Method, apparatus and device for deploying service function chain, and storage medium

Abstract

The application relates to a service function chain deployment method, device and equipment and a storage medium, and relates to the field of multi-domain network arrangement. The method comprises the following steps: obtaining a target service function sub-chain of a target domain network based on a chain cutting scheme of a service function chain; under the constraint of a first constraint condition, generating a plurality of candidate paths corresponding to the target service function sub-chain; taking the minimization of in-domain communication delay and resource occupation rate standard deviation as an optimization target, generating an optimal VNF instantiation scheme corresponding to each candidate path through a particle swarm optimization algorithm; determining a target path and an optimal VNF instantiation scheme of the target path based on the optimal VNF instantiation scheme corresponding to each candidate path; and deploying the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path. Through the above method, the allocation effect of an in-domain link is improved, and the stability and efficiency of the whole network are improved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the field of multi-domain network orchestration, and in particular to a service function chain deployment method and device, equipment and a storage medium. BACKGROUND

[0002] Network Function Virtualization (NFV) can achieve flexible deployment of network services on servers. A network service in NFV can be represented as a service function chain (SFC), which is composed of a series of virtual network functions (VNFs) in a given order. In order to provide efficient services, the SFC is deployed in multiple different servers and guides traffic through a set of VNFs in a specific order.

[0003] In the NFV network, the SFC can be flexibly deployed, but the flexible path selection and VNF deployment scheme selection also make the SFC deployment problem an NP-hard problem. SUMMARY

[0004] Embodiments of the present application provide a service function chain deployment method, device, equipment and storage medium, which can improve the allocation effect of intra-domain links, achieve balanced allocation of VNFs on each physical node, and improve the stability and efficiency of the network as a whole. The technical solution is as follows.

[0005] On the one hand, a service function chain deployment method is provided, which is executed by an intra-domain controller of a target domain network, the target domain network being any one of a plurality of domain networks in a multi-domain network, and the method comprising:

[0006] obtaining a target service function sub-chain of the target domain network based on a chain-cutting scheme of the service function chain; the chain-cutting scheme is used to indicate that the service function chain is cut into a plurality of service function sub-chains corresponding to the multi-domain network; the target service function sub-chain contains resource requirements of each virtual network function (VNF) allocated to the target domain network;

[0007] under the constraint of a first constraint condition, generating a plurality of candidate paths corresponding to the target service function sub-chain; the first constraint condition is constructed based on physical nodes in the target domain network and resource remaining amounts of each physical node;

[0008] The particle swarm optimization algorithm is used to generate an optimal VNF instantiation scheme corresponding to each candidate path, with the optimization objective of minimizing the intra-domain communication latency and the standard deviation of resource occupancy.

[0009] Based on the optimal VNF instantiation scheme corresponding to each candidate path, a target path and an optimal VNF instantiation scheme of the target path are determined.

[0010] Based on the target path and the optimal VNF instantiation scheme of the target path, the target service function sub-chain is deployed in the target domain network.

[0011] In another aspect, a method for deploying a service function chain is provided, which is executed by a master controller of a multi-domain network including at least two domain networks, and includes:

[0012] A chain splitting scheme of the service function chain is generated, which is used to indicate that the service function chain is split into a plurality of service function sub-chains corresponding to the multi-domain network.

[0013] The chain splitting scheme is sent to an intra-domain controller of each domain network, so that each intra-domain controller obtains a corresponding service function sub-chain, and for an intra-domain controller of any target domain network, a plurality of candidate paths corresponding to a target service function sub-chain are generated under the constraint of a first constraint condition, and an optimal VNF instantiation scheme corresponding to each candidate path is generated by a particle swarm optimization algorithm with the optimization objective of minimizing the intra-domain communication latency and the standard deviation of resource occupancy, and based on the optimal VNF instantiation scheme corresponding to each candidate path, a target path and an optimal VNF instantiation scheme of the target path are determined, and based on the target path and the optimal VNF instantiation scheme of the target path, the target service function sub-chain is deployed in the target domain network.

[0014] In another aspect, a device for deploying a service function chain is provided, which is applied in an intra-domain controller of a target domain network, and the target domain network is any domain network in a multi-domain network, and the device includes:

[0015] A link acquisition module is configured to acquire a target service function sub-chain of the target domain network based on a chain splitting scheme of a service function chain, the chain splitting scheme is used to indicate that the service function chain is split into a plurality of service function sub-chains corresponding to the multi-domain network, and the target service function sub-chain includes resource requirements of each virtual network function (VNF) allocated to the target domain network.

