Satellite communication multi-network bandwidth global dynamic allocation method for multi-level terminal group

Abstract

The application discloses a kind of satellite communication multi-network bandwidth global dynamic allocation methods for multi-level terminal group.It includes the construction of two-level bandwidth management architecture containing terminal group priority and terminal priority decoupling configuration;Collect and filter terminal bandwidth demand based on local configuration upper limit, generate global bandwidth application;Bandwidth global scheduling is executed by gathering global bandwidth application;Based on global scheduling result, execute local bandwidth allocation containing idle bandwidth allocation according to priority decoupling mechanism, and combine spectrum efficiency to map service bandwidth as channel bandwidth;According to terminal bandwidth satisfaction rate, determine the congestion state of terminal, and execute carrier switching based on competition benefit prediction for congested terminal;Satisfaction rate based on historical period generates negative feedback regulation signal to dynamically update terminal group configuration weight.The application improves the isolation problem of multi-level network resource allocation, improves the adaptive balancing ability of system to deal with sudden traffic and the comprehensive utilization rate of whole network channel resources.

Description

Technical Field

[0001] This invention relates to the field of satellite communication technology, and in particular to a method for global dynamic allocation of bandwidth across multiple satellite communication networks for multi-level terminal groups. Background Technology

[0002] With the development of high-throughput satellite communication technology, the scale of terminal access in satellite communication systems continues to expand, and service scenarios are characterized by multiple terminals, multiple services, and high bandwidth requirements. In complex communication scenarios encompassing multiple interactive networks and possessing multi-level terminal group architectures, how to scientifically coordinate and finely schedule limited satellite space band frequency resources is the underlying technology for ensuring the quality of concurrent communication among massive terminals. The ability to globally coordinate and dynamically allocate bandwidth resources directly determines the reliability of service transmission and the comprehensive utilization limit of spectrum resources under a multi-network collaborative architecture.

[0003] Current mainstream bandwidth management schemes typically employ a distributed, single-network, single-level local management mechanism. Each local network independently allocates channel resources to terminals within its jurisdiction based on preset static priorities or fixed weights. When facing cross-network macro-level resource competition, this allocation logic often places terminals and terminal groups on the same level for indiscriminate game-playing, easily leading to overlapping bandwidth competition logic between and within groups. Furthermore, when handling sudden congestion, it often relies on macro-level carrier-level overall throughput thresholds to passively trigger carrier scheduling or frequency switching actions. In addition, existing systems generally follow mechanical polling or a lack of targeted, fixed-ratio allocation strategies in the secondary reclamation and allocation of idle bandwidth capacity.

[0004] The aforementioned static and one-dimensional management and control model is prone to causing imbalances in cross-network resource distribution and the crowding out of high-priority services by edge nodes when dealing with the highly dynamic and sudden nature of modern satellite services. Furthermore, the lack of a micro-level individual state awareness and adaptive adjustment framework results in insufficient resource utilization and service response robustness under complex operating conditions.

[0005] Therefore, it is necessary to study a bandwidth management method that can improve the flexibility and fairness of channel resource scheduling in complex multi-network concurrent scenarios. Summary of the Invention

[0006] Purpose of the invention: To provide a global dynamic allocation method for bandwidth across multiple satellite communication networks for multi-level terminal groups, in order to solve the aforementioned problems in the prior art.

[0007] Technical solution: A method for global dynamic allocation of bandwidth across multiple satellite communication networks for multi-level terminal groups, comprising: Construct a two-level bandwidth management architecture and configure mutually decoupled terminal group priorities, terminal priorities, and terminal group configuration weights; Collect terminal bandwidth requirements, filter the terminal bandwidth requirements based on the preset local configuration limit, and generate a global bandwidth request; The global bandwidth requests are aggregated and global bandwidth scheduling is performed to obtain the global scheduling result; Based on the global scheduling result, local bandwidth allocation including idle bandwidth allocation is performed according to the terminal group priority and the terminal priority to obtain the service bandwidth, the terminal group bandwidth satisfaction rate and the terminal bandwidth satisfaction rate, and the service bandwidth is mapped to the channel bandwidth in combination with the obtained spectrum efficiency of each terminal. The terminal congestion state is determined based on the terminal bandwidth satisfaction rate, and carrier handover is performed on the terminal in the terminal congestion state based on competition benefit prediction. A negative feedback adjustment signal is generated based on the terminal group bandwidth satisfaction rate of historical periods, and the terminal group configuration weight is updated based on the negative feedback adjustment signal.

[0008] Beneficial effects: By constructing a priority-decoupled two-level bandwidth management architecture, physical isolation between macro-level competition between groups and micro-level allocation within groups is achieved, suppressing the phenomenon of preemption by higher levels and clearly defining the scheduling hierarchy boundaries of multi-level networks.

[0009] The introduction of a satisfaction rate balancing factor and a dynamic weight negative feedback mechanism mathematically couples static management intentions with real-time business gaps, enabling idle resources to automatically tilt to the node with the highest hunger level, which is conducive to automatically converging the satisfaction rate differences of multiple resource groups to a dynamic equilibrium level.

[0010] A frequency hopping decision model based on a competitive discount factor was designed, which refines the frequency hopping triggering and target selection mechanism into a micro net benefit prediction from an individual perspective, reducing the risk of terminals mistakenly entering false idle carriers and reducing the underlying signaling overhead caused by invalid handover.

[0011] By combining the spectral efficiency of the underlying modulation and coding scheme with cross-layer weighted mapping, high signal-to-noise ratio terminals can occupy fewer physical frequency bands under the same service requirements. This transforms simple service distribution fairness into spectral efficiency-based physical bearer fairness, thereby improving the throughput potential of high-throughput satellite carriers. Attached Figure Description

[0012] Figure 1 This is a flowchart of the overall method of the present invention.

[0013] Figure 2 A flowchart illustrating the construction of a two-level bandwidth management architecture for this invention.

[0014] Figure 3 This is a flowchart illustrating the global bandwidth allocation and scheduling process for the present invention.

[0015] Figure 4This is a flowchart of the local multi-round bandwidth allocation of the present invention. Detailed Implementation

[0016] Combination Figure 1 Description of Example 1 A method for global dynamic allocation of bandwidth across multiple satellite communication networks for multi-level terminal groups includes the following steps: Step 101: Construct a two-level bandwidth management architecture and configure mutually decoupled terminal group priorities and terminal priorities; In this embodiment, the two-level bandwidth management architecture specifically refers to a hierarchical management topology composed of a global network management system and multiple underlying interactive networks. The global network management system is responsible for coordinating the satellite communication frequency band resources of the entire network, while each interactive network is responsible for managing the satellite communication terminals within its coverage area.

[0017] In this architecture, there are multiple levels of business organization, namely terminal groups. To avoid the drawbacks of terminals within a group competing for resources in the traditional allocation model, independent scheduling parameters are set for each level of nodes.

[0018] Specifically, terminal group priority controls the order in which different terminal groups compete for global or local public bandwidth; terminal priority, on the other hand, determines the order in which each physical terminal within the same terminal group acquires its allocated bandwidth. This separate configuration enables physical isolation between inter-group competition and intra-group allocation.

[0019] Furthermore, in a specific system deployment scheme, the two-level bandwidth management architecture can be represented as two sets of mapped tree data structures, including a global tree that is effective across all interactive networks, and a local tree that is effective only within a single interactive network. By attaching physical terminals as leaf nodes to the corresponding branches, a strict resource inheritance and allocation path can be established.

[0020] In one hardware deployment method of this embodiment, the global network management system is deployed on the server cluster of the ground network control center, and the global scheduling results are sent to the gateway station controllers of each interactive network through the in-band management channel of the satellite communication system or the ground dedicated control network.

[0021] Step 102: Collect terminal bandwidth requirements, filter the terminal bandwidth requests based on the local configuration limit, and generate a global bandwidth request; Specifically, each interactive network periodically monitors the real-time traffic data of online terminals within its jurisdiction, and uses statistical algorithms to extract values ​​that accurately reflect the peak traffic conditions of the business as the bandwidth requirements of the terminals.

[0022] Before reporting the request to the global network management system, the system will retrieve the local configuration limit parameters pre-assigned to the terminal group and execute a pre-allocation logic.

[0023] This pre-allocation logic does not actually control the communication rate of physical terminals, but is used for compressing demand data. The system compares the terminal bandwidth demand with the local configuration limit, truncating and discarding invalid demand portions that exceed the configuration limit. The truncated valid data is then aggregated and packaged according to the interactive network, thereby forming the global bandwidth request.

[0024] By removing a specific proportion of extreme values, we can avoid the distortion of demand assessment caused by sudden interference traffic, thereby reducing the signaling overhead of the global network management system.

