Task-aware service dynamic deployment method

Abstract

The invention relates to a task-aware service dynamic deployment method, and belongs to the technical field of wireless communication. According to the method, a multifunctional time extension graph model is constructed, and a service request model, a time delay model and an energy cost model of a task are constructed on the model, so that effective management and control of network resources are realized. Meanwhile, a self-adaptive ant colony optimization algorithm based on dynamic volatilization is adopted, and heuristic weights are adjusted in real time according to service quality requirements of tasks, so that path selection and virtual network function deployment better meet task requirements, the utilization rate of network resources and the service completion rate are improved, system energy consumption is effectively reduced, and diversified requirements of different task flows are met.

Description

Technical Field

[0001] The invention belongs to the technical field of wireless communications and relates to a task-aware service dynamic deployment method. Background Art

[0002] With the rapid development of communication technology, traditional ground networks have achieved high capacity and low latency, and can be applied to application scenarios such as industrial Internet of Things and Internet of Vehicles. However, the infrastructure of traditional ground networks is fixedly deployed, so the network lacks flexibility and has limited coverage. Compared with ground networks, space and air networks are suitable for remote data perception, collection and dissemination with large coverage. However, the storage resources and computing resources of air nodes are scarce, making it difficult to efficiently process intensive computing tasks. The construction of air-space-ground network realizes the comprehensive integration of space network, air network and ground network. The air-space-ground network fully complements the advantages of the three network segments and provides many advantages such as wide coverage, high reliability, strong flexibility and high bandwidth. However, the air-space-ground network is highly dynamic and heterogeneous, which poses major challenges to network operators in terms of traffic distribution, routing protocol design and load balancing. In such a large-scale and dynamic network scenario, traditional proprietary network architecture and dedicated hardware can no longer cost-effectively and reliably orchestrate network resources for multi-dimensional business needs.

[0003] New network technologies such as Software-Defined Networking (SDN) and Network Function Virtualization (NFV) provide effective strategies to deal with the management and control issues of large-scale dynamic heterogeneous networks. The introduction of SDN / VNF in air-ground-space networks is crucial to improving network performance and enhancing the flexibility of management and control. With the continuous expansion of application scenarios, how to meet the diverse user QoS requirements in a complex and dynamically changing network environment has become one of the key and difficult points of current research. Under the SDN / VNF architecture, the virtualized network function (VNF) components are orderly combined in the form of logical links to form a service function chain (SFC), and then the task flow is guided to transmit through the SFC according to specific strategies to complete the service of the task flow. For air-ground-space networks, this architecture provides an effective technical approach to achieve dynamic resource management and diversified improvement of QoS.

[0004] In the air-ground-space network based on SDN / VNF, how to deploy SFC to improve resource utilization and ensure the QoS requirements of diverse tasks is an important research topic. In the ground network, the deployment of SFC has been studied more maturely. However, the deployment of SFC in the ground network is only for static network topology. In the air-ground-space network environment, how to effectively coordinate and allocate network resources to meet the QoS of different task flows, especially in the face of high load and dynamic changes, is still a challenge. Therefore, designing a task-aware service dynamic deployment method that can meet the diverse needs of users, improve the task completion rate and reduce the total cost of the network is of great research significance. Summary of the invention

[0005] In view of this, an object of the present invention is to provide a task-aware service dynamic deployment method.

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

[0007] A task-aware service dynamic deployment method, the method comprising the following steps:

[0008] S1: In the space-ground network that combines software-defined networking and network function virtualization, a multifunctional time-extended graph model is constructed for the dynamic heterogeneity of the network, and the service request model, delay model and energy cost model of the task are constructed on this model;

[0009] S2: Constructing a mission-aware service deployment model in air-space-ground networks;

[0010] S3: A dynamic volatilization-based adaptive ant colony optimization algorithm is proposed, which dynamically adjusts the heuristic weights according to the service quality requirements of the task, so that the path selection and virtual network function deployment are more in line with the task requirements.

