Network resource allocation method, device, electronic device and storage medium for slices

Abstract

The present application discloses a method, apparatus, electronic device, and storage medium for allocating network resources of slices, belonging to the field of satellite communication technology. In this method, a master controller receives intra-domain resource information of multiple satellite regions sent by multiple slave controllers, selects the regional paths of at least one slice based on the intra-domain resource information of the multiple satellite regions and the requirement information of the at least one slice, and sends a resource allocation instruction to the slave controller corresponding to the regional path to allocate network resources for the at least one slice within the corresponding satellite region. In this way, the master controller first preliminarily selects the regional paths of the slices, and then the slave controllers perform resource allocation within the satellite regions corresponding to the regional paths, which can minimize the number of satellites to be considered for resource allocation, thereby improving the speed of network resource allocation and reducing the impact of network resource reallocation on the continuous service provided by the slices.

Description

Technical Field

[0001] The present application relates to the field of satellite communication technology, and particularly to a method, apparatus, electronic device, and storage medium for allocating network resources of a slice. Background Art

[0002] Generally, a slice is a logically independent virtual network formed by partitioning a physical network, and each slice can be customized and optimized according to demand information such as quality requirements, latency, throughput, and other metrics.

[0003] The network resources of traditional terrestrial networks are relatively stable. Therefore, when applying slice technology in terrestrial networks, the network resources corresponding to the slices are relatively fixed and do not need to frequently switch the network resources carrying the slices. However, satellites are in high-speed motion, so the satellite network provided by satellites has the characteristics of dynamics and time-variance. These characteristics determine that the network resources in the satellite network are also dynamically changing. How to allocate network resources for slices based on the dynamically changing network resources in the satellite network is the key to ensuring continuous service provided by the slices. Summary of the Invention

[0004] Embodiments of the present application provide a method, apparatus, electronic device, and storage medium for allocating network resources of a slice, so as to provide a solution for allocating network resources for slices based on the dynamically changing network resources in the satellite network.

[0005] In a first aspect, embodiments of the present application provide a method for allocating network resources of a slice, including:

[0006] The total controller receives the intra-domain resource information of multiple satellite regions;

[0007] Based on the intra-domain resource information of the multiple satellite regions and the demand information of at least one slice, select the regional path of the at least one slice;

[0008] Send a resource allocation instruction to the sub-controller corresponding to the regional path to allocate network resources for the at least one slice within the satellite region corresponding to the sub-controller.

[0009] In some embodiments, based on the intra-domain resource information of the multiple satellite regions and the demand information of at least one slice, selecting the regional path of the at least one slice includes:

[0010] Extract features from the intra-domain resource information of the multiple satellite regions to obtain the intra-domain resource features of the multiple satellite regions;

[0011] Perform inter-domain correlation analysis on the intra-domain resource information of the multiple satellite regions to obtain the inter-domain resource features of the multiple satellite regions;

[0012] Based on the in-domain resource characteristics and the inter-domain resource characteristics, select a regional path that matches the requirement information of the at least one slice.

[0013] In some embodiments, based on the in-domain resource characteristics and the inter-domain resource characteristics, selecting a regional path that matches the requirement information of the at least one slice includes:

[0014] Based on the in-domain resource characteristics and the inter-domain resource characteristics, initialize the state of a pre-trained graph neural network model;

[0015] In the order of decreasing slice priority, input the requirement information of the at least one slice into the graph neural network model to obtain the regional path of the at least one slice. Among them, after obtaining the regional path of each slice, update the state of the graph neural network model based on the network resources required by the slice.

[0016] In some embodiments, the graph neural network model calculates the probabilities of multiple alternative regional paths between the starting satellite region and the ending satellite region of each slice, and determines the alternative regional path with the highest probability as the regional path of the slice.

[0017] In some embodiments, determine the slice priority of a slice according to the following steps:

[0018] Calculate the index values of multiple first indicators of the slice, and perform normalization processing on the index values of the multiple first indicators;

[0019] Perform weighted summation on the normalized index values of the multiple first indicators to obtain the priority score of the slice;

[0020] According to the priority score of the slice and the pre-established correspondence between the priority score and the slice priority, determine the slice priority of the slice.

[0021] In some embodiments, the in-domain resource information of the multiple satellite regions is sent after quantization, dimensionality reduction, and coding, and further includes:

[0022] Before selecting the regional path of the at least one slice based on the in-domain resource information of the multiple satellite regions and the requirement information of the at least one slice, perform decoding processing on the in-domain resource information of the multiple satellite regions.

[0023] In some embodiments, each type of in-domain resource information includes the index values of multiple second indicators. Quantize the in-domain resource information according to the following steps:

[0024] Determine the quantization method corresponding to the second indicator according to the value dispersion degree of each second indicator and the pre-established corresponding relationship between the value dispersion degree and the quantization method;

[0025] Quantize the indicator value of the second indicator according to the quantization method corresponding to the second indicator to obtain the quantization result of the second indicator.

[0026] In some embodiments, the dimensionality reduction process for each piece of in-domain resource information is performed according to the following steps:

[0027] Divide the quantized in-domain resource information into multiple data blocks;

[0028] Use the principal component analysis algorithm to perform dimensionality reduction on the multiple data blocks.

[0029] In some embodiments, the encoding process for each piece of in-domain resource information is performed according to the following steps:

[0030] For the multiple data blocks after dimensionality reduction, use the locality-sensitive hashing algorithm for hashing encoding, and use the binary encoding method to perform binary encoding on the hashing encoding result;

[0031] Perform decoding processing on the in-domain resource information of the multiple satellite regions, including:

[0032] For the binary encoding result, perform decoding using the binary decoding method, and use the locality-sensitive hashing algorithm to perform hashing decoding on the binary decoding result to obtain the multiple data blocks after dimensionality reduction.

[0033] In some embodiments, any sub-controller is deployed on the backbone satellite in the corresponding satellite region, and the backbone satellite is selected according to the following steps:

[0034] Obtain the capability characterization information of each satellite in the satellite region;

[0035] Score the service capabilities of the satellites according to the capability characterization information;

[0036] Select one satellite from the satellites as the backbone satellite according to the scoring results of the satellites.

[0037] In some embodiments, the capability characterization information includes at least one of in-domain coverage, available bandwidth, latency, and central processing unit (CPU) load.

[0038] In a second aspect, an embodiment of the present application provides a method for allocating network resources for slices, including:

[0039] The sub - controller sends the in - domain resource information of the corresponding satellite area to the master - controller. The master - controller selects the area path of the at least one slice based on the in - domain resource information of multiple satellite areas and the requirement information of at least one slice, and sends a resource allocation instruction to the sub - controller corresponding to the area path.

[0040] If the resource allocation instruction is received, then based on the resource allocation instruction, slice network resources are allocated within the corresponding satellite area.

[0041] In some embodiments, allocating slice network resources within the corresponding satellite area based on the resource allocation instruction includes:

[0042] Allocate the network resources within the satellite area to the corresponding slice in the order of slice priority from high to low.

[0043] In some embodiments, the slice priority of a slice is determined according to the following steps:

[0044] Calculate the index values of multiple first - type indicators of the slice, and perform normalization processing on the index values of the multiple first - type indicators.

[0045] Perform weighted summation on the normalized index values of the multiple first - type indicators to obtain the priority score of the slice.

[0046] Determine the slice priority of the slice according to the priority score of the slice and the pre - established correspondence between the priority score and the slice priority.

[0047] In some embodiments, before sending the in - domain resource information of the corresponding satellite area to the master - controller, it further includes:

[0048] Perform quantization, dimensionality reduction, and encoding processing on the in - domain resource information.

[0049] In some embodiments, the in - domain resource information includes the index values of multiple second - type indicators. The in - domain resource information is quantized according to the following steps:

[0050] According to the value dispersion degree of each second - type indicator and the pre - established correspondence between the value dispersion degree and the quantization method, determine the quantization method corresponding to the second - type indicator.