[0016] a path generation module, configured to generate a plurality of candidate paths corresponding to the target service function sub-chain under constraints of first constraint conditions; the first constraint conditions are constructed based on physical nodes in the target domain network and resource remaining amounts of the respective physical nodes;

[0017] a first scheme generation module, configured to generate, by a particle swarm optimization algorithm, an optimal VNF instantiation scheme corresponding to each candidate path as an optimization target of minimizing intra-domain communication latency and resource occupation rate standard deviation; the VNF instantiation scheme is used to indicate VNF deployment of each physical node on the candidate path;

[0018] a determination module, configured to determine a target path and an optimal VNF instantiation scheme of the target path based on the optimal VNF instantiation scheme corresponding to each candidate path;

[0019] a deployment module, configured to deploy the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path.

[0020] In a possible implementation, the first scheme generation module includes:

[0021] an initialization submodule, configured to initialize the particle swarm optimization algorithm based on a target candidate path; a particle population in the particle swarm optimization algorithm is a set of VNF instantiation schemes on the corresponding candidate path, and one particle in the particle population represents one VNF instantiation scheme; the target candidate path is any one of the candidate paths;

[0022] an iterative adjustment submodule, configured to perform iterative adjustment of particle positions of the respective particles based on a particle speed updating mechanism in the particle swarm optimization algorithm until the optimization target is met, so as to obtain the optimal VNF instantiation scheme corresponding to the target candidate path.

[0023] In a possible implementation, the iterative adjustment submodule is configured to,

[0024] in each iterative adjustment process, calculate fitness values of the respective particles by using a fitness function, and the fitness value is a comprehensive value of the intra-domain communication latency and the resource occupation rate standard deviation corresponding to the particle;

[0025] the VNF instantiation scheme represented by the particle with the minimum fitness value in the iterative adjustment is determined as the optimal VNF instantiation scheme corresponding to the target candidate path.

[0026] In a possible implementation, the iterative adjustment submodule is configured to,

[0027] Under constraint of the second constraint condition, each particle is adjusted in particle position iteration based on a particle speed updating mechanism in the particle swarm optimization algorithm;

[0028] The second constraint condition comprises:

[0029] The physical node of the VNF deployment is within a path range of the target candidate path;

[0030] An execution order between each VNF remains unchanged.

[0031] In a possible implementation, the path generation module comprises:

[0032] A node acquisition submodule is configured to acquire a source node and a destination node of the target service function sub-chain; the source node is a routing node connected with a previous domain network of the target service function sub-chain; the destination node is a routing node connected with a next domain network of the target service function sub-chain;

[0033] A path generation submodule is configured to generate a shortest path based on the source node and the destination node of the target service function sub-chain under constraint of the first constraint condition by using a KSP algorithm, to obtain a plurality of candidate paths corresponding to the target service function sub-chain.

[0034] In a possible implementation, the first constraint condition comprises:

[0035] The VNF is indivisible;

[0036] At most one VNF is deployed on each physical node;

[0037] The candidate path passes through a target physical node, which is a physical node for providing a VNF service for the target service function sub-chain;

[0038] A total resource consumption amount corresponding to the candidate path is less than or equal to a resource remaining amount of the target domain network.

[0039] In another aspect, a deployment apparatus of a service function chain is provided, the apparatus being applied in a master controller of a multi-domain network, the multi-domain network comprising at least two domain networks, and the apparatus comprising:

[0040] A second scheme generation module is configured to generate a chain-cutting scheme of a service function chain; the chain-cutting scheme is used to indicate that the service function chain is cut into a plurality of service function sub-chains corresponding to the multi-domain network;

[0041] The scheme sending module is configured to send the chain-cutting scheme to the in-domain controllers of each domain network, so that each in-domain controller obtains a corresponding service function sub-chain, and for the in-domain controller of any one target domain network, under the constraint of the first constraint condition, a plurality of candidate paths corresponding to the target service function sub-chain are generated; taking minimization of in-domain communication latency and resource occupation rate standard deviation as an optimization objective, an optimal VNF instantiation scheme corresponding to each candidate path is generated by using a particle swarm optimization algorithm, the target path and an optimal VNF instantiation scheme of the target path are determined based on the optimal VNF instantiation scheme corresponding to each candidate path, and the target path and the optimal VNF instantiation scheme of the target path are determined based on the optimal VNF instantiation scheme corresponding to each candidate path; and the target service function sub-chain is deployed in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path.