[0011] Further, in S1, based on the air-ground-space network combining software-defined networking and network function virtualization, a multifunctional time-extended graph model is constructed for the dynamic heterogeneity of the network, specifically:

[0012] In the space-ground network that combines software-defined networking and network function virtualization, the network state within the time range T is considered, and the time range T = [t 0 ,t Q ) is divided into Q consecutive time slots {τ 1 ,τ 2 ,…,τ q ,…,τ Q}, the network topology in time slot τ q can be considered as static. In the static time slot τ qIn the time slot τ, the connection relationship between the UAV, satellite and ground station nodes is mapped into the multifunctional time expansion graph, which includes the connection between UAV to satellite, UAV to ground station, satellite to satellite, satellite to ground station, ground station to satellite, and ground station to ground station. q and τ q+1 There is a storage link between each node for saving the information of the previous time slot. In this model, drones are considered as nodes for collecting ground tasks, and ground stations and satellites are considered as task processing nodes. Therefore, in the time-extended graph model, VNFs are deployed on satellite nodes and ground station nodes, and these nodes are virtualized into multiple virtual function nodes using virtualization technology. Each VNF is deployed on a different virtual function node to achieve the description of the same node embedded with multiple VNF scenarios.

[0013] Furthermore, in S1, according to the constructed function time extension graph model, a service request model of the task is constructed on the model, specifically:

[0014] With binary variables To represent the time slot τ q Internal Task Flow k Whether the cth VNF is embedded in the virtual function node In, it is expressed as:

[0015]

[0016] With binary variables To represent the time slot τ q Internal Task Flow k Whether the link (N i ,N j ), expressed as:

[0017]

[0018] Furthermore, in S1, according to the constructed functional time extension graph model, a task delay model and an energy cost model are constructed on the model, specifically:

[0019] The delay of each task mainly considers communication delay, calculation delay and propagation delay.

[0020] For LEO satellite and ground station nodes, the energy consumed mainly considers the energy consumption of data transceivers and data processing. For UAV source nodes, they are not embedded as VNFs, so they do not process received tasks, and the energy consumed mainly considers the energy consumption of data transmitters.

[0021] Furthermore, in S2, a task-aware service deployment model in the air-ground-space network is constructed to minimize the total energy cost of the network while satisfying the task QoS, specifically:

[0022]

[0023] in, N F,VP ,1q Q,1M M i} represents task flow l k The cth VNF Whether it is embedded in the virtual function node middle; E F ,1q Q} represents the task flow l k Whether through the link (N i ,N j )∈E F , For task flow k Delay requirements; represents the total energy cost of LEO satellite nodes, represents the total energy cost of the drone node, represents the total energy cost of the ground station node; E F,AS represents the transmission links of all input LEO satellite nodes, E F,SA Represents the transmission links of all output LEO satellite nodes; E F,AG and E F,GA Respectively represent the transmission links of all input and output ground station nodes. Constraint C1 is the capacity constraint of the transmission link; C2 is the computing resource constraint of the node; C3 is the delay constraint; C4 and C5 are traffic conservation constraints; C6 is the VNF deployment constraint; C7 is the constraint of binary variables X and Y.

[0024] Furthermore, in S3, a dynamic volatilization-based adaptive ant colony optimization algorithm is used, specifically:

[0025] At the beginning of each iteration, the pheromone volatility coefficient ρ is dynamically adjusted according to the current number of iterations k and the maximum number of iterations K. k , to increase exploration in the early stages of the search and focus on the optimal path in the later stages.

[0026] Place M ants at the starting node of the task flow. Based on the QoS heuristic mechanism, its heuristic function can be expressed as:

[0027]

[0028] Among them, Capacity(i,j) represents the link capacity from node i to node j, which is used to ensure bandwidth requirements; Delay(i,j) represents the transmission delay of the link, which mainly considers the transmission and propagation delays to select low-delay links; Reliability(i,j) represents the reliability of the link transmission task; w capacity 、w delay With w Reliability They are the weights of each factor, and the sum of the weights is 1. By dynamically adjusting the weights, the optimal balance is found between different task flows and network segments, and the optimal path selection and VNF deployment are completed.

[0029] The beneficial effects of the present invention are: by constructing a multifunctional time-extended graph model, combined with software-defined networking and network function virtualization technology, efficient deployment and management of task flows in air-space-ground networks are achieved. A dynamic volatile adaptive ant colony optimization algorithm is used to adjust the heuristic weights in real time based on the service quality requirements of the task, so that path selection and virtual network function deployment are more in line with task requirements. This method significantly improves the utilization rate and service completion rate of network resources, effectively reduces system energy consumption, and meets the diverse needs of different task flows. With its adaptability and resource optimization effect in dynamic heterogeneous environments, the invention has a wide range of application value in task-intensive scenarios.