[0051] Quantize the index value of the second - type indicator according to the quantization method corresponding to the second - type indicator to obtain the quantization result of the second - type indicator.

[0052] In some embodiments, the in - domain resource information is dimensionally reduced according to the following steps:

[0053] Divide the quantized in-domain resource information into multiple data blocks;

[0054] Use the principal component analysis algorithm to perform dimensionality reduction processing on the multiple data blocks.

[0055] In some embodiments, the following steps are used to perform encoding processing on each type of in-domain resource information:

[0056] For the multiple data blocks after dimensionality reduction, use the local hash-sensitive algorithm for hash encoding, and use binary encoding to perform binary encoding on the hash encoding results.

[0057] In some embodiments, any sub-controller is deployed on the backbone satellite in the corresponding satellite area, and the backbone satellite is selected according to the following steps:

[0058] Obtain the capability characterization information of each satellite in the satellite area;

[0059] According to the capability characterization information, score the service capabilities of the satellites;

[0060] According to the scoring results of the satellites, select one satellite from the satellites as the backbone satellite.

[0061] In some embodiments, the capability characterization information includes at least one of in-domain coverage, available bandwidth, delay, and central processing unit (CPU) load.

[0062] In a third aspect, an embodiment of the present application provides a sliced network resource allocation device, which is applied to a master controller and includes:

[0063] An acquisition module, configured to receive in-domain resource information of multiple satellite areas;

[0064] A selection module, configured to select the area path of the at least one slice based on the in-domain resource information of the multiple satellite areas and the demand information of the at least one slice;

[0065] A sending module, configured to send a resource allocation instruction to the sub-controller corresponding to the area path, so as to allocate network resources for the at least one slice in the satellite area corresponding to the sub-controller.

[0066] In a fourth aspect, an embodiment of the present application provides a sliced network resource allocation device, which is applied to a sub-controller and includes:

[0067] A sending module, configured to send the in-domain resource information of the corresponding satellite area to the master controller, and the master controller selects the area path of the at least one slice based on the in-domain resource information of the multiple satellite areas and the demand information of the at least one slice, and sends a resource allocation instruction to the sub-controller corresponding to the area path;

[0068] A distribution module, configured to, if receiving the resource allocation instruction, perform slice network resource allocation within a corresponding satellite area based on the resource allocation instruction.

[0069] In a fifth aspect, an embodiment of the present application provides an electronic device, including: at least one processor, and a memory communicatively connected to the at least one processor, wherein:

[0070] The memory stores a computer program executable by the at least one processor, and when the computer program is executed by the at least one processor, the at least one processor is enabled to execute the above-mentioned network resource allocation method for slices.

[0071] In a sixth aspect, an embodiment of the present application provides a storage medium, when the computer program in the storage medium is executed by a processor of an electronic device, the electronic device can execute the above-mentioned network resource allocation method for slices.

[0072] In a seventh aspect, an embodiment of the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, the above-mentioned network resource allocation method for slices is implemented.

[0073] In the embodiments of the present application, the satellite network is divided into multiple satellite areas, each satellite area corresponds to a sub-controller, the main controller receives the intra-domain resource information of multiple satellite areas sent by multiple sub-controllers, selects the area path of at least one slice based on the intra-domain resource information of multiple satellite areas and the demand information of at least one slice, and then sends a resource allocation instruction to the sub-controllers corresponding to each satellite area on the area path, so that each sub-controller allocates network resources for at least one slice within the corresponding satellite area. In this way, the main controller first preliminarily selects the area path of the slice, and then the sub-controller performs resource allocation within the satellite area corresponding to the area path, which can minimize the number of satellites to be considered for resource allocation, thereby improving the network resource allocation speed and reducing the impact of network resource reallocation on the continuous service provided by the slice. Description of the Drawings

[0074] The drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application, and do not constitute an improper limitation to the present application. In the drawings:

[0075] Figure 1 It is a schematic diagram of a slice in a satellite network provided by an embodiment of the present application;

[0076] Figure 2 It is another schematic diagram of a slice in a satellite network provided by an embodiment of the present application;

[0077] Figure 3 Flow chart of a network resource allocation method for slices provided by an embodiment of the present application;

[0078] Figure 4 Flow chart of a method for selecting an area path of a slice provided by an embodiment of the present application;

[0079] Figure 5 Flow chart of another network resource allocation method for slices provided by an embodiment of the present application;

[0080] Figure 6 Schematic diagram of the communication process between an inter-domain master controller and each intra-domain controller provided by an embodiment of the present application;

[0081] Figure 7 Schematic diagram of a method for determining the area path of each slice provided by an embodiment of the present application;

[0082] Figure 8 Schematic diagram of the structure of a network resource allocation device for slices provided by an embodiment of the present application;

[0083] Figure 9 Schematic diagram of the structure of another network resource allocation device for slices provided by an embodiment of the present application;

[0084] Figure 10 Schematic diagram of the hardware structure of an electronic device for implementing the network resource allocation method for slices provided by an embodiment of the present application. Detailed implementation manners

[0085] In order to be able to use slice technology in a dynamically changing satellite network, embodiments of the present application provide a network resource allocation method, device, electronic device, and storage medium for slices.

[0086] The following describes the preferred embodiments of the present application with reference to the accompanying drawings of the specification. It should be understood that the preferred embodiments described herein are only for explaining and illustrating the present application, and are not used to limit the present application. And without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0087] See Figure 1 , Figure 1A schematic diagram of slices in a satellite network provided by an embodiment of the present application, including multiple satellites and three slices: Slice 1 to Slice 3. Among them, the service characteristics of different slices are different. For example, the service characteristic of Slice 1 is large bandwidth, the service characteristic of Slice 2 is wide-area Internet of Things, and the service characteristic of Slice 3 is low latency. Generally, each slice corresponds to a number of virtual nodes, and these virtual nodes provide the network resources required by the slice. Moreover, each virtual node corresponds to a satellite, and one satellite can correspond to multiple virtual nodes. Taking satellite S1 as an example, S1 is mapped into virtual satellites S1-V1 and S1-V2. S1-V1 corresponds to Slice 3, and S1-V2 corresponds to Slice 2. Taking satellite S2 as an example, S2 is mapped into virtual satellites S2-V1, S2-V2, and S2-V3. S2-V1 corresponds to Slice 3, S2-V2 corresponds to Slice 2, and S2-V3 corresponds to Slice 1.

[0088] It should be noted that Figure 1 What is shown is only the correspondence between slices and satellites within a certain period of time. As the satellites are constantly moving rapidly, the correspondence between slices and satellites will also change continuously.

[0089] As the topology of the satellite network becomes more and more complex and the number of slices carried by the satellite network increases, the complexity of network resource allocation for slices in the satellite network also rises linearly. Moreover, each time resources are allocated to slices, all satellites in the global topology at the current time point / time slice need to be traversed in sequence according to the slice priority, calculate the relevant routing and Quality of Service (QoS), and give the optimal allocation plan through a specific resource optimization allocation strategy. Since the number of satellites in the global topology is large, the allocation speed is relatively slow, which is likely to affect the service quality of slices.

[0090] Therefore, in the embodiment of the present application, the satellite network is divided into multiple relatively independent distributed regions (i.e., multiple satellite regions), and the global complex problem is decomposed into multiple relatively easy-to-solve problems, thereby realizing distributed resource allocation for slices. In this way, it not only helps to reduce the network allocation complexity, but also can significantly improve the allocation efficiency, providing a strong guarantee for the stable and efficient operation of the satellite network.

