Intelligent resource scheduling method for satellite-ground convergence network based on QoS (Quality of Service)

Abstract

The invention discloses an intelligent resource scheduling method for a satellite-ground convergence network based on QoS (Quality of Service), and relates to the technical field of communication. Comprising the following steps: carrying out virtualization processing on computing resources, storage resources and communication resources in the satellite-ground convergence network; carrying out resource matching and scheduling through the virtual resource management layer according to the QoS (Quality of Service) demand and the resource demand of the user; the on-demand distribution and dynamic adjustment of resources are realized by utilizing an optimization algorithm; and resource collaborative scheduling of the multi-domain network is realized through the cross-domain control system. According to the invention, the software defined network SDN and the network function virtualization NFV technology are utilized to realize intelligent management and allocation of multi-dimensional resources such as computing resources, storage resources, communication bandwidth and the like in the satellite-ground network, so that the resource utilization efficiency is improved, and diversified service requirements are met.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and more particularly to a QoS-based satellite-ground fusion network intelligent resource scheduling method. Background Art

[0002] With the rapid development of global satellite communication networks, satellite-ground integrated networks have become the key to improving communication efficiency. However, traditional satellite-ground network resource management methods have problems such as poor flexibility and low utilization. Especially in the face of complex and diverse business needs, the existing static resource allocation method cannot fully meet user needs.

[0003] In recent years, the introduction of software-defined networking (SDN) and network function virtualization (NFV) technologies has provided new ideas for solving the above problems. SDN technology allows the separation of network control and forwarding, providing flexible network control capabilities, while NFV technology virtualizes physical resources to make it possible to dynamically allocate network resources. Existing technologies still lack effective solutions for how to apply SDN and NFV technologies in satellite-ground converged networks to achieve on-demand resource scheduling; existing satellite-ground network resource scheduling methods are difficult to adapt to the dynamic changes in network status and business needs, and cannot achieve on-demand allocation; traditional centralized control architectures are difficult to cope with large-scale, dynamically changing topological structures in satellite-ground networks, resulting in low resource allocation efficiency; in satellite-ground converged networks, resource scheduling and coordination between different domains are difficult, and it is difficult to achieve QOS-based resource optimization in multi-domain, multi-business scenarios.

[0004] Therefore, it is an urgent problem for those skilled in the art to propose a QoS-based satellite-ground fusion network intelligent resource scheduling method to solve the difficulties existing in the prior art. Summary of the invention

[0005] In view of this, the present invention provides a QoS-based satellite-ground fusion network intelligent resource scheduling method, which uses SDN and NFV technologies to realize intelligent management and allocation of multi-dimensional resources such as computing resources, storage resources and communication bandwidth in the satellite-ground network, thereby improving resource utilization efficiency and meeting diverse business needs.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A QoS-based satellite-ground fusion network intelligent resource scheduling method comprises the following steps:

[0008] S1. Virtualize the computing resources, storage resources and communication resources in the satellite-ground fusion network;

[0009] S2: According to the user's QoS and resource requirements, resources are matched and scheduled through the virtual resource management layer;

[0010] S3, using optimization algorithms to achieve on-demand allocation and dynamic adjustment of resources;

[0011] S4. Realize the coordinated resource scheduling of multi-domain networks through cross-domain control systems.

[0012] In the above method, optionally, the satellite-ground fusion network in S1 adopts a four-layer architecture:

[0013] User / application layer: includes users of different business categories and various applications;

[0014] Virtual resource management layer: Based on software-defined network (SDN) technology, the resources in the satellite-to-ground network are managed through a virtual resource controller;

[0015] Virtual resource layer: abstractly maps computing resources, storage resources, and communication resources in the satellite-to-ground network through network function virtualization (NFV) technology;

[0016] Physical network layer: includes physical nodes and links in the interstellar network SN and the terrestrial network TN.

[0017] In the above method, optionally, the specific content of the virtualization processing in S1 is:

[0018] The computing resources, storage resources and communication resources in the satellite-ground fusion network are virtualized and mapped into a virtual resource pool;

[0019] Computing resources include: onboard processor resources, satellite CPU, FPGA, and ASIC computing capabilities;

[0020] Transmission resources include: communication links between satellites and ground networks, bandwidth, transmission rate and frequency band;

[0021] Storage resources include: Satellite memory for temporary storage of data and on-orbit data processing.

