Finally, it should be noted that the above description only depicts some embodiments of this application. For those skilled in the art, various changes, modifications, substitutions, and variations can be conceived of these embodiments without departing from the principles and spirit of this application. The scope of protection of this application is defined by the appended claims and their equivalents, and all the above-mentioned behaviors should be covered within the scope of protection of this application.

Claims

1. A traffic multipath scheduling method for a low earth orbit satellite constellation, characterized in that, include: S1. In each routing update cycle, with the goal of discovering non-intersecting multipaths that meet bandwidth and energy constraints, execute a multipath routing algorithm based on ant colony optimization to obtain a global set of non-intersecting paths from the source satellite to the destination satellite. S2. In each time slice of the routing update cycle, based on the real-time latency, bandwidth utilization and energy remaining rate of each path in the global disjoint path set, calculate the comprehensive cost of each path and estimate the proportion of latency-sensitive and latency-tolerant traffic. S3. Based on the overall path cost and traffic ratio, with the goal of matching traffic latency requirements and balancing on-board resource utilization, a multi-path traffic scheduling mechanism is executed to determine the allocation ratio of the two types of traffic on each path in the global set of non-intersecting paths and complete data forwarding.

2. The traffic multipath scheduling method for low earth orbit satellite constellations of claim 1, wherein, The calculation of the comprehensive cost of each path and the estimation of the traffic ratio of latency-sensitive and latency-tolerant traffic are performed by mathematically modeling link latency, path bandwidth utilization, and path energy surplus rate, respectively. Construct a comprehensive path cost, balance the impact of bandwidth utilization and energy reserve rate through weighting coefficients, and comprehensively evaluate the transmission cost of the path in combination with the path hop count.

3. The traffic multipath scheduling method for low earth orbit satellite constellations of claim 1, wherein, The ant colony optimization-based multi-path routing algorithm divides ant packets into forward ant packets (FAP) and backward ant packets (BAP), and discovers multiple non-intersecting paths that meet the constraints of delay, bandwidth, and energy in each time slot. At the start of each time slot, the source node A batch of FAPs is generated and sent to the network; each node forwards the FAPs according to its local pheromone concentration; when the FAPs reach the destination node... When a path is discovered, the FAP of the first discovered path has the minimum latency, which is the optimal path; the set of node addresses is stored, and if the number of hops and the latency exceed the preset constraints, the FAP is destroyed at the current node; target satellite Upon receiving a FAP, a BAP is generated. The BAP is used for path information feedback and pheromone concentration updates. The BAP packet is returned to the source node along the path stored in the FAP. At the end, a set of globally disjoint paths from the source satellite to the destination satellite is obtained.

4. The traffic multipath scheduling method for a low earth orbit satellite constellation of claim 3, wherein, The forward ant packet (FAP) has the following message format: type field, delay field, hop number field, reserved field, source satellite IP address, destination satellite IP address, optimal path address set of the previous time slot, and satellite node address set already passed. The type field is used to distinguish between forward and backward ant packets. The delay field records the total cumulative delay of the FAP after it departs from the source satellite and passes through all links. The hop number field records the number of satellite nodes passed by the FAP. The reserved field is used for expansion.

5. The traffic multipath scheduling method for low-Earth orbit satellite constellations according to claim 3, characterized in that, The Backward Ant Packet (BAP) includes the following message format: type field, delay field, energy field, bandwidth field, source satellite IP address, fixed satellite IP address, and set of addresses of nodes already passed; the energy field records the minimum energy remaining rate of all satellite nodes on the FAP detection path, and the bandwidth field records the maximum bandwidth utilization of all links on the FAP detection path.

6. The traffic multipath scheduling method for low earth orbit satellite constellations of claim 3, wherein, Each node forwards FAP based on its local pheromone concentration. In the time slot Forward FAP to the node The probability is: in, Represents a node In the time slot Send to node The probability, Indicates link In the time slot pheromone concentration, Indicates the heuristic factor. They represent the weighting factors, Represents a node The set of neighboring nodes.

7. The traffic multipath scheduling method for low earth orbit satellite constellations of claim 3, wherein, The pheromone concentration is updated using the following formula: in, Indicates time slot After internal pheromone updates, Indicates the first The path latency of each record. Indicates the pheromone decay rate. Indicates the process The total number of BAPs.

8. The traffic multipath scheduling method for low earth orbit satellite constellations of claim 3, wherein, The source node employs the following interaction mechanism: the source node communicates with a maximum of four neighboring satellites, sending FAPs proportionally to each neighboring satellite; the source node records the optimal path as follows. If the next time slot to If the minimum hop count constraint remains unchanged, meaning the service latency requirement remains the same, then the source node will... Add it to the traversed node record of the FAP sent to non-optimal neighboring nodes; if the minimum hop count constraint changes, do not modify the node record.

9. The traffic multipath scheduling method for low earth orbit satellite constellations of claim 3, wherein, The multi-path traffic scheduling mechanism includes: based on traffic percentage With path weight A differentiated allocation rule is adopted, prioritizing the allocation of latency-sensitive traffic and then allocating latency-tolerant traffic. For latency-sensitive traffic, it is assigned to the optimal path regardless of the path weight; For latency-tolerant traffic, the allocation is divided into the following two cases: If , the flow is distributed as follows: in, Indicates the flow rate at the th Weights are assigned to each path. Represents path weight, This indicates the current percentage of traffic in a time slot. Represents the path weight for delay-tolerant traffic; like All latency-sensitive traffic is assigned to the optimal path, while latency-tolerant traffic is allocated according to the following formula: .

Images