[0091] See Figure 2 , Figure 2 Another schematic diagram of slices in a satellite network provided by an embodiment of the present application, including multiple satellites and three slices: Slice 1 to Slice 3, and the service characteristics of different slices are different. Figure 2Divide multiple satellites into four equal parts to obtain four distributed regions. Satellites within each distributed region can elect backbone satellites according to certain rules, such as the largest coverage area within the domain and the most available resources, and deploy in-domain controllers (i.e., sub-controllers) on the backbone satellites. Then, the in-domain controllers manage the satellites within the domain. For example, the in-domain controllers receive and distribute slice resources, and allocate network resources required for the corresponding slices based on the in-domain satellites, thereby achieving relative autonomy. At the same time, an inter-domain master controller (i.e., master controller) can be deployed on the ground network control center. The inter-domain master controller communicates with each in-domain controller to coordinate the relationships between domains, thereby achieving low-cost rapid response and optimized resource utilization.

[0092] See Figure 3 , Figure 3 is a flowchart of a method for allocating network resources of a slice provided by an embodiment of the present application. This method is applied to the master controller, and this method includes the following steps.

[0093] In step 301, receive the in-domain resource information of multiple satellite regions.

[0094] Among them, each sub-controller sends the in-domain resource information of its corresponding satellite region to the master controller. The in-domain resource information is used to describe the network resource status within the satellite region. The in-domain resource information includes, for example, slice identifiers, slice occupied resources, available resource amounts of each satellite, allocated resource amounts of each satellite, resource utilization rates of each satellite, link bandwidth, jitter, delay, packet loss rate, satellite topology, satellite interconnection conditions, etc.

[0095] In step 302, based on the in-domain resource information of multiple satellite regions and the requirement information of at least one slice, select the regional paths of the at least one slice.

[0096] In some embodiments, in order to simplify the data and reduce the quantity to be transmitted, the in-domain resource information of multiple satellite regions is sent after being quantified, dimension-reduced, and encoded by multiple sub-controllers.

[0097] The processes of quantization, dimension reduction, and encoding will be introduced separately below.

[0098] 1. Quantization.

[0099] Generally, each type of in-domain resource information may include the index values of multiple second indexes. Then, the corresponding sub-controller can perform quantization processing on the in-domain resource information according to the following steps:

[0100] According to the value dispersion degree of each second index and the pre-established corresponding relationship between the value dispersion degree and the quantization method, determine the quantization method corresponding to the second index. Then, according to the quantization method corresponding to the second index, perform quantization processing on the index value of the second index to obtain the quantization result of the second index.

[0101] In this way, for second indexes with different value dispersion degrees, different quantization methods are used for quantization. The quantization method is relatively flexible, which is convenient for minimizing the data complexity to the greatest extent.

[0102] 2. Dimensionality reduction.

[0103] During specific implementation, for each piece of in-domain resource information, the corresponding sub-controller can divide the quantized in-domain resource information into multiple data blocks. Then, use the principal component analysis algorithm to perform dimensionality reduction processing on these multiple data blocks to reduce the amount of data to be transmitted.

[0104] 3. Encoding.

[0105] For the multiple data blocks after dimensionality reduction, the corresponding sub-controller can first perform hash encoding using the local hash-sensitive algorithm, and then use the binary encoding method to perform binary encoding on the hash encoding result.

[0106] Therefore, before step 302, the master controller can also perform decoding processing on the in-domain resource information of these multiple satellite regions. The decoding process is as follows: for the binary encoding result, first perform decoding using the binary decoding method, and then use the local hash-sensitive algorithm to perform hash decoding on the binary decoding result, so as to obtain the multiple data blocks after dimensionality reduction. The data in these multiple data blocks is the in-domain resource information of the subsequent multiple satellite regions.

[0107] During specific implementation, the regional path of at least one slice can be selected according to the following steps.

[0108] The first step: Extract features from the in-domain resource information of multiple satellite regions to obtain the in-domain resource features of multiple satellite regions.

[0109] For example, extract features from the in-domain resource information of each satellite region to obtain the in-domain resource features of this satellite region. The in-domain resource features are such as the resource status of the in-domain satellites, the topological matrix of the in-domain satellites, labels related to the in-domain slice requirements, etc. Among them, the resource status of the in-domain satellites is such as CPU usage rate, memory usage rate, etc. The topological matrix of the in-domain satellites is used to reflect the connectivity between the in-domain satellites. The labels related to the in-domain slice requirements are such as whether to support a certain specific type of service, the priority of the corresponding slice, etc.

[0110] The second step: Perform inter-domain correlation analysis on the in-domain resource information of multiple satellite regions to obtain the inter-domain resource features of multiple satellite regions.

[0111] For example, perform inter-domain correlation analysis on the intra-domain resource information of every two satellite regions among multiple satellite regions to obtain the inter-domain resource characteristics of these two satellite regions. The inter-domain resource characteristics include other service constraints such as inter-domain bandwidth and latency.

[0112] Thirdly, based on the intra-domain resource characteristics and inter-domain resource characteristics of multiple satellite regions, select a regional path that matches the demand information of at least one slice.

[0113] Among them, the regional path refers to a path composed of satellite regions, and the regional path generally includes at least two satellite regions.

[0114] Generally, the graph neural network model can be expressed as G = (V, E), where V is the set of all satellite regions in the satellite network, E is the set of association links between satellite regions, the vertex represents the regional characteristics of the i-th satellite region, and the edge represents the association characteristics between satellite region i and satellite region j.

[0115] In the training stage, the graph neural network model learns the relationship characteristics between the demand information of the slice and the regional path in two cases: when the network resources are sufficient and when they are insufficient. In the prediction stage, each time a prediction is made, only the specific resource status of the graph neural network model (that is, the attribute values of the nodes and the attribute values of the edges of the graph neural network model) and the demand information of the slice need to be given, and the graph neural network model can select a regional path for the slice according to the corresponding resource status (sufficient or insufficient).

[0116] Specifically, during implementation, it can be based on Figure 4 the shown process to select the regional path of the slice, and this process includes the following steps:

[0117] In step 3021, select a target slice from at least one slice. The target slice has not been selected for satellite regions and has the highest priority.

[0118] That is, select a slice as the target slice from at least one slice in the order of decreasing slice priority.

[0119] For example, determine the slice priority of a slice according to the following steps:

[0120] Calculate the index values of the slice under multiple first indicators, then perform normalization processing on the index values of multiple first indicators, perform weighted summation on the normalized index values of multiple first indicators to obtain the priority score of the slice, and then determine the slice priority of the slice according to the priority score of the slice and the pre-established correspondence between the priority score and the slice priority. Among them, multiple first indicators are indicators used to measure the slice priority.

[0121] In step 3022, the requirement information of the target slice is input into the graph neural network model to obtain the regional path of the target slice. The initial state of the graph neural network model is determined according to the intra-domain resource characteristics and inter-domain resource characteristics of multiple satellite regions.

[0122] Here, the state of the graph neural network model includes the attribute values of nodes and the attribute values of edges. Since the intra-domain resource characteristics and inter-domain resource characteristics of multiple satellite regions can characterize the resource status of the current satellite network, such as available resources, existing resources, etc., therefore, at the initial state, the attribute value of each node in the graph neural network model can be the intra-domain resource characteristics of the corresponding satellite region, and the attribute value of each edge in the graph neural network model can be the inter-domain resource characteristics of the corresponding two satellite regions.

[0123] In step 3023, based on the network resources required by the target slice, the state of the graph neural network model is updated.

[0124] Generally, after the graph neural network model predicts the regional path of the target slice, the available network resources of the satellite network will decrease. Therefore, the state of the graph neural network model can be updated based on the network resources required by the target slice to prepare for predicting the regional path of the next target slice.

[0125] In step 3024, it is judged whether the region selection stop condition is satisfied. If not, return to step 3021. If so, enter step 3025.

[0126] Among them, the region selection stop conditions include that no target slice can be selected from these at least one slice (that is, the regional paths of all slices have been selected), the network resources of the entire satellite network are insufficient, etc.