[0025] Step 103: Aggregate the global bandwidth requests and execute global bandwidth scheduling to obtain the global scheduling result; In this embodiment, after receiving the global bandwidth requests reported by each interactive network, the global network management system will use the total available bandwidth of the entire satellite communication system as a resource pool and perform top-down macro-allocation.

[0026] The allocation process strictly adheres to the Quality of Service (QoS) agreement, prioritizing basic communication performance before allocating additional resources to enhance speed. Bandwidth quota limits are calculated for each terminal group on each interactive network. These quota limits are encapsulated into configuration commands and sent to the corresponding interactive networks, forming the global scheduling result.

[0027] Furthermore, if there is still unallocated remaining capacity in the total available bandwidth of the system after the global bandwidth scheduling is performed, the global network management system will continue to allocate this remaining capacity to the child nodes that still have potential demand, in order to ensure that the satellite space segment resources are not idle.

[0028] Step 104: Based on the global scheduling result, perform local bandwidth allocation including idle bandwidth allocation according to the terminal group priority and the terminal priority, obtain the terminal group bandwidth satisfaction rate of each terminal group and the terminal bandwidth satisfaction rate of each terminal, and map the service bandwidth to the channel bandwidth in combination with the spectrum efficiency. Specifically, after receiving the global scheduling result, each interactive network performs fine-grained scheduling within its local available bandwidth resource pool, using the quota limit specified therein as the boundary.

[0029] According to the terminal group priority, process the deterministic requirements of each terminal group; The system extracts unused bandwidth due to insufficient demand from some terminal groups and aggregates it into a public resource pool. For this public resource pool, the system performs idle bandwidth allocation, prioritizing idle resources for terminal groups whose demand has not been met.

[0030] During this allocation process, the ratio between the actual bandwidth acquired by each terminal and its original demand is recorded; this ratio is the terminal bandwidth satisfaction rate. After completing the bit rate allocation at the service level, the modulation and coding scheme corresponding to the current link environment of each terminal is obtained.

[0031] Based on a specific modulation order and coding rate, the corresponding spectral efficiency is extracted. The allocated bit rate value is then divided using the spectral efficiency to obtain the actual Hertz frequency range occupied by the physical layer, i.e., the channel bandwidth.

[0032] In some embodiments, the group-level channel bandwidth can be divided strictly according to the proportional ratio of the service bandwidth of each group; while when allocating channel bandwidth to terminals within a group, independent weighted calculation is performed based on the spectral efficiency of each terminal to ensure that channel resources are appropriately tilted towards terminals with high coding efficiency and to improve the overall spectrum utilization.

[0033] Step 105: Determine the terminal congestion state based on the terminal bandwidth satisfaction rate, and perform carrier handover for terminals in the terminal congestion state based on competition benefit prediction. In this embodiment, the changes in the terminal bandwidth satisfaction rate are continuously monitored.

[0034] When the ratio remains below a certain preset benchmark, the system determines that the local frequency band where the terminal is located is in short supply and marks it as being in a congested state.

[0035] Unlike traditional frequency hopping mechanisms that rely solely on overall carrier load, this scheme introduces a competition benefit prediction based on an individual perspective. It assesses the existing competition environment on candidate target carriers, calculates the actual resource share a congested terminal could acquire after switching to the target carrier, and infers the expected improvement in satisfaction rate based on this. Only when the expected improvement offsets the service interruption costs caused by the switch and still results in a positive benefit will the system trigger the physical layer carrier switching command.

[0036] As an anti-jitter mechanism, a cooling timer is started after the carrier handover is performed. Before the timer expires, the terminal will be prohibited from initiating another handover request, thereby avoiding a network-wide signaling storm caused by the terminal frequently hopping between multiple available carriers.

[0037] Step 106: Generate a negative feedback adjustment signal based on the terminal group bandwidth satisfaction rate of the historical period, and update the terminal group configuration weight.

[0038] Specifically, the system archives and statistically analyzes the bandwidth satisfaction rate of each terminal group and the actual allocated bandwidth and required bandwidth of each terminal within multiple consecutive scheduling cycles in the background. The system calculates the weighted average satisfaction level of all terminal groups in the entire network and uses it as a benchmark.

[0039] For each specific terminal group, if its historical satisfaction rate is lower than the network-wide baseline, it indicates that it has been at a disadvantage in past resource competition.

[0040] At this point, based on the difference between its deviation from the reference line, a positive negative feedback adjustment signal is generated to increase the initial static weight of the terminal group. Conversely, if its satisfaction rate is higher than the network-wide reference line, a negative negative feedback adjustment signal is generated to reduce its weight.

[0041] The adjusted signal value is overwritten into the system's configuration database and used as the input parameter for the next global scheduling cycle, i.e., the updated terminal group configuration weight.

[0042] Understandably, through the aforementioned negative feedback closed-loop mechanism, the originally static manually configured parameters are transformed into dynamic variables that evolve adaptively with the network situation, making the resource allocation among multi-level terminal groups approach dynamic equilibrium over a long time scale.

[0043] Example 2 describes the construction rules of the multi-level architecture in the satellite communication system, the node update mechanism of dynamic events, and the global quota scheduling logic to prevent demand inflation.

[0044] like Figure 2 and Figure 3 As shown, in one possible embodiment, a two-level bandwidth management architecture is constructed, and mutually decoupled terminal group priorities and terminal priorities are configured, including: Step 201: Construct a global bandwidth management tree that is effective across networks and an interactive network bandwidth management tree that is effective within a single interactive network; In this embodiment, the global bandwidth management tree is a multi-layered tree structure with the total bandwidth of the satellite system as the root node. This structure is divided from top to bottom into a root node, primary terminal group nodes, optional secondary terminal group nodes, and leaf nodes.

[0045] The number of levels in the global bandwidth management tree is determined by the nesting depth of the terminal group, typically supporting 2 to 4 levels of expansion. The interactive network bandwidth management tree is the physical projection of the global bandwidth management tree onto a single interactive network, with its root node configured with the available bandwidth of that interactive network.

[0046] By constructing these two types of tree structures, a two-layer logical channel for global resource coordination and local resource execution is established, alleviating the problem of lack of global bandwidth control in previous solutions.

[0047] In some alternative implementations, when facing large-scale network expansion, multiple levels of sub-group nodes can be added under the primary terminal group node. The cascading characteristics of the tree structure can be used to achieve unlimited hierarchical division, thereby adapting to the complex hierarchical communication network topology needs of enterprises or organizations.

[0048] Step 202: Configure global quality of service nodes and interactive quality of service nodes in each level of the global bandwidth management tree and the interactive network bandwidth management tree; Specifically, the global Quality of Service (QoS) node is attached to the global bandwidth management tree and is used to manage global bandwidth parameters that apply across all interactive networks. The interactive network QoS node is attached to the interactive network bandwidth management tree and applies only within a single network, used to further subdivide bandwidth for specific areas within the global quota.

[0049] In node configuration management, the system supports setting three policy levels, with priority from high to low: global configuration, interactive network independent configuration, and interactive network general configuration.

[0050] When the configuration parameters of the parent node change, if the child node has not been configured with the corresponding parameters separately, it will automatically inherit the new configuration value of the parent node; if the child node has been configured separately, it will remain unchanged, but the system will enforce that its configuration value must not exceed the upper limit of the parent node, and will automatically prune it to the upper limit of the parent node if it exceeds the limit.

[0051] Furthermore, to ensure the robustness of the configuration strategy, the system sets a timeout period for the globally distributed scheduling results. When a link interruption causes a timeout and the configuration is not re-distributed, the interactive network will automatically degrade and restore to the general configuration of the interactive network, striving to ensure that local communication services are not interrupted due to the failure of the global control plane.

[0052] Step 203: Mount the terminal as a leaf node to the corresponding global quality of service node or the interactive network quality of service node; In this embodiment, the system mandates that physical terminals can only be attached to the very end of the tree structure, i.e., exist as leaf nodes, and prohibits a terminal from being attached to both group nodes and non-group nodes simultaneously to participate in peer competition.

[0053] To address the highly dynamic characteristics of satellite networks, a dynamic tree structure update mechanism is designed.

[0054] Once a terminal completes its network registration on an interactive network, it locates its assigned Quality of Service (QoS) node based on its configuration information and creates leaf nodes under that node.

[0055] When a terminal leaves the network, the system removes the terminal from the interactive network bandwidth management tree. To prevent frequent tree structure rebuilding due to brief link interruptions, if a service quality node has no child nodes after removal, the system marks it as idle and starts a preset idle timeout timer.

[0056] As an improvement to the above steps, the threshold of the idle timeout timer can be set to 300 seconds. The system will only delete the node if no terminal is mounted after the timer is cleared.