[0042] In a possible implementation, the apparatus further includes:

[0043] The request receiving module is configured to receive a service function chain deployment request before the second scheme generating module generates the chain-cutting scheme of the service function chain, and the service function chain deployment request includes the number and execution order of VNFs, the source node and the destination node of the service function chain, and the resource requirement amount of each VNF.

[0044] The resource amount obtaining module is configured to obtain the inter-domain communication latency between each domain network and the resource remaining amount of each domain network.

[0045] The link determining module is configured to determine a service function chain from the source node to the destination node of the service function chain based on the inter-domain communication latency between each domain network, the resource remaining amount of each domain network, and the resource requirement amount of each VNF, and the total inter-domain communication latency of the service function chain is the shortest.

[0046] In another aspect, a computer device is provided, which includes a processor and a memory, and the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement any of the above service function chain deployment methods.

[0047] In another aspect, a computer readable storage medium is provided, which stores at least one computer program, and the computer program is loaded and executed by a processor to implement any of the above service function chain deployment methods.

[0048] In another aspect, a computer program product is provided, which includes at least one computer program loaded and executed by a processor to implement the deployment method of any service function chain provided in the above various optional implementations.

[0049] The technical solutions provided in the present application can include the following beneficial effects:

[0050] The deployment method of the service function chain provided in the embodiments of the present application can fully combine the actual physical node resource status of the target domain network to generate reasonable candidate paths, avoid the occurrence of deployment schemes that are not feasible due to resource limitations, and take into account both the intra-domain communication delay and the resource occupation rate standard deviation when determining the VNF instantiation scheme, so that the network can ensure the timeliness of data transmission and improve the allocation effect of intra-domain links during operation, achieve balanced allocation of VNFs on each physical node, and improve the stability and efficiency of the network as a whole.

[0051] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0052] The accompanying drawings incorporated in and forming a part of the specification, illustrate embodiments consistent with the present application and serve to explain the principles of the present application, together with the description.

[0053] Figure 1 A flowchart of the deployment method of the service function chain provided in an exemplary embodiment of the present application is shown;

[0054] Figure 2 A flowchart of the deployment method of the service function chain provided in another exemplary embodiment of the present application is shown;

[0055] Figure 3 A schematic diagram of the multi-domain network architecture provided in an exemplary embodiment of the present application is shown;

[0056] Figure 4 A flow chart of a method for deploying a service function chain is shown according to an example embodiment of the present application;

[0057] Figure 5 A flow chart of a process for optimal path selection based on KSP algorithm is shown according to an example embodiment of the present application;

[0058] Figure 6 A flow chart of a process for generating an optimal VNF scheme of candidate paths is shown according to an example embodiment of the present application;

[0059] Figure 7 A block diagram of a device for deploying a service function chain is shown according to an example embodiment of the present application;

[0060] Figure 8 A block diagram of a device for deploying a service function chain is shown according to another example embodiment of the present application;

[0061] Figure 9 A structural block diagram of a computer device is shown according to an example embodiment of the present application;

[0062] Figure 10 A structural block diagram of a computer device is shown according to another example embodiment of the present application. DETAILED DESCRIPTION

[0063] The example embodiments will be described in detail herein with reference to the attached drawings. When the description below refers to accompanying drawings, unless otherwise stated, the same numbers in different drawings refer to the same or similar elements. The embodiments described in the following example embodiments are not representative of all embodiments consistent with the present application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the present application as detailed in the appended claims.

[0064] The embodiments of the present application provide a deployment method of a service function chain, which relates to cutting of a service function chain and deployment of a service function sub-chain in a domain, Figure 1 A flow chart of a method for deploying a service function chain is shown according to an example embodiment of the present application, which is performed by a domain controller in a target domain network, the target domain network being any one of domain networks in a multi-domain network, as shown in Figure 1 The method can include the following steps:

[0065] In step 110, a target service function sub-chain of the target domain network is obtained based on a chain-cutting scheme of the service function chain; the chain-cutting scheme is used to indicate that the service function chain is cut into a plurality of service function sub-chains corresponding to the multi-domain network; the target service function sub-chain contains resource requirements of each virtual network function (VNF) allocated to the target domain network.