[0127] In step 3025, the region selection is stopped.

[0128] In step 303, a resource allocation instruction is sent to the sub-controller corresponding to the regional path to allocate network resources for these at least one slice within the satellite region corresponding to the sub-controller.

[0129] That is, for the regional path of each slice, a resource allocation instruction can be sent to the sub-controllers corresponding to each satellite region on the regional path. Subsequently, each sub-controller can allocate network resources for this slice in the corresponding satellite region based on the resource allocation instruction.

[0130] In practical applications, for each satellite area, the ability characterization information of each satellite within the satellite area can be obtained, such as the in-domain coverage range, available bandwidth, latency, CPU load, etc. According to the ability characterization information of each satellite, the service capabilities of each satellite are scored. Then, based on the scoring results of each satellite, a satellite is selected from each satellite as the backbone satellite. For example, the satellite with the highest scoring result is used as the backbone satellite. After that, a sub-controller can be deployed on the backbone satellite.

[0131] Taking the ability characterization information including the in-domain coverage range, available bandwidth, latency, and CPU load as an example, the process of selecting the backbone satellite will be introduced.

[0132] In specific implementation, for each satellite in any satellite area, the in-domain coverage range of the satellite can be scored according to the rule that the in-domain coverage range is positively correlated with the score, the available bandwidth of the satellite can be scored according to the rule that the available bandwidth is positively correlated with the score, the latency of the satellite can be scored according to the rule that the latency is negatively correlated with the score, and the CPU load of the satellite can be scored according to the rule that the CPU load is negatively correlated with the score. Then, the scores of the satellite can be weighted and averaged to obtain the final score (i.e., the scoring result) of the satellite. Furthermore, the satellite with the highest score in the satellite area is determined as the backbone satellite.

[0133] Next, the role of the sub-controller will be introduced.

[0134] See Figure 5 , Figure 5 which is a flowchart of another network resource allocation method for slices provided in an embodiment of this application. This method is applied in the sub-controller, and this method includes the following steps.

[0135] In step 501, the in-domain resource information corresponding to the satellite area is obtained.

[0136] In step 502, the in-domain resource information is quantized, dimension-reduced, and encoded.

[0137] In practical applications, the in-domain resource information may include the index values of multiple second-level indicators. Among them, the index values of some second-level indicators are decimals or integers with relatively large numerical differences, which consume a lot of resources during transmission and processing. To reduce the complexity of data processing, the index values of these indicators can be quantized.

[0138] For example, the in-domain resource information is quantified according to the following steps: according to the value dispersion degree of each second indicator and the pre-established corresponding relationship between the value dispersion degree and the quantization method, determine the quantization method corresponding to the second indicator, and then, according to the quantization method corresponding to the second indicator, quantify the indicator value of the second indicator to obtain the quantization result of the second indicator. In this way, different quantization methods can be used to quantify indicators with different value dispersion degrees, which can further improve the quantization effect.

[0139] After quantization, the in-domain resource information can also be subjected to dimensionality reduction and encoding processing to further reduce the amount of data to be transmitted and improve the data sending efficiency of the sub-controller. Among them, the dimensionality reduction process is to divide the quantized in-domain resource information into multiple data blocks, and then use the principal component analysis algorithm to perform dimensionality reduction processing on these multiple data blocks; the encoding process is to first perform hash encoding on the multiple data blocks after dimensionality reduction using the local hash-sensitive algorithm, and then use the binary encoding method to perform binary encoding on the hash encoding result.

[0140] In step 503, the in-domain resource information is sent to the master controller, and the master controller selects the area path of at least one slice based on the in-domain resource information of multiple satellite areas and the demand information of at least one slice, and sends a resource allocation instruction to the sub-controller corresponding to the area path.

[0141] In step 504, if a resource allocation instruction is received, then based on the resource allocation instruction, slice network resource allocation is performed within the corresponding satellite area of itself.

[0142] In specific implementation, the resource allocation instruction includes the slice identifier of at least one slice that needs to perform network resource allocation, and the amount of network resources that each slice needs to be allocated. Then, the network resources can be allocated for these slices in their respective satellite areas in the order of slice priority from high to low. Among them, the slice priority of a slice can be determined according to the following steps: calculate the indicator values of multiple first indicators of the slice, perform normalization processing on the indicator values of these multiple first indicators, and then perform weighted summation on the normalized indicator values of the multiple first indicators to obtain the priority score of the slice. Then, according to the priority score of the slice and the pre-established corresponding relationship between the priority score and the slice priority, determine the slice priority of the slice.

[0143] In the embodiment of the present application, the master controller first preliminarily selects the area path of the slice, and then the sub-controller performs resource allocation within the satellite area corresponding to the area path, which can minimize the number of satellites that need to be considered for resource allocation, thereby improving the network resource allocation speed and reducing the impact of network resource reallocation on the continuous service provided by the slice.

[0144] See Figure 6 ,Figure 6 A schematic diagram of the communication process between the inter-domain general controller and each intra-domain controller provided by an embodiment of this application, including an inter-domain general controller and multiple intra-domain controllers. Among them, the inter-domain general controller includes a global resource monitoring module, a slice priority division module, and a dynamic resource allocation module. Each intra-domain controller includes a local monitoring module, a resource quantization module, an information encoding module, a resource allocation and distribution module, and an inter-domain / intra-domain interaction module.

[0145] The communication process between the inter-domain general controller and each intra-domain controller will be introduced through these modules below.

[0146] The local monitoring module is used to obtain the original intra-domain resource information within the domain, including slice description information, satellite status information, and link status information. Among them, the slice description information includes slice identifiers, resources occupied by slices, etc.; the satellite status information includes available resource quantity, allocated resource quantity, resource utilization rate, etc.; the link status information includes link bandwidth, jitter, latency, packet loss rate, etc.

[0147] The resource quantization module is used to perform quantization processing on the original intra-domain resource information, that is, to convert the complex intra-domain resource information into measurable and comparable metrics to reduce the complexity of subsequent resource allocation.

[0148] In practical applications, the above intra-domain resource information will include the metric values of multiple second metrics, and the value dispersion degrees of different second metrics are different. To better perform quantization, a correspondence relationship between the value dispersion degree and the quantization method can be established in advance. Subsequently, for each second metric, according to the value dispersion degree of the second metric and the pre-established correspondence relationship between the value dispersion degree and the quantization method, determine the quantization method corresponding to the second metric, and then perform quantization processing on the metric value of the second metric according to the quantization method corresponding to the second metric to obtain the quantization result of the second metric.

[0149] The quantization process will be illustrated by examples below. Taking multiple second metrics including latency, bandwidth, and packet loss rate as examples.

[0150] The latency of the satellite is usually between 100ms and 1000ms, and the value dispersion degree is relatively large. Some approximation operations can be performed during quantization. For example, round 444ms to 400ms and round 487ms to 500ms.

[0151] The bandwidth of the satellite is usually between 100MB and 100GB, and the value dispersion degree is even larger. During quantization, the bandwidth values in different intervals can be mapped to different numbers with the help of a piecewise function. For example, map the bandwidth values within 10000MB to 1, map the bandwidth values between 10000MB and 10GB to 2, and map the bandwidth values between 10GB and 20GB to 3.

[0152] The packet loss rate of a satellite is generally a decimal between 0 and 1, and is taken to four decimal places. The value dispersion is relatively small, and the more digits after the decimal point, the higher the computational complexity. To reduce the computational complexity, when quantifying, the value of the packet loss rate can be directly truncated to two decimal places. For example, the packet loss rate of 0.343344% can be truncated to 0.34%.