[0057] For terminals migrating across networks, their affiliation to the global quality of service node on the global bandwidth management tree remains unchanged. The system only executes the de-network process in the original network and the network entry process in the new network in parallel, thereby achieving seamless logical roaming.

[0058] Step 204: Based on the hierarchical isolation structure of the global bandwidth management tree and the interactive network bandwidth management tree, separate the bandwidth competition between terminal groups from the terminal bandwidth competition within the terminal group, so as to decouple the terminal group priority from the terminal priority. Existing bandwidth allocation schemes often suffer from the logical flaw of allowing terminals and terminal groups to compete at the same level, which can easily lead to high-priority terminals within a group occupying bandwidth quotas from other groups. Therefore, this embodiment establishes a mandatory contention blocking mechanism that prioritizes groups over terminals.

[0059] In the first phase, only the terminal group priority configured for each group is used to perform group-level quota allocation in the global resource pool. In the second phase, within each group, only the terminal priority configured for each terminal is used to perform a secondary allocation of the acquired group-level quota. The secondary allocation logic for terminal bandwidth within a group has the same hierarchical processing structure as the aforementioned inter-group allocation logic. The system sorts terminals according to their configured priorities, prioritizing the fulfillment of the committed bandwidth requirements of high-priority terminals, and then fulfilling the maximum bandwidth requirement as needed. When the total committed bandwidth requirement within a group exceeds the group-level quota, the system uses the same maximum-minimum fair allocation algorithm as in step 402 to fairly allocate the quota among terminals within the group.

[0060] This decoupling mechanism helps ensure that sudden high demands from terminals only compete within the quota of their respective terminal groups, improving the isolation and fairness of multi-level network resource allocation.

[0061] Step 205: Aggregate the global bandwidth requests layer by layer from the leaf nodes to the root node of the global bandwidth management tree, compare the configuration limit of the non-leaf node with the sum of the requirements of all its child nodes, and extract the smaller value of the two as the node aggregation requirement of the corresponding node. In order to reduce the imbalance of the entire network scheduling caused by malicious false reporting of demand by the underlying nodes when aggregating demand across the entire network, a demand inflation prevention statistical rule was introduced.

[0062] For leaf nodes, their effective demand is directly equal to the bandwidth demand actually reported by the terminal; for non-leaf nodes, the system uses a recursive comparison algorithm.

[0063] Extract the configuration limit set by the administrator of the non-leaf node, and calculate the sum of the valid requirements of all its child nodes.

[0064] Perform a strict minimum value operation, specifically: The node aggregation requirement D for non-leaf nodes node_eff =min(C node_config ,D children_sum ); Among them, C node_config D is the preset configuration limit for this non-leaf node in the system. children_sum This is the sum of the effective demands of all child nodes under the non-leaf node, and min is the function to find the minimum value.

[0065] This statistical rule ensures that any valid request at any level passed to the root node does not exceed the physical constraint limit set by the administrator, nor is it greater than the sum of the actual physical demands of the terminals under its jurisdiction, which helps to ensure the rigor of global bandwidth request data.

[0066] Step 206: Based on the node aggregation requirements, allocate global committed bandwidth quotas and global maximum bandwidth quotas to nodes at all levels in a hierarchical order that prioritizes meeting the committed bandwidth requirements in the global bandwidth application and then meets the maximum bandwidth requirements in the global bandwidth application. Specifically, after obtaining the network-wide node aggregation requirements, the system initiates global bandwidth allocation.

[0067] The total network resource budget is calculated as the sum of the available bandwidth of all interactive networks. The allocation process is strictly divided into two phases.

[0068] In the first phase, the system allocates the committed bandwidth portion of the node aggregation demand equally and fairly according to the priority of each terminal group, generating the global committed bandwidth quota.

[0069] In the second phase, the system deducts the capacity allocated in the first phase, extracts the remaining total budget of network resources, and performs weighted allocation on the maximum bandwidth increment in the node aggregation demand according to the weight values ​​set for each terminal group, thereby generating the global maximum bandwidth quota.

[0070] Understandably, this hierarchical order helps to ensure that the basic guaranteed rate for all terminal groups is established with the highest priority at the network level, and then the relatively abundant bandwidth is allocated to improve the user experience according to management weight.

[0071] The CIR (Cost-Incremental Bandwidth) characterizes the basic guaranteed rate, while the MIR (Maximum Bandwidth) characterizes the peak rate, including incremental increases.

[0072] Step 207: For the same terminal group spanning multiple interactive networks, according to the demand ratio reported by each interactive network in the global bandwidth application, the global committed bandwidth quota and the global maximum bandwidth quota of the terminal group are decomposed across networks. Because certain global-level terminal groups may have their subordinate terminals physically distributed across multiple different satellite beams or interactive networks, the obtained global committed bandwidth quota is a total amount across networks. To allocate this total amount to specific networks, the system needs to perform quota decomposition calculations.

[0073] Extract the actual committed bandwidth demand reported by the terminal group in a specific interactive network, divide it by the sum of the actual committed bandwidth demands reported by the terminal group in all interactive networks across the entire network, and obtain a proportional coefficient.

[0074] The specific allocated bandwidth is calculated by multiplying the global committed bandwidth quota by the proportional coefficient. The specific calculation process is implemented using the following formula: Q _g_n_CIR =Q _g_CIR *(D _g_n_CIR / D _g_total_CIR ); Among them, Q _g_n_CIR Q is the locally committed bandwidth quota allocated to the g-th terminal group on the n-th interactive network. _g_CIR For the globally committed bandwidth quota of the g-th terminal group, D _g_n_CIR D represents the committed bandwidth requirement actually reported by the terminal group in the nth interactive network. _g_total_CIR This represents the total committed bandwidth demand of this terminal group across the entire network.

[0075] Furthermore, the system also adopts the same proportional decomposition logic based on demand for the global maximum bandwidth quota, which helps to strictly match the real-time business pressure of each physical region when allocating resources.

[0076] Step 208: Compare the total quota after decomposition in each interactive network with the available bandwidth of the interactive network itself. When the total quota exceeds the available bandwidth, perform proportional pruning on each quota in the interactive network and recover the quota difference released by pruning to the global resource pool to supplement the subsequent global idle resource reallocation. When the total quota does not exceed the available bandwidth, directly execute the issuance according to the decomposed quota.

[0077] Since the global allocation in the aforementioned steps is performed within a macro-level pool of the total network resource budget, quota decomposition across networks may lead to localized network physical capacity overload. Therefore, the system introduces a capacity over-allocation verification module at each interactive network layer.

[0078] The decomposed quotas of all terminal groups falling into a certain interactive network are summed to obtain the total quota. If the total quota is greater than the actual physical channel available bandwidth of the interactive network, the system initiates a pruning mechanism.

[0079] Calculate the ratio of available bandwidth to the total quota, and use this ratio to perform a uniform proportional reduction multiplication operation on each quota in the network, so that the issued quota can be carried by the physical channel.

[0080] Correspondingly, the bandwidth difference stripped away by the proportional throttling will not be discarded by the system. Instead, it will be reported and incorporated into the global resource pool of the central system. The recovered resources will be used in subsequent stages to compensate the remaining unsatisfied terminal groups across the network, in order to maintain the conservation of satellite network resources at both the mathematical and physical levels.

[0081] Step 209: Aggregate the global committed bandwidth quota and the global maximum bandwidth quota according to the interactive network dimension to generate the global scheduling result to be distributed to each interactive network.

[0082] In this embodiment, after rigorous anti-expansion statistics, cross-network demand decomposition, and physical capacity pruning verification, all quota data has met the conditions for conflict-free execution. Subsequently, using different physical interactive networks as index keys, the globally committed bandwidth quotas and the globally maximum bandwidth quotas belonging to the same network are structurally encapsulated. The encapsulated data set is the global scheduling result, which is pushed to each subordinate network by the global network management system through control signaling channels, serving as the constraint basis for subsequent local multi-round fine-grained allocation.

[0083] Example 3 describes the specific processing steps for achieving high-precision data collection and demand compression through a preset percentile value extraction and local pre-allocation mechanism before reporting terminal bandwidth requirements to the global scheduling layer.

[0084] In one possible embodiment, the process includes the following: Step 301: Extract the peak value of the real-time traffic of the terminal within the collection period as the initial terminal bandwidth requirement.

[0085] In one embodiment, specifically: the preset high percentile value of the real-time traffic of the terminal within the collection period is extracted as the initial terminal bandwidth requirement; The preset percentile value mentioned in this invention refers to a statistical value that can represent the normal peak of terminal business within the statistical period, including but not limited to the 95th percentile value, the 99th percentile value, or the maximum value within the statistical period.