[0030] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0032] Figure 1 It is a network system framework diagram of the present invention;

[0033] Figure 2 It is a multifunctional time expansion diagram based on the network system framework in the present invention;

[0034] Figure 3 A flow chart for the dynamic deployment of the service of the present invention. DETAILED DESCRIPTION

[0035] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0036] Among them, the drawings are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on the present invention. In order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0037] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "front", "rear", etc. indicate the orientation or position relationship, they are based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0038] See also Figure 1 to Figure 3 The task-aware service dynamic deployment method of the present invention specifically comprises the following steps:

[0039] Step 1: In the air-ground-space network that combines software-defined networking and network function virtualization, a multifunctional time-extended graph model is constructed according to the dynamic heterogeneity of the network, and the service request model, delay model and energy cost model of the task are constructed on this model;

[0040] Step 1.1: In the air-ground-space network that combines software-defined networking and network function virtualization, consider the network status within the time range T, and set the time range T = [t 0 ,t Q ) is divided into Q consecutive time slots {τ 1 ,τ 2 ,…,τ q ,…,τ Q}, the network topology in time slot τ q can be considered as static. In the static time slot τq In the time slot τ, the connection relationship between the UAV, satellite and ground station nodes is mapped into the multifunctional time expansion graph, which includes the connection between UAV to satellite, UAV to ground station, satellite to satellite, satellite to ground station, ground station to satellite, and ground station to ground station. q and τ q+1 There is a storage link between each node for saving the information of the previous time slot. In this model, drones are considered as nodes for collecting ground tasks, and ground stations and satellites are considered as task processing nodes. Therefore, in the time-extended graph model, VNFs are deployed on satellite nodes and ground station nodes, and these nodes are virtualized into multiple virtual function nodes using virtualization technology. Each VNF is deployed on a different virtual function node to achieve the description of the same node embedded with multiple VNF scenarios.

[0041] Step 1.2: Using binary variables To represent the time slot τ q Internal Task Flow k Whether the cth VNF is embedded in the virtual function node In, it is expressed as:

[0042]

[0043] With binary variables To represent the time slot τ q Internal Task Flow k Whether the link (N i ,N j ), expressed as:

[0044]

[0045] Step 1.3: The delay of each task mainly considers communication delay, calculation delay and propagation delay, using Represents the task flow l k Transmission delay on the transmission link, Indicates the computational delay of the virtual function node in processing the task. Represents the task flow l k The propagation delays are expressed as:

[0046]

[0047] in For task flow; k The data size, Transmission link in the current time slot The transmission rate on E F,TE is the transmission link in the multifunctional time expansion diagram, C congestionis the real-time congestion coefficient, with a value range of 0 to 1, where 0 indicates no congestion and 1 indicates extreme congestion. k is an adjustment coefficient that can adjust the degree of impact of congestion on communication delay according to actual conditions; Indicates that in time slot τ q Internal Task Flow k The amount of data computed by the cth VNF in The node corresponding to this time slot The computing power, N F,G and N F,S They represent the ground node and LEO satellite node in the multifunctional time expansion diagram, respectively. scheduling It is an intelligent resource scheduling factor, which dynamically adjusts the computing resource allocation method according to the real-time computing resource idle rate of the node and the complexity of the task. The value range is between 0 and 2, where 0 means that the resource allocation is extremely unreasonable and 2 means that the resource allocation reaches the optimal state, which can greatly improve the computing efficiency. F,A represents the air link set, Represents the time slot τ q Internal Task Flow k The corresponding transmission link through which the flow passes length, c is the speed of light in vacuum, T topo It is the network topology dynamic coefficient, and its value range is 0 to 1. 0 means that the topology is stable, and 1 means that the topology has undergone significant changes.

[0048] Step 1.4: For LEO satellite nodes, the energy consumed mainly considers the energy consumption of the data transceiver and the energy consumption of data processing. The energy cost of the LEO satellite receiver is expressed as The energy cost of the transmitter can be expressed as The energy cost of data processing on LEO satellite nodes is expressed as The total energy cost of the LEO satellite is:

[0049]

[0050] For the UAV source node, it is not embedded as a VNF, so it does not process the received tasks. The energy consumed by it mainly considers the energy consumption of the data transmitter. The energy cost of its transmitter is expressed as use To represent the basic energy cost of UAV to maintain operation, the total energy cost of the UAV node is:

[0051]