[0153] The information encoding module is used to first divide the quantized in-domain resource information into multiple data blocks. Each data block contains a part of the data, and these data are adjacent to each other in the original space. Then, apply the Principal Component Analysis (PCA) algorithm to each data block for dimensionality reduction. The PCA algorithm can extract the main components of the data, that is, the main direction of data change, so as to map the data from the original high-dimensional space to a low-dimensional space. Finally, use the Local hash sensitivity (LSH) algorithm to perform hash encoding on each dimensionality-reduced data block. LSH is a special hash function that can map similar data points to the same hash bucket with high probability, and map distant data points to different hash buckets with high probability. Further encode the hash-encoded data blocks (such as encoding using binary encoding).

[0154] The inter-domain / intra-domain interaction module is used to send the encoded in-domain resource information to the global resource monitoring module.

[0155] The global resource monitoring module is used to decode the in-domain resource information sent by each inter-domain / intra-domain interaction module to obtain the in-domain resource information from a global perspective. That is, use the same binary decoding method and LSH function to decode and reconstruct the received encoded data to restore an approximate representation of the original data.

[0156] The slice priority division module is used to determine the priority of each slice according to the priority characterization information of each slice. And the priority of each slice can be updated regularly or in real time.

[0157] Specifically, the priority of each slice can be determined according to the following steps:

[0158] Step 1: Define a series of first indicators for quantifying the priority. These first indicators can be defined based on the traffic characteristics and service characteristics of the slice. Among them, traffic characteristics such as traffic peak value, average value, variance, etc., and service characteristics such as delay sensitivity, bitstream quality of video slices, call duration distribution of voice slices, etc.

[0159] Here, three service characteristic indicators are provided for reference: 1. Latency-sensitive services such as real-time communication and remote control are very sensitive to latency and require quick responses; 2. High-bandwidth demand services, the bandwidth requirements of different services vary greatly. For example, high-definition video streams require higher bandwidth than text chats; 3. Critical services such as financial transactions and emergency services require highly reliable network connections.

[0160] Step 2: For each slice, according to the relevant information of the slice, calculate the indicator value of the slice under each first indicator, and normalize the indicator value to ensure comparability between different first indicators.

[0161] Step 3: Assign a weight value to each first indicator, which represents the importance of the first indicator when determining the priority. The weight value is positively correlated with the importance, and the sum of the weights of all first indicators can be 1.

[0162] Step 4: For each slice, perform a weighted sum of the indicator values of the slice under each first indicator to obtain a priority score. Among them, the weight of each first indicator is the weight determined in Step 3. Priority score formula:

[0163] ,

[0164] where P i is the priority score of slice i, w j is the weight of indicator j, and to ensure that the sum of the weights is 1 for easy comparison between different slices. x ij is the normalized score of slice i on indicator j, and M is the total number of first indicators.

[0165] Step 5: According to the calculated priority scores, divide the slices into different priority levels.

[0166] Taking three priorities as an example: high priority, medium priority, medium priority. Slices with a priority score higher than the first threshold can be divided into high priority, slices lower than the second threshold can be divided into low priority, and slices between the two can be divided into medium priority, where the first threshold is greater than the second threshold.

[0167] The dynamic resource allocation module is used to select the regional paths of each slice based on the resource information within each domain and the demand information of each slice, using a pre-trained graph neural network model, and then send resource allocation instructions to the inter-domain / intra-domain interaction modules corresponding to each satellite region on the regional path.

[0168] The training process of the graph neural network model is introduced below.

[0169] 1. To achieve dynamic resource allocation, a graph neural network should be defined first.

[0170] For example, define the graph G = (V, E), where V is the set of all satellite regions in a large-scale space bearer network (i.e., a satellite network), E is the set of associated links between satellite regions, the node features of V include the resource status of multiple satellites within the domain (such as CPU and memory usage rates), the topological matrix representation of multiple satellites within the domain, labels related to slice requirements within the domain (e.g., whether a certain specific type of service needs to be supported), and the priorities of the corresponding slices. The inter-domain association features of E include the bandwidth, latency, and service constraints of inter-domain links. The vertex represents the regional features of the i-th satellite region, and the edge represents the association features between the domain satellite region i and the satellite region j.

[0171] 2. Initialize the model.

[0172] Initialization involves setting initial parameters and configurations for the graph neural network model, specifically including determining the number of layers in the network, the number of neurons in each layer, the type of activation function, etc. In addition, an appropriate learning rate and optimization algorithm can be selected for the model, and the initial parameter values can be given.

[0173] Here, a Graph Attention Network (GAT) model can be selected for training, and the input layer, hidden layer, and output layer can be defined. The input layer accepts the attribute values of nodes, the attribute values of edges, and slice requirements, and the output layer predicts path selection. The cross-entropy loss can be selected as the objective function to transform the regional path selection problem of slices into a classification problem. In addition, the backpropagation algorithm and an optimizer (such as Adam) can be selected to update the model parameters.

[0174] 3. Train the model.

[0175] Use the collected data to train the graph neural network model. During the training process, update the weights and biases of the model through the backpropagation algorithm to minimize the objective function. Training may require multiple epochs until the model converges or reaches a predetermined performance standard.

[0176] 4. Conduct model validation and testing.

[0177] Evaluate the model performance on the validation set, adjust the hyperparameters (such as the learning rate, hidden layer size, etc.), evaluate the generalization ability of the model on the test set, and ensure that the model can work effectively on unseen data.

[0178] Then, use the trained model for prediction.

[0179] According to the in-region resource information of each satellite region currently, generate the in-region features and inter-region features of each satellite region. Use the in-region features and inter-region features as the attribute values of nodes and the attribute values of edges in the graph neural network model respectively. Then, input the demand information of each slice into the model in the order of decreasing slice priority to obtain the regional path of the slice. Among them, after obtaining the regional path of each slice, the attribute values of nodes and the attribute values of edges in the graph neural network model can be updated according to the network resources required by this slice, so that the nodes and edges of the graph neural network model can reflect the changes in network resources in real time and ensure that the model always makes decisions based on the latest network resources.

[0180] See Figure 7 , Figure 7 FIG. is a schematic diagram for determining the regional path of each slice provided by an embodiment of the present application. Four circles represent four satellite regions. Assume that the numbering rule of satellite regions is: starting from the upper left corner and numbered clockwise in sequence as: A, B, C, D. It can be seen that the regional path of slice 1 is: D -> C -> B, the regional path of slice 2 is: A -> B -> C, and the regional path of slice 3 is: D -> C.

[0181] The resource allocation and distribution module is used to determine the corresponding slices of the corresponding satellite region after the corresponding inter-region / intra-region interaction module receives the resource allocation instruction. Select the satellite corresponding to each slice from the satellites in the satellite region in the order of decreasing priority of these slices, and allocate network resources for the slice from the network resources that these satellites can provide.

[0182] The advantages of the embodiments of the present application are as follows:

[0183] 1) The hierarchical and domain-based management and control technology reduces complexity: In the slice resource allocation of a large-scale space bearer network, the hierarchical and domain-based management and control technology is adopted, effectively reducing the complexity of system management and control. By dividing different management domains and levels, more efficient management and control of network resources are achieved, solving the problem that traditional terrestrial centralized management and control strategies are difficult to handle large-scale networks;

[0184] 2) The information compression technology optimizes the utilization of satellite resources: In response to the challenge of limited satellite resources, the information compression technology is used to optimize the utilization of satellite resources. Through effective information compression and processing, the demand for satellite resources is reduced, the resource utilization efficiency is improved, and the best service quality is ensured under limited resources;

[0185] 3) Implementing dynamic resource allocation using a graph neural network model: With the help of graph neural network technology, fine-grained optimization of inter-domain resources and global load balancing are achieved. By constructing a network topology graph and applying graph neural network algorithms, the global resource status and service requirements can be grasped more accurately, enabling more refined resource allocation and optimization, and improving the utilization rate of network resources and service quality.