[0022] In the above method, optionally, the QoS requirements of the users in S2 include: latency, bandwidth, and computing power;

[0023] Resource matching and scheduling are performed through the virtual resource management layer, including:

[0024] Model the resource requirements of each user node and represent its resource requirements as a three-dimensional vector of communication, computation, and storage resources;

[0025] According to the needs of users and the status of the network, determine whether the service network is the interstellar network SN or the terrestrial network TN, and select nodes in the corresponding network to provide communication services for users;

[0026] The static resources in the satellite-ground fusion network are represented as a cube graph; the static resources include: computing resource cycle / s, storage resource Kbits, and communication resource bandwidth Kbps; the total amount of virtual resources allocated to a task is equal to the volume of one of the cubes, and the volume of the entire cube is the total resource constraint;

[0027] Determine the on-demand scheduling plan for resources within the domain, match virtual resources with physical network elements, and allocate resources to corresponding tasks within the volume constraints of the entire cube.

[0028] In the above method, optionally, the optimization algorithm in S3 includes: resource allocation algorithm and dynamic resource adjustment;

[0029] Resource allocation algorithm: realize on-demand scheduling of virtual resources through linear programming or dynamic programming;

[0030] Dynamic resource adjustment: adjust resource allocation in real time through feedback mechanism.

[0031] In the above method, optionally, the cross-domain control system in S4 adopts a centralized-distributed hierarchical and domain-based dynamic control architecture, which is composed of a main control system, a space-based control system, and a ground-based control system;

[0032] Main control system: Based on the global situation of the satellite-ground fusion network and the corresponding business needs, it generates the global optimal resource scheduling strategy to realize on-demand scheduling of end-to-end transmission resources. The main control system sends control strategies and QoS requirements to the space-based control system and the ground-based control system through the southbound interface to realize global control of the satellite-ground fusion network;

[0033] Space-based / ground-based control system: Based on the control strategy and QoS requirements issued by the main control system, it generates the optimal control strategy in the domain in combination with the network situation in the domain and sends it to the network elements in the domain through the southbound interface;

[0034] The main control system, space-based control system and ground-based control system provide logical interconnection between systems through north-south / east-west interfaces, respectively, to achieve multi-level control and cross-domain collaborative management.

[0035] It can be seen from the above technical solutions that, compared with the prior art, the present invention provides a QoS-based satellite-ground fusion network intelligent resource scheduling method, which has the following beneficial effects:

[0036] The present invention uses SDN and NFV technologies to achieve intelligent management and deployment of multi-dimensional resources such as computing resources, storage resources and communication bandwidth in satellite-to-ground networks, thereby improving resource utilization efficiency and meeting diverse business needs, reducing resource waste; achieving resource collaboration between networks in different domains and adapting to diverse business scenarios; and providing a flexible control architecture that can cope with dynamic changes in network topology and changes in service needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0038] Figure 1 A flow chart of a QoS-based satellite-ground fusion network intelligent resource scheduling method provided by the present invention;

[0039] Figure 2 This is a schematic diagram of the satellite-to-ground network system model architecture of the present invention;

[0040] Figure 3 A physical node and virtual resource mapping diagram of the present invention;

[0041] Figure 4 This is a schematic diagram of the satellite-ground fusion network resource allocation framework of the present invention;

[0042] Figure 5 It is a static resource cube representation diagram of the present invention;

[0043] Figure 6 It is a schematic diagram of the centralized-distributed hierarchical and domain-based dynamic control architecture of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] See also Figure 1 As shown, a QoS-based satellite-ground fusion network intelligent resource scheduling method includes the following steps:

[0046] S1. Virtualize the computing resources, storage resources and communication resources in the satellite-ground fusion network;

[0047] S2: According to the user's QoS and resource requirements, resources are matched and scheduled through the virtual resource management layer;

[0048] S3, using optimization algorithms to achieve on-demand allocation and dynamic adjustment of resources;

[0049] S4. Realize the coordinated resource scheduling of multi-domain networks through cross-domain control systems.