[0086] Specifically, the monitoring unit of the interactive network acquires the real-time data transmission rate of each physical terminal according to a set acquisition period. In order to filter out transient spikes caused by network jitter, the distribution characteristics of the real-time traffic sequence are extracted over a relatively long statistical period.

[0087] All traffic samples within the statistical period are sorted in ascending order, and the values ​​at specific percentiles are calculated. The system selects the preset percentile value corresponding to the 95th percentile as the normal peak value, while discarding extreme values ​​in the top 5%. The extracted 95th percentile value is then identified as the initial terminal bandwidth requirement.

[0088] In another alternative, if the physical link environment of the terminal is extremely stable, the system can also be configured to extract the 99th percentile value or directly take the maximum value within the statistical period as the terminal's bandwidth requirement, in order to adapt to leased line service scenarios with more stringent bandwidth guarantee requirements.

[0089] Step 302: Obtain the pre-configured committed bandwidth limit and maximum bandwidth limit; perform local pre-allocation of the terminal bandwidth demand within the interactive network based on the committed bandwidth limit and the maximum bandwidth limit; filter out the demand portion exceeding the limit configuration to obtain the effective bandwidth demand. In this embodiment, the system locally stores the policy parameters of the QoS nodes to which each terminal belongs on the interactive network. The system obtains the pre-configured committed bandwidth limit and maximum bandwidth limit, and uses the committed bandwidth limit and the maximum bandwidth limit to construct a digital filtering mechanism.

[0090] The initial terminal bandwidth requirement is compared with the aforementioned upper limit parameter. This comparison operation is local pre-allocation, and its purpose is to compress the demand and generate a global request volume, rather than to perform actual underlying physical traffic control.

[0091] The portion of the terminal bandwidth requirement that does not exceed the configured upper limit is retained, while any portion exceeding the committed bandwidth upper limit or the maximum bandwidth upper limit is truncated and discarded. The sequence output after this numerical comparison and truncation operation is the effective bandwidth requirement for upward transmission.

[0092] Furthermore, this local pre-allocation mechanism does not involve the allocation of channel idle bandwidth, and its pre-allocation result is not used as the basis for the final channel bandwidth allocation, nor will it send the pre-allocation result to the hardware execution unit of the communication terminal.

[0093] By performing this pre-screening operation at the edge of the interactive network, unreasonable and excessive demands that exceed local configuration constraints can be directly filtered out, thereby avoiding the reporting of redundant data to the global network management system and reducing the signaling transmission overhead of cross-level communication links.

[0094] Step 303: Summarize the effective bandwidth requirements in each interactive network and generate the global bandwidth request.

[0095] Specifically, each interactive network will aggregate the effective bandwidth requirements obtained after local pre-allocation. The system will categorize and summarize the data according to the physical layer attributes of the terminals and the logical dimension of their respective terminal groups.

[0096] The terminal's unique identifier, the globally requested committed bandwidth value calculated during the pre-allocation phase, the globally requested maximum bandwidth value, the interactive network identifier to which the terminal currently belongs, and the associated global quality of service group identifier are extracted, encapsulated, and integrated.

[0097] The data set containing the above parameters is converted into an internal standard signaling format to generate the global bandwidth request.

[0098] Accordingly, the generated global bandwidth request is periodically reported to the information collection module of the global network management system via a specific control channel.

[0099] The global network management system parses the structured data reported by each interactive network and stores it in an in-memory database, which serves as a discrete input parameter for subsequent calculation of global bandwidth resource quotas across networks.

[0100] Example 4 describes how, after receiving data distributed globally, the interactive network sequentially performs initial resource allocation based on demand and weight locally, and how it completes the skewed allocation of remaining system capacity by introducing a satisfaction rate balancing mechanism that reflects dynamic resource gaps, thereby overcoming the uneven allocation defects caused by a single static weight.

[0101] like Figure 4 As shown, in one possible embodiment, the following steps are included: Step 401: Based on the quota limit provided by the global scheduling result, execute the committed bandwidth on-demand allocation round and the maximum bandwidth on-demand allocation round in sequence to establish the initial service bandwidth quota; In this embodiment, the bandwidth management controller of the interactive network first extracts the quota limit parameters assigned to each terminal group from the global scheduling result, and integrates the local quality of service configuration policy to divide the bandwidth allocation process into four sequentially executed rounds.

[0102] The first two basic rounds aim to establish the initial service bandwidth quota for each terminal group. In the round of committed bandwidth allocation on demand, the system uses the priority of the terminal group as the sorting criterion, prioritizing the data transmission and reception needs of basic, guaranteed communication services. In the maximum bandwidth on-demand allocation round, the system uses the management weight set by the terminal group as a guiding coefficient to perform resource allocation for incremental demands used to improve communication rates.

[0103] Furthermore, in the two basic rounds mentioned above, if the sum of the actual demands reported by each terminal group is less than the currently available physical transmission capacity of the interactive network, then the corresponding demand amount will be issued in full. If a resource shortage occurs that prevents the full fulfillment of the needs of all terminal groups, the system will switch to a controlled equal or weighted cutoff process, which is the subsequent anti-over-allocation mechanism.

[0104] Step 402: When the promised bandwidth allocation round is executed and the remaining available bandwidth is insufficient, the bandwidth demand of each terminal group is sorted from smallest to largest and an equal share is calculated. For terminal groups whose demand is less than or equal to the equal share, the demand is fully satisfied. For terminal groups whose demand is greater than the equal share, fair allocation is performed according to the equal share to obtain the promised bandwidth allocation result for each terminal group. When there is sufficient remaining available bandwidth in the interactive network, the committed bandwidth requirements of each terminal group will be fully met. Specifically, when the total committed bandwidth demand exceeds the network's carrying capacity, the system triggers the maximum-minimum fair allocation algorithm to prevent high-demand groups from excessively crowding out resources.

[0105] Obtain the bandwidth requirements of all terminal groups participating in the current level of competition, and arrange them in ascending order to form a monotonically increasing sequence of requirements.

[0106] Extract the total available resources for the current level, divide it by the number of terminal groups participating in the competition, and obtain a preliminary equal share value.

[0107] The system performs an iterative comparison operation. The system compares the demand of the terminal group at the top of the ranking with the equivalent share. If the former's value is less than or equal to the latter's, the system determines that the terminal group's demand can be fully met, issues an allocation equivalent to its original demand, subtracts the issued amount from the total available resources, updates the number of remaining terminal groups, and recalculates the new equivalent share.

[0108] If the comparison finds that the demand of a certain terminal group is greater than the current updated equal share, the system determines that the resources have entered the scarcity threshold. For that terminal group and all terminal groups in the sequence that are ranked after it, the system forcibly distributes a bandwidth value equal to the equal share, thereby performing strict fair allocation and terminating the iterative calculation of the current level.

[0109] Step 403: When executing the maximum bandwidth on-demand allocation round, extract the remaining available bandwidth after completing the promised bandwidth on-demand allocation round, divide the theoretical allocation amount according to the terminal group configuration weight ratio of each terminal group, perform quota pruning for terminal groups whose theoretical allocation amount exceeds their actual maximum bandwidth requirements, and iteratively redistribute the excess bandwidth released by pruning in terminal groups whose requirements are not met until no pruning occurs or the available bandwidth is exhausted, and obtain the maximum bandwidth allocation result of each terminal group; In this embodiment, the data source used to allocate the maximum bandwidth increment requirement is the remaining available bandwidth after the previously committed bandwidth has been allocated.

[0110] Since maximum bandwidth typically corresponds to non-guaranteed services, a strict equal-quota approach is no longer used; instead, a weighted mechanism is introduced. The system obtains the pre-set configuration weight values ​​for each terminal group and calculates the sum. Then, it multiplies the configuration weight ratio of each terminal group by the aforementioned remaining available bandwidth to output the theoretical allocation for each terminal group.

[0111] In most cases, the theoretical allocation calculated based on the configuration weights of pure terminal groups may deviate from the actual data transmission needs of the terminal groups. The theoretical allocation for each terminal group is compared with its reported actual maximum bandwidth requirement. For terminal groups whose theoretical allocation exceeds their actual maximum bandwidth requirement, the system performs a quota pruning operation, issuing only the quota equal to their actual demand, and extracting and aggregating the bandwidth saved due to pruning into the available resource pool.

[0112] For terminal groups whose actual needs are still not met, the system re-allocates the aforementioned surplus bandwidth according to their relative terminal group configuration weight ratios. This pruning and compensation process iterates frequently within the system until no terminal group is pruned or the available bandwidth decreases to zero, thereby maximizing network resource utilization.