[0066] Service function chain is a network service architecture concept, which concatenates a series of virtual network functions in a specific order to achieve customized processing of network traffic, such as firewall detection, intrusion detection, traffic optimization and other different functions of data packets in turn, to ensure that network services can meet diversified needs. Virtual network function is to realize the function of traditional physical network equipment (such as firewall, router, etc.) through software virtualization, running on general server hardware platform, which can be flexibly deployed and allocated, providing basic components for building flexible network services.

[0067] The chain cutting scheme of service function chain is a scheme for dividing the complete service function chain according to certain rules and strategies to adapt to multi-domain network environment, the purpose is to divide the whole service function chain into multiple parts, so that each part (i.e. service function sub-chain) can correspond to different domain networks for subsequent processing, which is convenient for multi-domain to complete the overall network service function.

[0068] After the domain controller of the target domain network obtains the chain cutting scheme of the service function chain, it can determine the target service function sub-chain belonging to itself from the chain cutting scheme. The target service function sub-chain records the resource demand amount of each virtual network function allocated to the target domain network, including the computing resource demand amount and the bandwidth resource demand amount. The computing resource demand amount is a quantitative index of the consumed computing power resource required to implement the virtual network function, such as CPU core number, etc. The bandwidth resource demand amount is a quantitative value of the network bandwidth resource required to implement the virtual network function, which can be measured by the amount of data transmitted per second, representing the size of network bandwidth expected to be used by VNF in unit time.

[0069] Step 120, under the constraint of the first constraint condition, a plurality of candidate paths corresponding to the target service function sub-chain are generated; the first constraint condition is constructed based on the physical nodes in the target domain network and the resource remaining amount of each physical node.

[0070] The first constraint condition is a restriction rule constructed by comprehensively considering the physical node situation in the target domain network and the remaining available resource amount of each physical node, to ensure that the generated candidate paths are within the actual resource carrying capacity of the target domain network, avoid generating paths that do not conform to the actual resource situation, and ensure the feasibility of subsequent path selection and VNF deployment.

[0071] The physical node is a network device node in the network environment, such as router, server, switch and other hardware devices, which is the actual carrier of VNF and network traffic transmission.

[0072] After the target service function sub-chain is acquired, the in-domain controller generates a plurality of candidate paths corresponding to the target service function sub-chain according to the first constraint condition. In one possible implementation, each candidate path needs to satisfy that the number of physical nodes passed through is greater than or equal to the number of VNFs allocated to the target domain network, so as to ensure that there are sufficient physical nodes to carry the VNFs, and to provide a plurality of possible selection paths for subsequent reasonable instantiation of the VNFs, thereby facilitating selection of an optimal solution.

[0073] In step 130, an optimal VNF instantiation scheme corresponding to each candidate path is generated by a particle swarm optimization algorithm, with minimization of in-domain communication latency and resource occupancy standard deviation as the optimization target. The VNF instantiation scheme is used to indicate the VNF deployment of each physical node on the candidate path.

[0074] The in-domain communication latency refers to the time delay of network data transmission from a source point to a target point through a series of physical nodes in the target domain network. The communication latency is affected by multiple factors such as processing capability of the physical nodes, link bandwidth, network congestion, etc. The smaller the communication latency, the higher the real-time performance and service quality of the network.

[0075] The resource occupancy standard deviation is a statistical index for measuring the dispersion degree of resource occupancy of each physical node, i.e., an index for measuring the load balancing state of the network. For example, there are differences in the occupancy proportions of different physical nodes for computing resources and bandwidth resources, etc. The standard deviation can reflect the balancing degree of such resource allocation. The smaller the standard deviation, the more balanced the resource allocation on each physical node, which is beneficial to stable and efficient operation of the entire network.

[0076] The particle swarm optimization algorithm (PSO) is a kind of optimization algorithm based on swarm intelligence, which simulates the characteristics of bird foraging and other group behaviors. Multiple particles (which can be analogous to different candidate path schemes in this scenario) move and update their own positions (i.e., continuously adjust the VNF instantiation scheme corresponding to the candidate path) in the search space. That is, the VNF scheme is a scheme for planning the VNF deployment of each physical node on the candidate path, which specifies which VNFs should be deployed on each physical node in the candidate path, and the specific configuration parameters (such as the amount of allocated computing resources, the required network bandwidth quota, etc.) of each VNF. According to the pre-set optimization target (here, the minimization of in-domain communication latency and resource occupancy standard deviation), the optimal solution is found, and the relatively optimal VNF instantiation scheme under each candidate path is finally determined.