[0050] Further, see Figure 2 As shown in the figure, the satellite-ground fusion network in S1 adopts a four-layer architecture:

[0051] User / application layer: includes users of different business categories and various applications;

[0052] Virtual resource management layer: Based on software-defined network (SDN) technology, the resources in the satellite-to-ground network are managed through a virtual resource controller;

[0053] Virtual resource layer: abstractly maps computing resources, storage resources, and communication resources in the satellite-to-ground network through network function virtualization (NFV) technology;

[0054] Physical network layer: includes physical nodes and links in the interstellar network SN and the terrestrial network TN.

[0055] Further, refer to Figure 3 As shown, the specific contents of virtualization processing in S1 are:

[0056] The computing resources, storage resources and communication resources in the satellite-ground fusion network are virtualized and mapped into a virtual resource pool;

[0057] Computing resources include: onboard processor resources, satellite CPU, FPGA, and ASIC computing capabilities;

[0058] Transmission resources include: communication links between satellites and ground networks, bandwidth, transmission rate and frequency band;

[0059] Storage resources include: satellite memory, which is used for temporary storage of data and for processing data on orbit;

[0060] By building virtual resources to achieve the logical representation of resources, further integration forms a multi-resource collection called a virtual resource pool, which supports the upper layer to adopt centralized control and coordination to allocate resources, thereby supporting multi-system collaborative task planning. The mapping of physical nodes and virtual resources is as follows: Figure 3 shown.

[0061] Further, see Figure 4 As shown, the QoS requirements of users in S2 include: latency, bandwidth, and computing power;

[0062] Resource matching and scheduling are performed through the virtual resource management layer, including:

[0063] Model the resource requirements of each user node and represent its resource requirements as a three-dimensional vector of communication, computation, and storage resources;

[0064] According to the needs of users and the status of the network, determine whether the service network is the interstellar network SN or the terrestrial network TN, and select nodes in the corresponding network to provide communication services for users;

[0065] Reference Figure 5 As shown in the figure, the static resources in the satellite-ground fusion network are represented as a cubic graph; the static resources include: computing resource cycle / s, storage resource Kbits, and communication resource bandwidth Kbps; the total amount of virtual resources allocated to a task is equal to the volume of one of the cubes, and the volume size of the entire cube is the total resource constraint;

[0066] Determine the on-demand scheduling plan for resources within the domain, match virtual resources with physical network elements, and allocate resources to corresponding tasks within the volume constraints of the entire cube.

[0067] Furthermore, the optimization algorithms in S3 include: resource allocation algorithm and dynamic resource adjustment;

[0068] Resource allocation algorithm: virtual resources are scheduled on demand through an optimization algorithm (such as linear programming or dynamic programming). The goal of the algorithm is to minimize resource waste and ensure that the user's QoS requirements are met;

[0069] Dynamic resource adjustment: Taking into account the time-varying nature of network status, resource allocation is adjusted in real time through a feedback mechanism to ensure network robustness and service quality.

[0070] Further, see Figure 6 As shown, in S4, the resource collaborative scheduling of multi-domain networks is realized through the cross-domain control system, specifically:

[0071] In view of the multi-layer and multi-domain nature of the satellite-ground fusion network, the centralized-distributed hierarchical and domain-based dynamic control architecture can take into account the integration of satellite-ground access and space bearing, and has higher flexibility and scalability, so that the system can achieve high-precision real-time control. The specific architecture is shown in Figure 6; the architecture is mainly composed of the main control system, the space-based control system and the ground-based control system, which provide logical interconnection between systems through north-south / east-west interfaces, respectively, to achieve multi-level control and cross-domain collaborative management;

[0072] Main control system: Based on the global situation of the satellite-ground fusion network and the corresponding business needs, it generates the global optimal resource scheduling strategy to realize on-demand scheduling of end-to-end transmission resources. The main control system sends control strategies and QoS requirements to the space-based control system and the ground-based control system through the southbound interface to realize global control of the satellite-ground fusion network;

[0073] Space-based / ground-based control system: Based on the control strategy and QoS requirements issued by the main control system, it generates the optimal control strategy in the domain in combination with the network situation in the domain and sends it to the network elements in the domain through the southbound interface.