[0052] For ground nodes, the energy consumed mainly comes from the energy consumption of data transceivers and the energy consumption of data processing. The energy cost of the ground station receiver is expressed as The energy cost of the ground station transmitter is expressed as When the ground node can be used as a VNF embedded node to process the task, the energy cost of its data processing is expressed as use represents the energy cost per unit computing resource on the ground node, represents the basic energy cost required to maintain the normal operation of the ground node, then the total energy cost of the ground node is:

[0053]

[0054] Step 1.5: Construct the service deployment model of task awareness in the air-ground network as follows:

[0055]

[0056] in, N F,VP ,1q Q,1M M i} represents the task flow; k The cth VNF Whether it is embedded in the virtual function node middle; E F ,1q Q} represents the task flow; k Whether through the link (N i ,N j )∈E F , For task flow; k The delay requirement of E F,AS represents the transmission links of all input LEO satellite nodes, E F,SA Represents the transmission links of all output LEO satellite nodes; E F,AG and E F,GA Respectively represent the transmission links of all input and output ground station nodes. Constraint C1 is the capacity constraint of the transmission link; C2 is the computing resource constraint of the node; C3 is the delay constraint; C4 and C5 are traffic conservation constraints; C6 is the VNF deployment constraint; C7 is the constraint of binary variables X and Y.

[0057] Step 1.6: For the proposed dynamic volatilization-based adaptive ant colony optimization algorithm:

[0058] (1) Initialization settings:

[0059] Set the parameters of the dynamic volatile adaptive ant colony optimization algorithm, including the number of ants M, the maximum number of iterations K, the heuristic factor α, the expectation factor β, the initial pheromone concentration τ 0 and the initial volatility coefficient ρ initialInput the service request of the task flow and its QoS requirements, and define the service function chain of each task flow, that is, the sequence of virtual network functions required by the task flow.

[0060] (2) Dynamic adjustment of pheromones:

[0061] At the beginning of each iteration, the pheromone volatility coefficient ρ is dynamically adjusted according to the current number of iterations k and the maximum number of iterations K. k , to increase exploration in the early stages of the search and focus on the optimal path later:

[0062]

[0063] (3) Path selection and VNF deployment:

[0064] Place M ants at the starting point of the task flow. Based on the QoS heuristic mechanism, its heuristic function can be expressed as:

[0065]

[0066] Among them, Capacity(i,j) represents the link capacity from node i to node j, which is used to ensure bandwidth requirements; Delay(i,j) represents the transmission delay of the link, which mainly considers the transmission and propagation delays to select low-delay links; Reliability(i,j) represents the reliability of the link transmission task; w capacity 、w delay With w Reliability They are the weights of each factor, and the sum of the weights is 1. By dynamically adjusting the weights, the optimal balance is found between different task flows and network segments, and the optimal path selection and VNF deployment are completed.

[0067] Ants use heuristic information and pheromone concentration τ ij Calculate the path selection probability and select the next node according to the probability to complete the path deployment of the task flow. On the selected path, deploy VNFs according to the computing resources of the nodes, and give priority to nodes with sufficient resources to meet QoS requirements. Record the deployment plan and update the computing and transmission delay of the task flow according to the selected node resources.

[0068] (4)Pheromone update:

[0069] After each ant has selected a path and deployed VNF, it will perform a local pheromone update on the path it has passed. The pheromone update formula is:

[0070] τ ij =(1-ρ k )×τ ij +Δτ ij

[0071] where Δτ ij is the pheromone increment of the path, which is related to the QoS satisfaction and energy consumption of the path.

[0072] After each iteration, all paths found by ants are evaluated, and the best path is globally updated with pheromones to increase its pheromone concentration and improve the subsequent ants' preference for the preferred path.

[0073] (5) Record the optimal path and VNF deployment plan:

[0074] After each iteration, record the optimal path found in this iteration and its corresponding VNF deployment plan. If the optimal path meets the QoS requirements of the task flow and the global cost is the smallest, save the plan. Determine whether the convergence condition is met. If so, end the iteration.

[0075] (6) Output the optimal SFC deployment plan:

[0076] After the iteration, the final SFC deployment plan is output, including the optimal path for each task flow, the VNF deployment location on the path, the total resource consumption, and the QoS satisfaction.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A task-aware service dynamic deployment method, characterized by: The following steps are involved: S1: In the space-ground network that combines software-defined networking and network function virtualization, a multifunctional time-extended graph model is constructed for the dynamic heterogeneity of the network, and the service request model, delay model and energy cost model of the task are constructed on this model; S2: Construct a task-aware service dynamic deployment model in air-ground network; S3: A dynamic volatility-based adaptive ant colony optimization algorithm is used to solve the task-aware service dynamic deployment model in the air-ground network.