[0077] For each generated candidate path, a particle swarm optimization algorithm is used to generate the corresponding optimal VNF instantiation scheme, with the optimization objectives of minimizing the intra-domain communication latency and the standard deviation of resource occupancy. In this process, each candidate path can be regarded as a search space, and the particles in the particle swarm optimization algorithm (representing different VNF deployment attempts) constantly explore and adjust in the search space. According to the resource conditions of each physical node on the current path, the connection relationship between nodes, and other information, different VNF deployment location combinations (i.e., different VNF instantiation schemes) are tried. Through repeated iterative calculations, the VNF instantiation scheme that makes the intra-domain communication latency as small as possible and the standard deviation of resource occupancy as small as possible is found. In this way, the performance of the optimized candidate path is more in line with the requirements of efficient and stable network operation. Finally, the optimal VNF instantiation scheme corresponding to each path is obtained.

[0078] Step 140, based on the optimal VNF instantiation scheme corresponding to each candidate path, determine the target path and the optimal VNF instantiation scheme of the target path.

[0079] After obtaining the optimal VNF instantiation scheme corresponding to each candidate path through the particle swarm optimization algorithm, these candidate paths need to be compared and a most suitable target path and its corresponding optimal VNF instantiation scheme are determined. This determination process can be a comprehensive comparison of the performance of each candidate path in terms of communication latency, resource occupancy balance, etc. For example, the values of the intra-domain communication latency and the standard deviation of resource occupancy corresponding to different candidate paths are compared, and the candidate path with the best overall performance is selected as the target path, and its corresponding VNF instantiation scheme is determined as the optimal VNF instantiation scheme of the target path, preparing for subsequent actual deployment in the target domain network.

[0080] Step 150, based on the target path and the optimal VNF instantiation scheme of the target path, deploy the target service function sub-chain in the target domain network.

[0081] Based on the determined target path and the optimal VNF instantiation scheme of the target path, the intra-domain controller of the target domain network performs actual deployment operations within the target domain network, i.e., according to the optimal VNF instantiation scheme, the corresponding VNF is deployed to each physical node involved in the target path, ensuring that the target service function sub-chain can normally operate in the target domain network in the expected manner, implementing its corresponding function in the entire multi-domain network service function chain system, thereby guaranteeing the service quality and business processing capacity of the entire multi-domain network.

[0082] In summary, the service function chain deployment method provided by the embodiments of the present application, the domain controller of any one target domain network in the multi-domain network obtains a target service function sub-chain of the target domain network from the chain cutting scheme of the service function chain, generates a plurality of candidate paths corresponding to the target service function sub-chain under the constraint of the first constraint condition constructed by comprehensively considering the target domain network physical nodes and the resource remaining amount of each physical node, then takes the minimization of the domain communication time delay and the resource occupation rate standard deviation as the optimization target, generates the optimal VNF scheme corresponding to each candidate path by means of the particle swarm optimization algorithm, and then determines the optimal target path of the target domain network and the optimal VNF scheme corresponding thereto from the optimal VNF scheme, and deploys the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path. Through the above method, reasonable candidate paths can be generated by fully combining the actual physical node resource status of the target domain network, and deployment schemes that are not feasible due to resource limitations can be avoided. When determining the VNF instantiation scheme, both the domain communication time delay and the resource occupation rate standard deviation are taken into account, the performance of the scheme is determined by the particle swarm optimization algorithm, the timeliness of data transmission can be ensured and the allocation effect of the intra-domain link can be improved in the running process of the network, the balanced allocation of VNF on each physical node is realized, and the stability and efficiency of the network as a whole are improved.

[0083] The service function chain deployment method provided by the present application includes two stages of service function chain cutting scheme formulation and intra-domain service function sub-chain deployment. When formulating the service function chain cutting scheme, the generation of the chain cutting scheme can be constrained by the load balancing of the domain network, Figure 2 A flowchart of the service function chain deployment method provided by another example embodiment of the present application is shown. The method can be used for service function chain cutting and intra-domain service function sub-chain deployment, as shown in Figure 2 The method can include the following steps.

[0084] Step 210, generating a chain cutting scheme of the service function chain; the chain cutting scheme is used to indicate that the service function chain is divided into a plurality of service function sub-chains corresponding to the multi-domain network.