[0186] Based on the same technical concept, the embodiments of this application also provide network resource allocation devices for two slices. The principle of solving problems by the network resource allocation devices for slices is similar to the above-mentioned network resource allocation method for slices. Therefore, the implementation of the network resource allocation devices for slices can refer to the implementation of the network resource allocation method for slices, and the repeated parts will not be elaborated.

[0187] Figure 8 The following is a schematic structural diagram of a network resource allocation device for a slice provided by the embodiments of this application, including:

[0188] An acquisition module 801, configured to receive intra-domain resource information of multiple satellite regions;

[0189] A selection module 802, configured to select the regional paths of the at least one slice based on the intra-domain resource information of the multiple satellite regions and the requirement information of the at least one slice;

[0190] A sending module 803, configured to send a resource allocation instruction to the sub-controller corresponding to the regional path, so as to allocate network resources for the at least one slice within the satellite region corresponding to the sub-controller.

[0191] In some embodiments, the selection module 802 is specifically configured to:

[0192] Extract features from the intra-domain resource information of the multiple satellite regions to obtain the intra-domain resource features of the multiple satellite regions;

[0193] Perform inter-domain correlation analysis on the intra-domain resource information of the multiple satellite regions to obtain the inter-domain resource features of the multiple satellite regions;

[0194] Select the regional paths that match the requirement information of the at least one slice based on the intra-domain resource features and the inter-domain resource features.

[0195] In some embodiments, the selection module 802 is specifically configured to:

[0196] Initialize the state of a pre-trained graph neural network model based on the intra-domain resource features and the inter-domain resource features;

[0197] Input the demand information of the at least one slice into the graph neural network model in the order of decreasing slice priority, and obtain the regional paths of the at least one slice. After obtaining the regional path of each slice, update the state of the graph neural network model based on the network resources required by the slice.

[0198] In some embodiments, the graph neural network model calculates the probabilities of multiple alternative regional paths between the starting satellite region and the ending satellite region of each slice, and determines the regional path with the highest probability as the regional path of the target slice.

[0199] In some embodiments, determine the slice priority of a slice according to the following steps:

[0200] Calculate the index values of multiple first indicators of the slice, and perform normalization processing on the index values of the multiple first indicators;

[0201] Perform weighted summation on the normalized index values of the multiple first indicators to obtain the priority score of the slice;

[0202] Determine the slice priority of the slice according to the priority score of the slice and the pre-established correspondence between the priority score and the slice priority.

[0203] In some embodiments, the intra-domain resource information of the multiple satellite regions is sent after being quantified, dimension-reduced, and encoded by the multiple sub-controllers, and further includes a decoding module 804 for:

[0204] Before selecting the regional paths of the at least one slice by using the pre-trained graph neural network model based on the intra-domain resource information of the multiple satellite regions and the demand information of the at least one slice, perform decoding processing on the intra-domain resource information of the multiple satellite regions.

[0205] In some embodiments, each type of intra-domain resource information includes the index values of multiple second indicators. Quantify the intra-domain resource information according to the following steps:

[0206] According to the value dispersion degree of each second indicator and the pre-established correspondence between the value dispersion degree and the quantization method, determine the quantization method corresponding to the second indicator;

[0207] Quantify the index values of the second indicator according to the quantization method corresponding to the second indicator to obtain the quantization result of the second indicator.

[0208] In some embodiments, perform dimension reduction processing on each type of intra-domain resource information according to the following steps:

[0209] Divide the quantized in-domain resource information into multiple data blocks;

[0210] Use the principal component analysis algorithm to perform dimensionality reduction processing on the multiple data blocks.

[0211] In some embodiments, the following steps are used to perform encoding processing on each type of in-domain resource information:

[0212] For the multiple data blocks after dimensionality reduction, use the locality-sensitive hashing algorithm to perform hash encoding, and use binary encoding to perform binary encoding on the hash encoding result;

[0213] The decoding module 804 is specifically configured to, for the binary encoding result, use binary decoding to perform decoding, and use the locality-sensitive hashing algorithm to perform hash decoding on the binary decoding result to obtain the multiple data blocks after dimensionality reduction.

[0214] In some embodiments, any sub-controller is deployed on the backbone satellite in the corresponding satellite area, and the backbone satellite is selected according to the following steps:

[0215] Obtain the capability characterization information of each satellite in the satellite area;

[0216] According to the capability characterization information, score the service capabilities of the satellites;

[0217] According to the scoring results of the satellites, select one satellite from the satellites as the backbone satellite.

[0218] In some embodiments, the capability characterization information includes at least one of in-domain coverage, available bandwidth, delay, and central processing unit (CPU) load.

[0219] Figure 9 The structure diagram of another slice network resource allocation device provided by the embodiments of the present application includes:

[0220] The sending module 901 is configured to send the in-domain resource information of the corresponding satellite area to the master controller, and the master controller selects the area path of the at least one slice based on the in-domain resource information of multiple satellite areas and the demand information of the at least one slice, and sends a resource allocation instruction to the sub-controller corresponding to the area path;

[0221] The allocation module 902 is configured to, if receiving the resource allocation instruction, perform slice network resource allocation in the corresponding satellite area based on the resource allocation instruction.

[0222] In some embodiments, the allocation module 902 is specifically configured to:

[0223] Allocate network resources within the satellite area for corresponding slices in descending order of slice priority.

[0224] In some embodiments, determine the slice priority of a slice according to the following steps:

[0225] Calculate the index values of multiple first indicators of the slice, and normalize the index values of the multiple first indicators;

[0226] Perform weighted summation on the normalized index values of the multiple first indicators to obtain the priority score of the slice;

[0227] Determine the slice priority of the slice according to the priority score of the slice and the established correspondence between the priority score and the slice priority.

[0228] In some embodiments, it further includes:

[0229] The processing module 903 is configured to perform quantization, dimensionality reduction, and encoding processing on the intra-domain resource information before sending the intra-domain resource information of the corresponding satellite area to the master controller.

[0230] In some embodiments, the intra-domain resource information includes index values of multiple second indicators. The processing module 903 is specifically configured to perform quantization processing on the intra-domain resource information according to the following steps:

[0231] Determine the quantization method corresponding to the second indicator according to the value dispersion degree of each second indicator and the established correspondence between the value dispersion degree and the quantization method;

[0232] Quantize the index value of the second indicator according to the quantization method corresponding to the second indicator to obtain the quantization result of the second indicator.

[0233] In some embodiments, the processing module 903 is specifically configured to perform dimensionality reduction processing on the intra-domain resource information according to the following steps:

[0234] Divide the quantized intra-domain resource information into multiple data blocks;

[0235] Adopt the principal component analysis algorithm to perform dimensionality reduction processing on the multiple data blocks.

[0236] In some embodiments, the processing module 903 is specifically configured to perform encoding processing on each intra-domain resource information according to the following steps:

[0237] For the multiple data blocks after dimensionality reduction, perform hash encoding using the local hash-sensitive algorithm, and perform binary encoding on the hash encoding result using the binary encoding method.

[0238] In some embodiments, any sub - controller is deployed on the backbone satellite in the corresponding satellite area, and the backbone satellite is selected according to the following steps:

[0239] Obtain the capability characterization information of each satellite in the satellite area;

[0240] Score the service capabilities of each satellite according to the capability characterization information;

[0241] Select one satellite from each satellite as the backbone satellite according to the scoring results of each satellite.

[0242] In some embodiments, the capability characterization information includes at least one of in - domain coverage, available bandwidth, latency, and central processing unit (CPU) load.

[0243] The division of modules in the embodiments of the present application is illustrative. It is only a logical function division. In actual implementation, there may be other division methods. In addition, each functional module in the embodiments of the present application can be integrated in one processor, can exist separately physically, or two or more modules can be integrated in one module. The coupling between each module can be achieved through some interfaces, and these interfaces are usually electrical communication interfaces, but it does not exclude the possibility of being mechanical interfaces or other forms of interfaces. Therefore, the modules described as separate components may or may not be physically separated, and can be located in one place or distributed to different positions of the same or different devices. The above - integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0244] After introducing the network resource allocation method and device of the slice of the exemplary embodiment of the present application, next, an electronic device according to another exemplary embodiment of the present application is introduced.