[0074] In a specific embodiment, in a certain satellite-ground fusion network, assuming that the service demand proposed by the user node is a low-latency and high-bandwidth communication task, the system first models the user's demand through a virtual resource controller and represents its resource demand as a set of three-dimensional vectors; then, the resource scheduling algorithm compares the resource status of the SN and TN networks and selects a node in the SN network to provide communication services for the user; finally, based on the user's latency requirements, the allocation ratio of computing resources and communication bandwidth is dynamically adjusted to ensure that QoS requirements are met.

[0075] First, the resource demand problem is modeled. For user node u j For example, its task requirements are defined as

[0076]

[0077] in, is a 3x1 column vector, the three elements in the vector Represents node u j The demand for total latency, the amount of data to be processed, and resource preferences due to different business needs, among which The elements in the vector represent the node u j Demand preferences for three types of resources.

[0078] Correspondingly, after being processed by the virtual resource controller, the STN network can j The service capability indicators provided are expressed as:

[0079]

[0080] in They represent the STN network for node u j Provides communication, computing, and storage resources.

[0081] For node u j , whose service delay can be defined as U j,TD =t j +f j , where tj ,f j They represent propagation delay and computational processing delay respectively.

[0082] Define node u j The distance between the nth TN network node is The distance between the mth SN network node is For the sake of brevity, unless otherwise specified, ln will be used in the following sections. j Represents user node u j The distance between the node and the service node, and thus the propagation delay can be calculated as

[0083] t j =θl nj

[0084] Here, θ is a scalar representing a weight factor.

[0085] Furthermore, node u j The computational processing delay is

[0086]

[0087] In summary, node u j The overall service delay should meet the following requirements:

[0088]

[0089] According to the above modeling constraints, node u j When the allocated resources meet their needs, the quantity and proportion of various resources can be flexibly and dynamically allocated and adjusted by the virtual resource controller according to the existing resource status in the virtual resource layer; that is, the virtual resource controller allocates all user nodes according to The allocated a j The total amount of resources is not greater than the total amount of virtual resources owned by the entire STN network, and the total delay calculated based on the allocated resources is not greater than the delay requirement in its business needs, then in theory, multiple resource allocation methods that meet the requirements can be found.

[0090] Therefore, when a certain type of resource in the virtual resource layer is scarce and cannot meet When , we can meet its latency and resource requirements by reasonably adding other types of resources, that is, there is a corresponding substitution relationship between resources. Therefore, we introduce a new objective function

[0091]

[0092] Among them, ||x||2 represents the norm-2 norm of vector x, and the goal is to minimize R(a j ), that is, the allocated resources a jWith each user node u j Resource Preferences The gap between them should be as small as possible;

[0093] Therefore, based on the above analysis, the problem of on-demand scheduling of satellite-ground fusion network resources can be modeled as follows:

[0094]

[0095] Since this section considers the intelligent matching mapping between domain resources and QoS requirements, it is only necessary to propose a unified algorithm strategy to meet the resource allocation in the satellite network and ground network domains respectively. Therefore, the original problem can be expressed as follows:

[0096]

[0097] Among them, F∈{S,N},K FN ∈{K SN , K TN},A=[a1,...,a J ].

[0098] According to the literature, the L-2 norm is a convex function. At the same time, the first three constraints are linear constraints, and the last constraint is a convex constraint. Therefore, the entire problem is a convex problem and can be solved directly using CVX.

[0099] Since the virtual resource pool integrates the physical resources of each actual physical network element, the resource scheduling plan obtained based on the virtual resource pool capacity may result in a situation where the resource supply that a physical network element should bear exceeds the physical resources it actually owns. Therefore, further adjustments need to be made to the given initial resource scheduling plan.

[0100] After the physical network resources are abstracted into virtual resources, the total virtual resources occupied by the user nodes they serve must not exceed the physical resources of the physical network element nodes. Therefore, the static resource block allocation must meet the following requirements:

[0101]

[0102] Where j∈K STN ↑t n,n' , j∈K STN ↑s m,m' and j∈K STN ↑s n,m Indicates that the link resources allocated to the user node are mapped to t n,n' ,s m,m' or Physical link resources, j∈K TN ↑t n and j∈K SN↑s m Represents K TN and K SN Mapped to node t n and m physical resources.