[0085] In a possible implementation, the master controller can generate a chain cutting scheme of the service function chain with load balancing as a constraint. The number of service function sub-chains in the chain cutting scheme is the same as the number of domain networks in the multi-domain network.

[0086] Load balancing is a resource allocation strategy that evenly distributes network load (such as network traffic and computing tasks) across multiple network resources (such as servers and links), preventing performance degradation due to overuse of some resources while others remain idle. In the process of generating service function chain slicing schemes, load balancing as a constraint means ensuring that the resource allocation of each service function sub-chain after slicing is relatively balanced across different domain networks. This allows each domain network to reasonably undertake a portion of the service function chain's workload, preventing any domain network from being overloaded or underloaded, thereby guaranteeing the stable operation and efficient service of the entire multi-domain network.

[0087] In this embodiment, the main controller can divide the service function chain according to the number of domain networks, and the division formula can be expressed as:

[0088]

[0089] Where k represents the number of domain networks, and the domain networks are ordered. This represents the first step in achieving network load balancing. The amount of computing resources that a domain network should provide. Indicates the first The amount of computing resources remaining in a domain network. This represents the remaining computing resources of the entire domain network. The ratio of this to the remaining resources can be used to determine the share of resources that the domain network should provide. This indicates the total amount of computing resources required for this service function chain.

[0090] Suppose the computational resource requirements of each VNF in a service function chain of length M are expressed as follows: To reduce inter-domain link bandwidth and latency, the chain-splitting algorithm in this embodiment comprehensively considers the remaining computing resources of each domain network and each VNF, dividing the SFC into multiple sub-chains equal to the number of domain networks, thereby determining the SFC sub-chain allocated to each domain network. While ensuring that the number of service function sub-chains is the same as the number of domain networks, the computing resource requirements of each VNF are traversed sequentially. and Then it can be determined that the service function subchain assigned to the first domain network is from the first VNF ​​to the second VNF. One VNF, and so on, ultimately achieving maximum load balancing across multiple domain networks and reasonable SFC partitioning.

[0091] Step 220, send the chain cutting scheme to the in-domain controllers of each domain network, so that each in-domain controller obtains the corresponding service function sub-chain, for any one in-domain controller of a target domain network, generate a plurality of candidate paths corresponding to the target service function sub-chain under the constraint of the first constraint condition; taking the minimization of the in-domain communication delay and the resource occupation rate standard deviation as the optimization objective, generate the optimal VNF instantiation scheme corresponding to each candidate path respectively through the particle swarm optimization algorithm, determine the target path and the optimal VNF instantiation scheme of the target path based on the optimal VNF instantiation scheme corresponding to each candidate path respectively, determine the target path and the optimal VNF instantiation scheme of the target path based on the optimal VNF instantiation scheme corresponding to each candidate path respectively; deploy the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path.

[0092] The master controller sends the chain cutting scheme to the in-domain controllers of each domain network, and after receiving the chain cutting scheme, the in-domain controllers of each domain network can perform in-domain service function sub-chain deployment based on the scheme shown in the embodiment shown in Figure 1 After the deployment of the corresponding service function sub-chain in each domain network is completed, the deployment of the service function chain is completed. The above process of determining the service function chain can consider the inter-domain communication delay between each domain network, the resource remaining amount of each domain network, and the resource demand amount of each VNF in the service function chain in a multi-domain network environment. After receiving a service function chain deployment request containing the number of VNFs, the execution order, and the source node and destination node of the service function chain, the inter-domain communication delay between each domain network and the resource remaining amount of each domain network are obtained. Based on this information, the service function chain with the shortest total inter-domain communication delay from the source node to the destination node of the service function chain is determined. The chain cutting scheme for indicating that the service function chain is cut into a plurality of service function sub-chains corresponding to the multi-domain network is generated under the constraint of load balancing. This can effectively improve the rationality and efficiency of resource utilization of the entire multi-domain network, enhance the ability of the multi-domain network to cope with diversified business loads, and ensure the stability of the entire network system operation.

[0093] In a possible implementation, before generating the chain cutting scheme of the service function chain, the master controller is further configured to, after receiving the service function chain deployment request, generate the corresponding service function chain, and the process can be implemented as follows:

[0094] S201, receive a service function chain deployment request, which contains the number and execution order of VNFs, the source node and destination node of the service function chain, and the resource demand amount of each VNF.