[0113] Step 404: After the promised bandwidth on-demand allocation round and the maximum bandwidth on-demand allocation round are completed, extract the allocation quota that has not been used due to the actual demand being lower than the quota limit, and release the allocation quota to the public resource pool of the interactive network to form idle bandwidth. Specifically, after the first and second rounds of precise segmentation, each terminal group has obtained a basic quota that matches its original reported data.

[0114] At this time, because the actual service traffic generated by some terminal groups is lower than the upper limit of resource configuration parameters set by the global network management system, unused channel resources will remain in the system.

[0115] Extract the configuration limit value of each terminal group, subtract the actual allocation quota value it received in the on-demand allocation round, and the calculated positive difference is the unused allocation quota.

[0116] In addition, if the total demand of the entire network is less than the total physical network capacity when the global network management system issues the scheduling results, a portion of network-level unallocated bandwidth will also be generated at the global scheduling level that is not issued to any terminal group.

[0117] The unused allocated quotas within each terminal group, along with the unallocated bandwidth at the network level, are summed and transferred to the public resource pool node of the interactive network, thus generating idle bandwidth for the third and fourth rounds of supplementary scheduling.

[0118] Step 405: Identify terminal groups whose actual bandwidth demand is not fully met by the initial service bandwidth quota, and treat them as terminal groups in a state of unmet demand. Based on the idle bandwidth in the public resource pool, execute the idle committed bandwidth reallocation round and the idle maximum bandwidth reallocation round in sequence. In this embodiment, the idle bandwidth data that has been aggregated is read from the public resource pool, and a secondary resource scheduling mechanism is initiated. The recipients at this stage are strictly limited to terminal groups whose actual business needs were not fully allocated in previous rounds, i.e., terminal groups in a state of unmet demand.

[0119] Prioritize initiating supplementary calculations for the committed bandwidth gap, i.e., the idle committed bandwidth reallocation round; if there is still a surplus in the public resource pool after processing, initiate the idle maximum bandwidth reallocation round to fill the maximum bandwidth gap.

[0120] Step 406: Based on the allocated bandwidth and corresponding bandwidth requirements of each terminal group, calculate the bandwidth satisfaction rate of each terminal group in the current allocation round. Specifically, to avoid the polarization caused by the traditional static weight allocation model, the system introduces a dynamic measurement indicator. Before allocating any idle resources, the hunger level of each participating entity must be accurately calculated. The system extracts the allocated bandwidth value accumulated by the terminal group in previous rounds as the numerator, and extracts the total actual bandwidth demand reported by all hierarchical child nodes as the denominator. Dividing the two yields a small value between 0 and 1. This small value is the bandwidth satisfaction rate of the terminal group. This parameter objectively reflects the relative proportion of each physical node's current service transmission intentions being satisfied.

[0121] According to one aspect of this application, the resource gap corresponding to the bandwidth satisfaction rate of the terminal group is coupled and mapped with the pre-acquired configuration weight of the terminal group to generate a satisfaction rate balancing factor for each terminal group, as follows: Step 407: The difference between the value of one and the current group bandwidth satisfaction rate is taken as the resource gap, and the resource gap is subjected to exponential operation using the pre-configured fairness adjustment index; In this embodiment, the system derives a dynamic variable reflecting the urgency of demand based on the bandwidth satisfaction rate calculated above. A subtraction operation is performed, subtracting the bandwidth satisfaction rate from a constant 1, and the resulting difference is defined as the resource gap for the terminal group.

[0122] Furthermore, to enable the system to provide non-linear, sensitive feedback to varying degrees of resource scarcity, the system extracts a fairness adjustment index preset in the database by operations and maintenance personnel. The system then performs a power operation using the resource shortage as the base and the fairness adjustment index as the exponent. The choice of the power and the base can be optimized by engineers based on actual circumstances.

[0123] In some optional implementations, the fairness adjustment index has a predetermined system control significance. When the system configures this index to gradually increase, the computational results have a significantly amplified multiplier effect on terminal groups with extremely low satisfaction rates. In a typical high-throughput satellite communication system, in this embodiment, the fairness adjustment index is set to 1 to achieve a balance between algorithm efficiency and fairness convergence.

[0124] Step 408: Multiply the result of the exponentiation operation by the terminal group configuration weight to obtain the satisfaction rate balancing factor that couples the configuration weight with the dynamic gap. In order to form a fusion allocation parameter that both respects the top-level will of the administrator and adapts to the dynamic fluctuations of the underlying layer, the system extracts the initial static configuration weight parameters of the stored terminal groups.

[0125] The result of the exponentiation operation output in the previous step is multiplied by the configured weight of the terminal group. The final product output constitutes the satisfaction rate balancing factor. Because the calculation of this factor incorporates the real-time service dissatisfaction level, the system automatically grants terminal groups that have not obtained sufficient transmission opportunities for a long time a relatively higher bidding capability. The calculation process in this step is implemented using the following formula: ω _g = w _g * (1 - η _g ) β ; Where, ω _g w is the satisfaction rate balancing factor for the g-th terminal group. _g Configure weights for the g-th terminal group pre-acquired by the system, with initial values ​​set by the administrator as w. _g_0 Further updates will follow. _g β is the bandwidth satisfaction rate of the g-th terminal group calculated above, and β is the fairness adjustment index set above.

[0126] In some embodiments, the total available idle bandwidth in the current public resource pool of the system is assumed to be 5.0. There are currently three terminal groups competing for resources within the system, denoted as terminal group A, terminal group B, and terminal group C. After calculations in the previous on-demand allocation rounds, the bandwidth satisfaction rate of terminal group A is 0.5, the bandwidth satisfaction rate of terminal group B is 0.667, and the bandwidth satisfaction rate of terminal group C is 0.6. The initial terminal group configuration weights are set to 3, 2, and 1, respectively. The system's pre-set fairness adjustment index is set to 1. The parameters of each terminal group are calculated according to the above linear formula calculation mechanism. The satisfaction rate balancing factor for terminal group A is calculated as 3*(1-0.5)=1.5. The satisfaction rate balancing factor for terminal group B is calculated as 2*(1-0.667) = 0.666. The satisfaction rate balancing factor for terminal group C is calculated as 1*(1-0.6) = 0.4. The system sums the satisfaction rate balancing factors for the three terminal groups to obtain 1.5 + 0.666 + 0.4 = 2.566. Based on this, the subsequent allocation share is determined according to the numerical proportion of each factor. The final distribution amount obtained by terminal group A is 5.0 * (1.5 / 2.566) = 2.92.

[0127] Without adjusting the original physical weights, terminal group A, whose initial bandwidth satisfaction rate is at its absolute minimum, automatically receives more resource allocation, which is beneficial to improving the adaptive equilibrium convergence of state indicators among multi-level terminal groups.

[0128] Step 409: Based on the satisfaction rate balancing factor of each terminal group, determine the redistribution ratio of the idle bandwidth in the public resource pool, and allocate the idle bandwidth preferentially to terminal groups whose bandwidth satisfaction rate is lower than a set threshold according to the redistribution ratio, thereby obtaining the supplementary idle bandwidth quota for each terminal group. This ensures that the idle bandwidth is preferentially allocated to terminal groups with relatively lower bandwidth satisfaction rates.

[0129] Specifically, a dynamic reference array is constructed using the satisfaction rate balancing factors of all unmet demand terminal groups obtained through calculation.

[0130] The system sums all the factor values ​​in the array above to obtain a normalized sum. It then extracts the satisfaction rate balancing factor for a single terminal group and divides it by this normalized sum to calculate the redistribution ratio for that specific terminal group.

[0131] The idle bandwidth stored in the public resource pool is extracted and divided and distributed according to the aforementioned redistribution ratio. Through this closed-loop operation, terminal groups whose bandwidth satisfaction rate has severely declined due to a surge in business volume can automatically intercept and absorb idle physical resources in the system until their satisfaction rate catches up with the average level of the entire network.

[0132] Example 5 describes how to break down the barrier between the network layer and the physical layer in a satellite communication system. By combining the dynamically changing modulation and coding scheme of the terminal, the service bandwidth based on the data transmission rate is accurately mapped to the channel bandwidth based on the frequency occupancy, and the specific physical execution logic of reverse service pruning is executed when the link deteriorates.

[0133] In one possible embodiment, the following steps are included: Step 501: Obtain the modulation and coding scheme corresponding to the current link state of each terminal, and extract the spectral efficiency associated with the modulation and coding scheme according to the pre-configured lookup table; Because satellite communication system terminals operate in mobile or climate-changing environments, the signal-to-noise ratio (SNR) of their physical links fluctuates in real time. The adaptive coding and modulation controller of the underlying physical system continuously collects SNR estimates for each communication terminal.

[0134] When the signal-to-noise ratio changes, the adaptive coding and modulation controller dynamically switches the modulation and coding scheme currently used by the terminal. To achieve the conversion between network layer bit rate and physical layer frequency, a standard mapping table is pre-stored in the memory of the global network management system and each interactive network.