[0245] Next, refer to Figure 10 to describe the electronic device 130 implemented according to this embodiment of the present application. Figure 10 The shown electronic device 130 is only an example and should not bring any limitations to the functions and usage scope of the embodiments of the present application.

[0246] As Figure 10 shown, the electronic device 130 is presented in the form of a general - purpose electronic device. The components of the electronic device 130 may include but are not limited to: the above - mentioned at least one processor 131, the above - mentioned at least one memory 132, and a bus 133 connecting different system components (including the memory 132 and the processor 131).

[0247] The bus 133 represents one or more of several types of bus architectures, including a memory bus or memory controller, a peripheral bus, a processor, or a local bus using any of the various bus architectures.

[0248] The memory 132 may include a readable medium in the form of volatile memory, such as random access memory (RAM) 1321 and / or cache memory 1322, and may further include read-only memory (ROM) 1323.

[0249] The memory 132 may also include a program / utilities 1325 having a set (at least one) of program modules 1324. Such program modules 1324 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment.

[0250] The electronic device 130 may also communicate with one or more external devices 134 (such as a keyboard, a pointing device, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 130, and / or may communicate with any device that enables the electronic device 130 to communicate with one or more other electronic devices (such as a router, a modem, etc.). Such communication may be through an input / output (I / O) interface 135. Also, the electronic device 130 may communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 136. As shown in the figure, the network adapter 136 communicates with other modules for the electronic device 130 through the bus 133. It should be understood that although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 130, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0251] In an exemplary embodiment, a storage medium is also provided. When the computer program in the storage medium is executed by a processor of an electronic device, the electronic device can execute the network resource allocation method of any of the above slices. Optionally, the storage medium may be a non-transitory computer-readable storage medium. For example, the non-transitory computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage devices, etc.

[0252] In an exemplary embodiment, the electronic device of the present application may at least include at least one processor and a memory communicatively connected to the at least one processor. Among them, the memory stores a computer program executable by the at least one processor. When the computer program is executed by the at least one processor, the at least one processor can execute the steps of the network resource allocation method for any slice provided by the embodiments of the present application.

[0253] In an exemplary embodiment, a computer program product is further provided. When the computer program product is executed by an electronic device, the electronic device can implement any exemplary method provided by the present application.

[0254] Moreover, the computer program product can adopt any combination of one or more readable media. The readable media can be a readable signal medium or a readable storage medium. The readable storage medium can, for example, but not be limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, a RAM, a ROM, an Erasable Programmable Read-Only Memory (EPROM), a flash memory, an optical fiber, a Compact Disk Read Only Memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0255] The program product for network resource allocation for slicing in the embodiments of the present application can adopt a CD-ROM and include program code, and can run on a computing device. However, the program product of the present application is not limited thereto. In this document, the readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device.

[0256] The readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries the readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable signal medium can also be any readable medium other than the readable storage medium, and this readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device.

[0257] The program code included on the readable medium can be transmitted by any appropriate medium, including but not limited to wireless, wired, optical cable, Radio Frequency (RF), etc., or any suitable combination of the above.

[0258] Program code for performing the operations of this application can be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and also including conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, executed as a stand-alone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user's computing device through any kind of network such as a Local Area Network (LAN) or a Wide Area Network (WAN), or it can be connected to an external computing device (e.g., by connecting through the Internet using an Internet service provider).

[0259] It should be noted that although several units or subunits of the device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more of the above-described units can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units.

[0260] In addition, although the operations of the method of this application are described in a specific order in the drawings, this does not require or imply that the operations must be performed in that specific order, or that all of the operations shown must be performed to achieve the desired result. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step for execution, and / or one step can be decomposed into multiple steps for execution.

[0261] Those skilled in the art should understand that the embodiments of this application can be provided as a method, a system, or a computer program product. Therefore, this application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0262] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, as well as the combination of flows and / or blocks in the flowchart and / or block diagram. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in one or more flows Figure 1 one or more flows and / or blocks Figure 1 or a plurality of blocks.

[0263] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions specified in one or more flows Figure 1 one or more flows and / or blocks Figure 1 or a plurality of blocks.

[0264] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows Figure 1 one or more flows and / or blocks Figure 1 or a plurality of blocks.

[0265] Although the preferred embodiments of the present application have been described, those skilled in the art can make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0266] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also includes these modifications and variations.

Claims

1. A sliced network resource allocation method, characterized in that Including: The total controller receives the in-domain resource information of multiple satellite regions; Based on the in-domain resource information of the multiple satellite regions and the requirement information of at least one slice, select the regional path of the at least one slice; Send a resource allocation instruction to the sub-controller corresponding to the regional path to allocate network resources for the at least one slice within the satellite region corresponding to the sub-controller; Among them, based on the in-domain resource information of the multiple satellite regions and the requirement information of at least one slice, selecting the regional path of the at least one slice includes: Extract features from the in-domain resource information of the multiple satellite regions to obtain the in-domain resource features of the multiple satellite regions; Conduct inter-domain correlation analysis on the in-domain resource information of the multiple satellite regions to obtain the inter-domain resource features of the multiple satellite regions; Initialize the state of a pre-trained graph neural network model based on the in-domain resource features and the inter-domain resource features. The graph neural network model is G = (V, E), where V is the set of all satellite regions in the satellite network and E is the set of association links between satellite regions. The vertex represents the in-domain resource features of the i-th satellite region, and the edge represents the inter-domain resource features between satellite region i and satellite region j; Input the requirement information of the at least one slice into the graph neural network model in the order of slice priority from high to low to obtain the regional path of the at least one slice. Among them, after obtaining the regional path of each slice, update the state of the graph neural network model based on the network resources required by the slice.

2. The method according to claim 1, characterized in that, The graph neural network model calculates the probabilities of multiple alternative regional paths between the starting satellite region and the ending satellite region of each slice, and determines the alternative regional path with the highest probability as the regional path of the slice.

3. The method according to claim 1, characterized in that Determine the slice priority of a slice according to the following steps: Calculate the index values of multiple first indicators of the slice, and perform normalization processing on the index values of the multiple first indicators; Perform weighted summation on the normalized index values of the multiple first indicators to obtain the priority score of the slice; Determine the slice priority of the slice according to the priority score of the slice and the pre-established correspondence between the priority score and the slice priority.

4. The method according to claim 1, characterized in that, The in-domain resource information of the multiple satellite regions is sent after quantization, dimensionality reduction, and encoding. It also includes: Before selecting the regional path of the at least one slice based on the in-domain resource information of the multiple satellite regions and the requirement information of at least one slice, perform decoding processing on the in-domain resource information of the multiple satellite regions.

5. The method according to claim 4, characterized in that Each type of in-domain resource information includes the index values of multiple second indicators. Quantize the in-domain resource information according to the following steps: Determine the quantization method corresponding to the second indicator according to the value dispersion degree of each second indicator and the pre-established correspondence between the value dispersion degree and the quantization method; Quantize the index values of the second indicator according to the quantization method corresponding to the second indicator to obtain the quantization result of the second indicator.

6. The method according to claim 4 or 5, characterized in that, Perform dimensionality reduction processing on each type of in-domain resource information according to the following steps: Divide the quantized in-domain resource information into multiple data blocks; Use the principal component analysis algorithm to perform dimensionality reduction processing on the multiple data blocks.