2. According to claim 1, a task-aware service dynamic deployment method is characterized by: In S1, a multifunctional time-extended graph model is constructed for the dynamic heterogeneity of the network in combination with the air-ground network of software-defined networking and network function virtualization, specifically: In the space-ground network that combines software-defined networking and network function virtualization, the network state within the time range T is considered, and the time range T = [t0,t Q ) is divided into Q consecutive time slots {τ1,τ2,…,τ q ,…,τ Q }, the network topology in time slot τ q is considered static; in the static time slot τ q In the time slot τ, the connection relationship between the UAV, satellite and ground station nodes is mapped into the multifunctional time expansion graph, which includes the connection between UAV to satellite, UAV to ground station, satellite to satellite, satellite to ground station, ground station to satellite, and ground station to ground station. q and τ q+1 Between each node, there is a storage link for saving the information of the previous time slot; in this model, the UAV is used as the node for collecting ground tasks, and the ground station and satellite are used as task processing nodes; In the time-extended graph model, VNFs are deployed on satellite nodes and ground station nodes, and virtualization technology is used to virtualize these nodes into multiple virtual function nodes. Each VNF is deployed on a different virtual function node, realizing the description of the scenario of embedding multiple VNFs in the same node.

3. The task-aware service dynamic deployment method according to claim 2, characterized in that: In S1, according to the constructed function-time extended graph model, a service request model of the task is constructed on the model, specifically: With binary variables To represent the time slot τ q Internal Task Flow k Whether the cth VNF is embedded in the virtual function node In, it is expressed as: With binary variables To represent the time slot τ q Internal Task Flow k Whether the link (N i ,N j ), expressed as:

4. The task-aware service dynamic deployment method according to claim 3, characterized in that: In S1, according to the constructed functional time extension graph model, a task delay model and an energy cost model are constructed on the model, specifically: For each task’s delay, communication delay, computation delay, and propagation delay are considered; For LEO satellites and ground station nodes, the energy consumed includes the energy consumption of data transceivers and the energy consumption of data processing; For the UAV source node, it does not serve as an embedded VNF and does not process the received tasks, and its energy consumption includes the energy consumption of the data transmitter.

5. The task-aware service dynamic deployment method according to claim 4, characterized in that: In S2, a task-aware service deployment model in the air-ground-space network is constructed to minimize the total energy cost of the network while satisfying the task QoS, specifically: C7:X,Y∈{0,1} in, Represents the task flow l k The cth Whether it is embedded in the virtual function node middle; Represents the task flow l k Whether through the link (N i ,N j )∈E F , For task flow k Delay requirements; represents the total energy cost of LEO satellite nodes, represents the total energy cost of the drone node, represents the total energy cost of the ground station node; E F,AS represents the transmission links of all input LEO satellite nodes, E F,SA Represents the transmission links of all output LEO satellite nodes; E F,AG and E F,GA They represent the transmission links of all input and output ground station nodes respectively; constraint C1 is the capacity constraint of the transmission link; constraint C2 is the computing resource constraint of the node; constraint C3 is the delay constraint; constraints C4 and C5 are traffic conservation constraints; constraint C6 is the VNF deployment constraint; constraint C7 is the constraint of binary variables X and Y.

6. A task-aware service dynamic deployment method according to claim 5, characterized in that: In S3, a dynamic volatilization-based adaptive ant colony optimization algorithm is used, specifically: At the beginning of each iteration, the pheromone volatility coefficient ρ is dynamically adjusted according to the current number of iterations k and the maximum number of iterations K. k , to increase exploration in the early stages of the search and focus on the optimal path in the later stages; Place M ants at the starting node of the task flow; Based on the QoS heuristic mechanism, its heuristic function is expressed as: Among them, Capacity(i,j) represents the link capacity from node i to node j, which is used to ensure bandwidth requirements; Delay(i,j) represents the transmission delay of the link, which takes into account the transmission and propagation delays to select low-delay links; Reliability(i,j) represents the reliability of the link transmission task; w capacity 、w delay With w Reliability They are the weights of each factor, and the sum of the weights is 1. By dynamically adjusting the weights, the optimal balance is found between different task flows and network segments, and the optimal path selection and VNF deployment are completed.

Images