[0135] Using the extracted modulation and coding scheme as an index key, a search is performed in the lookup table to obtain a theoretical value that characterizes the data carrying capacity per unit frequency band.

[0136] Specifically, the aforementioned reference table is constructed based on common satellite communication standards, listing theoretical spectral efficiency reference values ​​for different combinations of modulation order and coding rate. For example, a modulation and coding scheme using quadrature phase shift keying (QPSK) with a coding rate of 1 / 4 has a theoretical spectral efficiency of 0.49. When calculating the spectral efficiency for practical applications, the system introduces a roll-off factor for the physical channel to correct for theoretical errors.

[0137] The spectral efficiency ξ in actual terminal applications _actual (m_t) =ξ(m_t) / (1 + ρ); Where ξ(m_t) is the theoretical spectral efficiency obtained by looking up the table, and ρ is the roll-off coefficient of the physical system.

[0138] By configuring the specific value of the roll-off factor, the system can output more accurate actual occupied spectrum parameters.

[0139] Step 502: In the group-level channel allocation stage, the channel capacity of the current carrier obtained is divided into group-level channel bandwidths in proportion to the service bandwidth ratio of each terminal group. After completing the logical bandwidth allocation of the network layer in the early stage, the abstract total bit rate needs to be implemented on the physical Hertz spectrum of the satellite carrier.

[0140] To prevent terminal groups with poor overall link conditions from consuming excessive physical channel resources and crowding out available frequencies for other groups, the system adopts a strict principle of fair service bandwidth ratio at the terminal group level.

[0141] The system extracts the inherent total channel capacity of a single carrier and obtains the service bandwidth quota previously allocated to each terminal group. The system then divides the service bandwidth quota of a specific terminal group by the sum of the service bandwidth quotas of all terminal groups on that carrier to calculate a group-level quota ratio coefficient.

[0142] By multiplying the total channel capacity of the carrier with the quota ratio coefficient of the group level, the physical frequency limit that the specific terminal group is allowed to use, i.e., the group-level channel bandwidth, is obtained.

[0143] Step 503: In the terminal-level channel allocation stage, within the group-level channel bandwidth, the service bandwidth is weighted and calculated using the spectral efficiency corresponding to each terminal to complete the channel bandwidth allocation for terminals within the group.

[0144] Specifically, after dividing the physical frequency limits of each terminal group, frequency bands are allocated to specific communication entities within each terminal group.

[0145] Since different physical terminals within the same terminal group may use different modulation and coding schemes due to differences in geographical location or antenna size, the system performs accurate conversion at the terminal level.

[0146] Extract the specific service bandwidth value previously allocated to a terminal and divide it by the actual spectral efficiency required by that terminal. The calculated quotient is the channel bandwidth actually needed by that terminal.

[0147] Through the above mechanism, the system achieves dynamic resource adaptation at the physical layer. Terminals using higher-order modulation and coding schemes and with good signal-to-noise ratios will occupy less channel bandwidth resources while obtaining the same service bandwidth. This mechanism facilitates fair allocation based on service logic, avoiding excessive penalties for terminals with low signal-to-noise ratios.

[0148] As an important supplement and anomaly handling mechanism to this embodiment, when the terminal encounters deteriorating weather conditions that cause a significant drop in the signal-to-noise ratio, its underlying modulation and coding scheme will undergo a down-order switch, resulting in a sharp decrease in its corresponding spectral efficiency. If its originally allocated service bandwidth remains unchanged, the channel bandwidth required for its conversion will increase exponentially.

[0149] At this point, the system performs a dynamic check of the remaining carrier channel capacity. If the remaining carrier channel bandwidth is insufficient to accommodate the new demand calculated by the terminal, the system will initiate a reverse pruning protection procedure.

[0150] Extract the maximum available channel bandwidth that the terminal is allowed to acquire on the carrier, and multiply it by the reduced-order updated spectral efficiency. The product is the actual service bandwidth of the terminal after pruning.

[0151] This reverse pruning operation effectively blocks the risk of carrier channel overload caused by the deterioration of a single terminal link. The pruned difference will be recorded as the unmet demand of the terminal to trigger the subsequent congestion determination and carrier switching assessment process of the system.

[0152] Example 6 describes how, in a satellite communication network, the system accurately identifies restricted terminals by constructing a composite congestion index, and how it uses a competitive discount factor to predict the actual available capacity of the target carrier, thereby completing the specific execution process of frequency hopping net benefit assessment and anti-oscillation cooling control.

[0153] In one possible embodiment, the following steps are included: Step 601: Extract the terminal bandwidth satisfaction rate in the committed bandwidth dimension and the maximum bandwidth dimension of the local bandwidth allocation, respectively; By using preset weighting coefficients to weight and fuse the terminal bandwidth satisfaction rates of the two dimensions mentioned above, a composite congestion index is calculated to characterize the degree of terminal limitation. When the composite congestion index is greater than the preset congestion threshold, the terminal is determined to be in a congested state and is added to the handover queue.

[0154] Specifically, the system obtains the satisfaction rate of the terminal at the current moment in the committed bandwidth dimension and the satisfaction rate in the maximum bandwidth dimension. The system pre-configures different weighting coefficients, for example, setting the weighting coefficient for the committed bandwidth dimension to 0.7 and the weighting coefficient for the maximum bandwidth dimension to 0.3. The system obtains the corresponding gap ratio by subtracting the satisfaction rate of each dimension from 1, and then multiplies the above gap ratios by the corresponding weighting coefficients and sums them to calculate and generate the composite congestion index.

[0155] Terminal Composite Congestion Index (CI) _terminal = α _1 * (1 - η _t_CIR ) + α _2 * (1 - η _t_MIR ); Where, α _1 η is the weighting coefficient for the committed bandwidth dimension. _t_CIR For the terminal bandwidth satisfaction rate in the committed bandwidth dimension, α _2 η represents the weighting coefficient for the maximum bandwidth dimension. _t_MIR The terminal bandwidth satisfaction rate is the maximum bandwidth dimension.

[0156] Furthermore, the system compares the calculated composite congestion index with a preset congestion threshold (0.3~0.5) for determination. When the composite congestion index of the terminal is greater than the congestion threshold, the system determines that the frequency band resources acquired by the terminal are severely limited, i.e., the terminal is in a congested state, and immediately stores its identifier in the handover queue for processing. Through a micro-determination mechanism based on the individual terminal satisfaction rate, the system can effectively identify hidden congestion scenarios where the overall carrier load is low but individual terminal resources are severely insufficient.

[0157] Step 602: For the candidate target carrier, count the number of existing terminals currently in an unmet demand state on the target carrier, and calculate the competition discount factor to characterize the difficulty of obtaining the remaining resources. The remaining capacity of the target carrier is calculated and evaluated using the competition discount factor, and combined with the terminal's own bandwidth requirements, the estimated satisfaction rate after the terminal switches to the target carrier is predicted. Calculate the difference between the estimated satisfaction rate and the current terminal bandwidth satisfaction rate, and subtract the pre-configured handover cost penalty factor to generate the net frequency hopping benefit; Under the condition that the net gain from frequency hopping is greater than zero, the target carrier corresponding to the maximum net gain from frequency hopping is selected to perform the carrier handover; When the net gain from frequency hopping is not greater than zero for all candidate target carriers, the terminal remains on the current carrier and does not perform a handover.

[0158] In this embodiment, when evaluating a candidate target carrier, the total capacity of the target carrier is obtained by summing the total capacity of the target carrier with the currently allocated bandwidth, and the apparent remaining capacity is calculated by subtracting the two. To avoid resource contention caused by multiple terminals concurrently jumping to the same spurious idle carrier, the system detects and counts the number of terminals with a current satisfaction rate of less than 1 on the target carrier, and defines them as the number of existing terminals in an unmet demand state.

[0159] Among them, the competition discount factor δ of the target carrier _c' = 1 / (1 + N _c'_unsatisfied ); N _c'_unsatisfied This represents the number of existing terminals on the target carrier that are currently in a state of unmet demand.

[0160] Multiplying the apparent remaining capacity by the competition discount factor yields the effective remaining capacity that the jumping terminal is actually expected to obtain.

[0161] Based on the bandwidth requirements reported by the terminal and the available remaining capacity, the expected satisfaction rate in terms of committed bandwidth and maximum bandwidth is calculated respectively.