[0103] According to user resource preferences and STN network resource status, after executing the intra-domain resource allocation algorithm, the virtual resource manager matches each user node to its corresponding optimal STN network node, and obtains a preliminary stable many-to-many matching.

[0104] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0105] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A QoS-based satellite-ground fusion network intelligent resource scheduling method, characterized in that: The following steps are involved: S1. Virtualize the computing resources, storage resources and communication resources in the satellite-ground fusion network; S2: According to the user's QoS and resource requirements, resources are matched and scheduled through the virtual resource management layer; S3, using optimization algorithms to achieve on-demand allocation and dynamic adjustment of resources; S4. Realize the coordinated resource scheduling of multi-domain networks through cross-domain control systems.

2. The method for intelligent resource scheduling of satellite-ground fusion network based on QoS according to claim 1, characterized in that: The satellite-ground fusion network in S1 adopts a four-layer architecture: User / application layer: includes users of different business categories and various applications; Virtual resource management layer: Based on software-defined network (SDN) technology, the resources in the satellite-to-ground network are managed through a virtual resource controller; Virtual resource layer: abstractly maps computing resources, storage resources, and communication resources in the satellite-to-ground network through network function virtualization (NFV) technology; Physical network layer: includes physical nodes and links in the interstellar network SN and the terrestrial network TN.

3. The method for intelligent resource scheduling of satellite-ground fusion network based on QoS according to claim 2, characterized in that: The specific contents of virtualization processing in S1 are as follows: The computing resources, storage resources and communication resources in the satellite-ground fusion network are virtualized and mapped into a virtual resource pool; Computing resources include: onboard processor resources, satellite CPU, FPGA, and ASIC computing capabilities; Transmission resources include: communication links between satellites and ground networks, bandwidth, transmission rate and frequency band; Storage resources include: Satellite memory for temporary storage of data and on-orbit data processing.

4. The method for intelligent resource scheduling of satellite-ground fusion network based on QoS according to claim 1, characterized in that: The QoS requirements of users in S2 include: latency, bandwidth, and computing power; Resource matching and scheduling are performed through the virtual resource management layer, including: Model the resource requirements of each user node and represent its resource requirements as a three-dimensional vector of communication, computation, and storage resources; According to the needs of users and the status of the network, determine whether the service network is the interstellar network SN or the terrestrial network TN, and select nodes in the corresponding network to provide communication services for users; The static resources in the satellite-ground fusion network are represented as a cube graph; the static resources include: computing resource cycle / s, storage resource Kbits, and communication resource bandwidth Kbps; the total amount of virtual resources allocated to a task is equal to the volume of one of the cubes, and the volume of the entire cube is the total resource constraint; Determine the on-demand scheduling plan for resources within the domain, match virtual resources with physical network elements, and allocate resources to corresponding tasks within the volume constraints of the entire cube.

5. The method for intelligent resource scheduling of satellite-ground fusion network based on QoS according to claim 1, characterized in that: The optimization algorithms in S3 include: resource allocation algorithm and dynamic resource adjustment; Resource allocation algorithm: realize on-demand scheduling of virtual resources through linear programming or dynamic programming; Dynamic resource adjustment: adjust resource allocation in real time through feedback mechanism.

6. The method for intelligent resource scheduling of satellite-ground fusion network based on QoS according to claim 1, characterized in that: The cross-domain control system in S4 adopts a centralized-distributed hierarchical and domain-based dynamic control architecture, which consists of a main control system, a space-based control system, and a ground-based control system; Main control system: Based on the global situation of the satellite-ground fusion network and the corresponding business needs, it generates the global optimal resource scheduling strategy to realize on-demand scheduling of end-to-end transmission resources. The main control system sends control strategies and QoS requirements to the space-based control system and the ground-based control system through the southbound interface to realize global control of the satellite-ground fusion network; Space-based / ground-based control system: Based on the control strategy and QoS requirements issued by the main control system, it generates the optimal control strategy in the domain in combination with the network situation in the domain and sends it to the network elements in the domain through the southbound interface; The main control system, space-based control system and ground-based control system provide logical interconnection between systems through north-south / east-west interfaces, respectively, to achieve multi-level control and cross-domain collaborative management.

Images