7. The method according to claim 6, wherein Perform encoding processing on each type of in-domain resource information according to the following steps: For the multiple data blocks after dimensionality reduction, use the local hashing sensitive algorithm for hashing encoding, and use the binary encoding method to perform binary encoding on the hashing encoding result; Decode the in-domain resource information of the multiple satellite regions, including: For the binary coding result, perform decoding using binary decoding method, and perform hash decoding on the binary decoding result using the local hash-sensitive algorithm to obtain multiple reduced-dimensional data blocks.

8. The method according to claim 1, characterized in that Any sub-controller is deployed on the backbone satellite in the corresponding satellite region, and the backbone satellite is selected according to the following steps: Obtain the capability characterization information of each satellite in the satellite region; According to the capability characterization information, score the service capabilities of the satellites; According to the scoring results of the satellites, select one satellite from the satellites as the backbone satellite.

9. The method according to claim 8, wherein, The capability characterization information includes at least one of the in-domain coverage range, available bandwidth, latency, and central processing unit (CPU) load.

10. A method for allocating sliced network resources, characterized in that, Including: The sub-controller sends the in-domain resource information of the corresponding satellite region to the master controller. The master controller selects the regional path of the at least one slice based on the in-domain resource information of the multiple satellite regions and the demand information of the at least one slice, and sends a resource allocation instruction to the sub-controller corresponding to the regional path; If the resource allocation instruction is received, based on the resource allocation instruction, perform slice network resource allocation in the corresponding satellite region; Among them, the master controller selects the regional path of the at least one slice based on the in-domain resource information of the multiple satellite regions and the demand information of the at least one slice in the following way: Extract features from the in-domain resource information of the multiple satellite regions to obtain the in-domain resource features of the multiple satellite regions; Perform inter-domain correlation analysis on the in-domain resource information of the multiple satellite regions to obtain the inter-domain resource features of the multiple satellite regions; Initialize the state of a pre-trained graph neural network model based on the in-domain resource features and the inter-domain resource features. The graph neural network model is G = (V, E), where V is the set of all satellite regions in the satellite network, and E is the set of association links between satellite regions. The vertex represents the in-domain resource features of the i-th satellite region, and the edge represents the inter-domain resource features between satellite region i and satellite region j; Input the demand information of the at least one slice into the graph neural network model in the order of slice priority from high to low to obtain the regional path of the at least one slice. Among them, after obtaining the regional path of each slice, update the state of the graph neural network model based on the network resources required by the slice.

11. The method according to claim 10, wherein, Based on the resource allocation instruction, perform slice network resource allocation in the corresponding satellite region, including: Allocate the network resources in the satellite region to the corresponding slice in the order of slice priority from high to low.

12. The method according to claim 11, wherein Determine the slice priority of a slice according to the following steps: Calculate the index values of multiple first indicators of the slice, and perform normalization processing on the index values of the multiple first indicators; Perform weighted summation on the normalized index values of the multiple first indicators to obtain the priority score of the slice; According to the priority score of the slice and the pre-established correspondence between the priority score and the slice priority, determine the slice priority of the slice.

13. The method according to claim 10, wherein Before sending the in-domain resource information of the corresponding satellite region to the master controller, it also includes: Perform quantization, dimensionality reduction, and coding processing on the in-domain resource information.

14. The method according to claim 13, wherein The in-domain resource information includes the index values of multiple second indicators. Quantize the in-domain resource information according to the following steps: Determine the quantization method corresponding to the second indicator according to the value dispersion degree of each second indicator and the pre-established correspondence between the value dispersion degree and the quantization method; Quantize the indicator value of the second indicator according to the quantization method corresponding to the second indicator to obtain the quantization result of the second indicator.

15. The method according to claim 13 or 14, wherein Perform dimensionality reduction processing on the in-domain resource information according to the following steps: Divide the quantized in-domain resource information into multiple data blocks; Use the principal component analysis algorithm to perform dimensionality reduction processing on the multiple data blocks.

16. The method according to claim 15, characterized in that, Perform encoding processing on each type of in-domain resource information according to the following steps: For the multiple data blocks after dimensionality reduction, use the local hashing sensitive algorithm for hashing encoding, and use the binary encoding method to perform binary encoding on the hashing encoding result.

17. The method according to claim 10, characterized in that Any sub-controller is deployed on the backbone satellite in the corresponding satellite area. Select the backbone satellite according to the following steps: Obtain the capability characterization information of each satellite in the satellite area; Score the service capabilities of the satellites according to the capability characterization information; Select a satellite from the satellites as the backbone satellite according to the scoring results of the satellites.

18. The method according to claim 17, wherein The capability characterization information includes at least one of the in-domain coverage range, available bandwidth, delay, and central processing unit (CPU) load.

19. A sliced network resource allocation device, characterized in that, Applied to the master controller, it includes: An acquisition module, configured to receive the in-domain resource information of multiple satellite areas; A selection module, configured to select the area path of the at least one slice based on the in-domain resource information of the multiple satellite areas and the requirement information of the at least one slice; A sending module, configured to send a resource allocation instruction to the sub-controller corresponding to the area path to allocate network resources for the at least one slice in the satellite area corresponding to the sub-controller; Among them, the selection module is specifically configured to: Extract features from the in-domain resource information of the multiple satellite areas to obtain the in-domain resource features of the multiple satellite areas; Perform inter-domain correlation analysis on the in-domain resource information of the multiple satellite areas to obtain the inter-domain resource features of the multiple satellite areas; Initialize the state of a pre-trained graph neural network model based on the intra-domain resource features and the inter-domain resource features. The graph neural network model is G = (V, E), where V is the set of all satellite regions in the satellite network, and E is the set of association links between satellite regions. The vertex represents the intra-domain resource features of the i-th satellite region, and the edge represents the inter-domain resource features between satellite region i and satellite region j; Input the requirement information of the at least one slice into the graph neural network model in the order of slice priority from high to low to obtain the area path of the at least one slice. After obtaining the area path of each slice, update the state of the graph neural network model based on the network resources required by the slice.

20. A sliced network resource allocation device, characterized in that, Applied to the sub-controller, it includes: A sending module, configured to send the in-domain resource information of the corresponding satellite area to the master controller. The master controller selects the area path of the at least one slice based on the in-domain resource information of the multiple satellite areas and the requirement information of the at least one slice, and sends a resource allocation instruction to the sub-controller corresponding to the area path; An allocation module, configured to, if receiving the resource allocation instruction, perform slice network resource allocation in the corresponding satellite area based on the resource allocation instruction; Among them, the master controller selects the area path of the at least one slice based on the in-domain resource information of the multiple satellite areas and the requirement information of the at least one slice in the following manner: Extract features from the in-domain resource information of the multiple satellite regions to obtain the in-domain resource features of the multiple satellite regions; Perform inter-domain correlation analysis on the in-domain resource information of the multiple satellite regions to obtain the inter-domain resource features of the multiple satellite regions; Initialize the state of a pre-trained graph neural network model based on the intra-domain resource characteristics and the inter-domain resource characteristics. The graph neural network model is G = (V, E), where V is the set of all satellite regions in the satellite network, and E is the set of association links between satellite regions. A vertex represents the intra-domain resource characteristics of the i-th satellite region, and an edge represents the inter-domain resource characteristics between satellite region i and satellite region j; Input the demand information of the at least one slice into the graph neural network model in the order of decreasing slice priority to obtain the regional path of the at least one slice. After obtaining the regional path of each slice, update the state of the graph neural network model based on the network resources required by the slice.

21. An electronic device, characterized in that, Comprising: At least one processor, and a memory communicatively connected to the at least one processor, wherein: The memory stores a computer program executable by the at least one processor. The computer program is executed by the at least one processor to enable the at least one processor to execute the method according to any one of claims 1-18.

22. A storage medium, characterized in that, When the computer program in the storage medium is executed by the processor of the electronic device, the electronic device can execute the method according to any one of claims 1-18.

23. A computer program product, characterized in that, Comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-18.

Images