[0162] In one optional implementation, the system prioritizes allocating a larger proportion of the effective remaining capacity to the committed bandwidth demand. Specifically, the system takes the smaller value between the aforementioned proportion of the effective remaining capacity and the terminal's committed bandwidth demand and divides it by the committed bandwidth demand to obtain the expected satisfaction rate of the committed bandwidth dimension. The expected satisfaction rate for the maximum bandwidth dimension is obtained by taking the smaller value between the remaining proportion of effective remaining capacity and the maximum bandwidth increment requirement of the terminal, and then dividing it by the maximum bandwidth increment requirement.

[0163] The remaining portion is allocated to the maximum bandwidth requirement. Those skilled in the art can adjust this allocation ratio based on the system's priority requirements for guaranteed committed bandwidth. The system weights and fuses the expected satisfaction rates of these two dimensions using the same weighting coefficients as in the preceding steps, calculating and outputting the comprehensive estimated satisfaction rate.

[0164] Based on this assessment, the system obtains the current terminal bandwidth satisfaction rate before the terminal frequency hopping, subtracts the current terminal bandwidth satisfaction rate from the estimated satisfaction rate, and then subtracts the handover cost penalty factor representing the physical loss of communication interruption to calculate the net benefit of frequency hopping.

[0165] Among them, the net frequency hopping gain Δ of the terminal switching to the target carrier _G_t_c' = η _t_c' - η_t_cur - γ; η _t_c' For the predicted estimated satisfaction rate, η _t_cur γ represents the current terminal bandwidth satisfaction rate, and γ is a preset handover cost penalty factor.

[0166] Optionally, the handover cost penalty factor can be set within the range of 0.02 to 0.10. The system iterates through and calculates the net frequency hopping gain of all candidate target carriers. Only when the calculated net frequency hopping gain is greater than 0 is the physical handover confirmed to have positive technical improvement value, and the target carrier with the largest net frequency hopping gain value is extracted and control signaling is sent to drive the radio frequency front-end to perform the carrier handover.

[0167] Step 603: Obtain the cumulative number of frequency hoppings of the terminal within a preset time window in real time; based on the pre-configured basic cooling time, set an adaptive cooling time that increases positively with the frequency hopping frequency according to the cumulative number of frequency hoppings; before the adaptive cooling time is reached, suspend triggering new carrier switching for the terminal to suppress network oscillations caused by frequent frequency hopping.

[0168] Specifically, after the terminal completes the frequency switching action at the physical layer, the system immediately records the timestamp of the action and triggers the backoff cooling logic. The system sets a preset time window of 10 minutes and continuously counts the total number of carrier switching actions performed by the specific terminal within this preset time window, extracting it as the cumulative frequency hopping count.

[0169] The system extracts the base cooldown time parameter and the penalty coefficient used to define the slope of time growth from the configuration library. It multiplies the cumulative number of frequency hopping counts by the penalty coefficient, adds a constant 1 to the product, and finally multiplies the sum by the base cooldown time to calculate the adaptive cooldown time.

[0170] That is, adaptive cooling time T cool_t = T0* (1 + κ * n hop_t ); Where T0 is the base cooldown time, κ is the pre-configured penalty coefficient, and n hop_t The cumulative number of frequency hopping times for the terminal within the preset time window.

[0171] According to the aforementioned backoff mechanism, when a terminal performs a carrier switch for the first time, its cooling-off time is 45 seconds; if the terminal performs a second switch within the preset time window, its cooling-off time will be linearly extended to 60 seconds.

[0172] Before the internal countdown timer reaches the adaptive cooldown time value, the system will intercept and discard any new carrier switching requests triggered by the terminal. By introducing this positively increasing time penalty strategy, the ping-pong frequency hopping phenomenon formed by the terminal between multiple carriers with similar remaining capacity is suppressed, which is beneficial to maintaining the signaling transmission stability of the satellite communication network.

[0173] Example 7 describes how, in a satellite communication network, historical allocation data is collected to calculate the weighted average satisfaction rate of the entire network, and the configuration weights are dynamically updated through a negative feedback control loop to achieve closed-loop optimization. It also provides a specific implementation method for data visualization and timeout recovery mechanisms.

[0174] In one possible embodiment, the following steps are included: Step 701: Periodically obtain the bandwidth satisfaction rate of each terminal group and the total actual bandwidth demand of all terminals in each terminal group within the historical period as a scale, and calculate the weighted average of the bandwidth satisfaction rate of each terminal group based on the actual demand scale to obtain the weighted average satisfaction rate of the entire network. Specifically, the global network management system uses a set scheduling period as the sampling interval to continuously record the bandwidth allocation data of each interactive network within a specific historical period.

[0175] Extract the actual allocated bandwidth value of each terminal group on the entire interactive network as the numerator, and extract the corresponding maximum bandwidth demand value as the denominator. Calculate the bandwidth satisfaction rate of each terminal group within the historical period by dividing the two values.

[0176] To prevent disproportionate fluctuations in satisfaction rates from smaller terminal groups from affecting the overall network evaluation benchmark, the system does not use an arithmetic average. Instead, it extracts the total actual demand of all terminals within each terminal group as the actual demand scale. The system performs a summation calculation across the entire network, dividing the total allocated bandwidth of all terminals by the total demand bandwidth of all terminals, and generates the weighted average satisfaction rate for the entire network.

[0177] The weighted average satisfaction rate η of the entire network calculated in the current scheduling cycle _ avg = ∑A _t / ∑R _t ; Where, ∑A _t The total bandwidth actually allocated to all terminals in the network during this scheduling period, ∑R _t This represents the total actual bandwidth demand of all terminals across the network during this scheduling period.

[0178] Step 702: Calculate the satisfaction rate deviation between the bandwidth satisfaction rate of each terminal group and the weighted average satisfaction rate of the entire network, and generate the negative feedback adjustment signal based on the satisfaction rate deviation to reduce the satisfaction difference between each terminal group. In this embodiment, the system compares the bandwidth satisfaction rate of each terminal group with the network-wide weighted average satisfaction rate generated in the previous steps. The system subtracts the terminal group's own bandwidth satisfaction rate from the network-wide weighted average satisfaction rate, and the resulting difference is the satisfaction rate deviation. This satisfaction rate deviation objectively characterizes the direction and magnitude of the deviation of a specific terminal group's resource acquisition level from the network-wide average benchmark.

[0179] Furthermore, when the bandwidth satisfaction rate of a specific terminal group is lower than the weighted average satisfaction rate of the entire network, the calculated satisfaction rate deviation is a positive value.

[0180] The negative feedback adjustment signal is generated based on the positive value to provide the terminal group with more bandwidth acquisition capability in the next cycle.

[0181] Conversely, when the bandwidth satisfaction rate of a specific terminal group is higher than the weighted average satisfaction rate of the entire network, the satisfaction rate deviation is a negative value. Based on this, the system generates a negative feedback adjustment signal to suppress the excess resources occupied by the terminal group and release them for use by other scarce groups.

[0182] Step 703: Adjust the pre-acquired initial configuration weights using the negative feedback adjustment signal, and truncate them within the pre-configured upper and lower limit constraint range of the weights to obtain the updated terminal group configuration weights for the next scheduling cycle.

[0183] Specifically, the system extracts the initial configuration weights of the terminal group pre-set by the administrator, combines them with the negative feedback adjustment signal generated in the previous steps, and performs update calculations with the set step size parameters.

[0184] The relative deviation rate is obtained by dividing the satisfaction rate deviation by the weighted average satisfaction rate of the entire network. The system multiplies the relative deviation rate by the configured adjustment step size, adds a constant of 1, and then multiplies it by the initial configured weights of the terminal group. To prevent the weights from diverging and becoming ineffective after multiple iterations, the system extracts the pre-configured lower and upper weight limits as cutoff boundaries. The system compares the above product result with the upper and lower limit boundaries. If the result is lower than the lower limit, it is assigned the lower limit value; if the result is higher than the upper limit, it is assigned the upper limit value.

[0185] The updated terminal group configuration weight w g_t1 = clip(w g_0 * (1 + μ * (η _avg- η g_t ) / max(η _avg , ε)), w min , w max ); Among them, w g_0 The initial configuration weights configured for the administrator, μ is the pre-configured adjustment step size, and η is the weights. g_t η represents the current terminal group bandwidth satisfaction rate for this specific terminal group. _avg The weighted average satisfaction rate of the entire network calculated in the above steps; ε is a preset zero-prevention constant, w min w is the pre-configured lower bound of the weights. max The pre-configured upper limit of weights is defined by `clip`, which is the truncation function, and `max`, which is the function to maximize the weight. In one optional implementation, the adjustment step size `μ` is set to 0.3. The lower limit of weights is defined by `w`. min Set as the initial configuration weight w g_0 0.5 times, the upper limit of the weight w max Set as the initial configuration weight w g_0 The updated terminal group configuration weights are used in the next scheduling cycle for the local idle bandwidth reallocation calculation of the interactive network, thereby forming a complete closed loop of network-wide state awareness and dynamic weighting.

[0186] Beyond the aforementioned core closed-loop control process, this embodiment also provides an auxiliary management mechanism for data visualization and timeout rollback. In the multi-dimensional situational statistics phase, the system statistically analyzes bandwidth allocation and utilization data at three levels: the global network, the interactive network, and the terminal. The system extracts these statistical results and generates a tree structure diagram on the console interface to display the allocated links. Simultaneously, it renders trend charts based on historical sampling points to display the bandwidth change trajectory. For network areas where the bandwidth satisfaction rate is lower than a preset standard, the system control display unit applies specific color markings to the corresponding graphical nodes on the interface to alert maintenance personnel to areas where current needs are not being met.

[0187] Furthermore, to address the risk of unexpected interruptions to the global network control link, the system configures a timeout countdown timer on each interactive network node. This countdown timer is reset when the interactive network receives a global scheduling result. If, after the countdown timer reaches zero and a timeout occurs, the interactive network still has not received a new global scheduling instruction, the system will determine that the control link is abnormal. In this case, the interactive network will automatically freeze the use of expired global scheduling results and extract the locally stored common quality of service (QoS) configuration parameters to take over the local bandwidth allocation logic, thereby ensuring that the basic services of the underlying satellite communication terminals are not paralyzed due to a single point of failure in the global management system.

Claims

1. A method for global dynamic allocation of bandwidth across multiple satellite communication networks for multi-level terminal groups, characterized in that, include: Construct a two-level bandwidth management architecture and configure mutually decoupled terminal group priorities, terminal priorities, and terminal group configuration weights; Collect terminal bandwidth requirements, filter terminal bandwidth requirements based on preset local configuration limits, and generate a global bandwidth request; Aggregate global bandwidth requests, execute global bandwidth scheduling, and obtain the global scheduling result; Based on the global scheduling results, local bandwidth allocation, including idle bandwidth allocation, is performed according to the terminal group priority and the terminal priority to obtain the service bandwidth, and the bandwidth satisfaction rate of the terminal group and the terminal bandwidth satisfaction rate are obtained. The service bandwidth is then mapped to the channel bandwidth in combination with the obtained spectrum efficiency of each terminal. The terminal congestion status is determined based on the terminal bandwidth satisfaction rate, and carrier handover is performed for terminals in the terminal congestion status based on competition revenue prediction. A negative feedback adjustment signal is generated based on the terminal group bandwidth satisfaction rate of historical periods, and the terminal group configuration weight is updated based on the negative feedback adjustment signal.

2. The method according to claim 1, characterized in that, Construct a two-level bandwidth management architecture and configure mutually decoupled terminal group priorities and terminal priorities, including: Construct a global bandwidth management tree that is effective across networks and an interactive network bandwidth management tree that is effective within a single interactive network; Configure global quality of service nodes and interactive quality of service nodes at each level of the global bandwidth management tree and the interactive network bandwidth management tree; Mount the terminal as a leaf node to the corresponding global quality of service node or interactive network quality of service node; Based on the hierarchical isolation structure of the global bandwidth management tree and the interactive network bandwidth management tree, bandwidth competition between terminal groups is separated from terminal bandwidth competition within a terminal group, thereby decoupling terminal group priority from terminal priority.

3. The method according to claim 2, characterized in that, The global bandwidth request is aggregated and global bandwidth scheduling is performed to obtain the global scheduling results, including: Global bandwidth requests are aggregated layer by layer from the leaf nodes to the root node of the global bandwidth management tree. The configuration limit of a non-leaf node is compared with the sum of the requirements of all its child nodes, and the smaller value is extracted as the node aggregation requirement of the corresponding node. Based on the hierarchical order of prioritizing the fulfillment of committed bandwidth requirements and then the fulfillment of maximum bandwidth requirements, global committed bandwidth quotas and global maximum bandwidth quotas are allocated to nodes at all levels according to their aggregation requirements. Aggregate the global committed bandwidth quota and the global maximum bandwidth quota at the interactive network level to generate global scheduling results that are distributed to each interactive network.

4. The method according to claim 1, characterized in that, In the local bandwidth allocation process, which involves allocating idle bandwidth based on the global scheduling results and according to the terminal group priority and terminal priority, the following is included: Based on the quota limit provided by the global scheduling results, the promised bandwidth on-demand allocation round and the maximum bandwidth on-demand allocation round are executed sequentially to establish the initial service bandwidth quota; After the promised bandwidth on-demand allocation round and the maximum bandwidth on-demand allocation round are completed, the unused allocation quota due to actual demand being lower than the quota limit is extracted and released to the public resource pool of the interactive network to form idle bandwidth; Identify terminal groups whose actual bandwidth demand is not fully met by the initial service bandwidth quota, and treat them as terminal groups in a state of unmet demand. Based on the idle bandwidth in the public resource pool, execute the idle committed bandwidth reallocation round and the idle maximum bandwidth reallocation round in sequence to obtain the updated service bandwidth quota.

5. The method according to claim 4, characterized in that, The process of sequentially executing the idle committed bandwidth reallocation round and the idle maximum bandwidth reallocation round includes: Based on the allocated bandwidth and corresponding bandwidth requirements of each terminal group, calculate the bandwidth satisfaction rate of each terminal group in the current allocation round. The resource gap corresponding to the bandwidth satisfaction rate of the terminal group is coupled and mapped with the pre-acquired terminal group configuration weight to generate the satisfaction rate balancing factor for each terminal group. Based on the satisfaction rate balancing factor of each terminal group, the redistribution ratio of idle bandwidth in the public resource pool is determined, and the idle bandwidth is preferentially allocated to terminal groups whose bandwidth satisfaction rate is lower than a set threshold according to the redistribution ratio, so as to obtain the idle bandwidth supplement quota of each terminal group.

6. The method according to claim 1, characterized in that, Determining terminal congestion status based on terminal bandwidth satisfaction rate includes: Extract the terminal bandwidth satisfaction rate in the dimensions of committed bandwidth and maximum bandwidth in the local bandwidth allocation, respectively; By using preset weighting coefficients to weight and fuse the terminal bandwidth satisfaction rates of the two dimensions mentioned above, a composite congestion index is calculated to characterize the degree of terminal limitation. When the composite congestion index is greater than the preset congestion threshold, the terminal is determined to be in a congested state and is added to the handover queue.

7. The method according to claim 6, characterized in that, Carrier handover based on competitive revenue prediction includes: For each candidate target carrier, the number of existing terminals currently in an unmet demand state on the target carrier is counted, and a competition discount factor is calculated to characterize the ease of obtaining remaining resources. The remaining capacity of the detected target carrier is evaluated by using a competitive discount factor, and the estimated satisfaction rate after the terminal switches to the target carrier is predicted by combining the terminal's own bandwidth requirements. Calculate the difference between the estimated satisfaction rate and the current terminal bandwidth satisfaction rate, and subtract the pre-configured handover cost penalty factor to generate the net frequency hopping benefit; If the net gain from frequency hopping is greater than zero, select the target carrier corresponding to the maximum net gain from frequency hopping to perform carrier handover.

8. The method according to claim 7, characterized in that, After performing carrier switching, the following is also included: Real-time acquisition of the cumulative frequency hopping count of the terminal within a preset time window; Based on the pre-configured base cooldown time, an adaptive cooldown time that increases positively with the frequency of frequency hopping is set according to the cumulative number of frequency hopping. Before the adaptive cooling time is reached, new carrier handover for this terminal is suspended to suppress network oscillations caused by frequent frequency hopping.

9. The method according to claim 5, characterized in that, The process of coupling and mapping the resource gap corresponding to the terminal group bandwidth sufficiency rate with the pre-acquired terminal group configuration weights includes: The difference between the value 1 and the current terminal group bandwidth satisfaction rate is taken as the resource gap, and the resource gap is subjected to exponential operation using the pre-configured fairness adjustment index. The result of the exponentiation operation is multiplied by the terminal group configuration weight to obtain the satisfaction rate balancing factor that couples the configuration weight with the dynamic gap.

10. The method according to claim 3, characterized in that, The aggregation of global committed bandwidth quotas and global maximum bandwidth quotas by interactive network dimension includes: For the same terminal group spanning multiple interactive networks, the global committed bandwidth quota and global maximum bandwidth quota for the terminal group are decomposed across networks according to the demand ratio reported by each interactive network in the global bandwidth application; By comparing the total quota after decomposition within each interactive network with the available bandwidth of that interactive network itself, when the total quota exceeds the available bandwidth, each quota within the interactive network is proportionally pruned, and the quota difference released by the pruning is recovered to the global resource pool to supplement the subsequent global idle resource reallocation.

Images