[0095] The service function chain deployment request can be initiated by an external system or a user, and is a set of instruction information for creating and deploying a specific service function chain in a multi-domain network environment. The service function chain deployment request contains parameters of the service function chain, so that the multi-domain network can perform subsequent resource planning and deployment operations according to the parameters. The number of VNFs specifies the number of VNF components that need to be connected in the entire service function chain. The execution order of VNFs specifies the order of network traffic when flowing through each VNF. Different execution orders will produce different network service effects. The source node and destination node information of the service function chain determines the entry and exit positions of the network traffic. The resource requirement of each VNF, including the computing resource requirement (such as the number of CPU cores and memory capacity required for each VNF to run) and the bandwidth resource requirement (the network bandwidth size required by each VNF when processing data traffic), etc., provides a basis for subsequent reasonable allocation of resources and planning of service function chain paths in the multi-domain network.

[0096] The source node is the starting point of network traffic entering the service function chain processing flow. It can be a specific server, router interface or other network access device in the network. The destination node is the final output point of network traffic after a series of virtual network functions (VNFs) in the service function chain. After the data packet enters the service function chain through the source node and is processed in the specified VNF order, it finally flows to the destination node, completes the entire service function chain processing flow and enters the subsequent network transmission link.

[0097] In one possible implementation, the main control network can first model the underlying physical network (including intra-domain and inter-domain network topology) to define the underlying physical network as a weighted undirected graph wherein, N represents a set of physical network nodes, E represents a set of links between physical nodes. The resource capacity of the physical node is represented as For a physical link , the bandwidth resource and the delay are represented as and . Figure 3 A schematic diagram of a multi-domain network architecture provided by an example embodiment of the present application is shown as Figure 3As shown, taking five domain networks in a multi-domain network as an example, in the process of establishing a network, physical nodes include two types of nodes, i.e., routing nodes and processing nodes, the routing nodes are responsible for forwarding data streams between various domain networks, and the processing nodes are responsible for providing virtual network functions, each of which can contain at least two routing nodes; the underlying physical network contains multiple networks and inter-domain links, in the embodiment of the present application, the network can be established by representing the network structure of the th domain network, wherein represents a set of nodes in the th domain network, represents a set of links in the th domain network, and represents a set of inter-domain links. Therefore, the entire multi-domain network can be represented by a set of various domain networks and inter-domain links ; a service function chain is defined as a linear chain composed of virtual network functions (VNFs) and links therebetween in sequence, and is represented as , wherein respectively represent the source node and the destination node of the service function chain, and the source node and the destination node of the SFC are initialized according to the nodes in the network at the beginning; represents the requirement of the SFC for network functions, including the number and sequence of virtual network functions (VNFs) required by the SFC, and the computing resources required by each VNF, and can be represented as ; respectively represent the requirement of the SFC for bandwidth resources, although each VNF has different requirements for the computing resources of the physical node, each virtual link of a service function chain has the same bandwidth resource requirement.

[0098] Then, the main controller can construct a service function chain cross-domain deployment problem based on the underlying modeling of the multi-domain network and the service function chain, which includes two sub-problems of service function chain cutting and intra-domain sub-chain deployment,

[0099] S202, obtaining the inter-domain communication delay between various domain networks and the resource remaining amount of each domain network.

[0100] The inter-domain communication delay refers to the time delay generated when data is transmitted between different domain networks under the multi-domain network architecture; since different domain networks may differ in geographical location, network topology structure, network device performance, etc., data transmission across domain boundaries will be affected by various factors such as link bandwidth, intermediate node processing speed, transmission distance, etc., which comprehensively lead to the generation of inter-domain communication delay, and the inter-domain communication delay is an important indicator for measuring the efficiency of the cooperative work of the multi-domain network.

[0101] For each domain network, the resource residual quantity represents the quantity of various network resources currently available for allocation and use in the domain network, and the resource residual quantity includes a computing resource residual quantity (such as a number of idle CPU cores, unoccupied memory space, etc.) and a network bandwidth resource residual quantity (such as a bandwidth capacity not used by each link).

[0102] In S203, a service function chain from a source node to a destination node of the service function chain is determined based on the inter-domain communication delay between the various domain networks, the resource residual quantity of each domain network, and the resource requirement quantity of each VNF.

[0103] The main controller determines the service function chain according to the inter-domain communication delay between the various networks, the resource residual quantity of each domain network, and the resource requirement quantity of each VNF in the service function chain deployment request. When determining the service function chain, the main controller can obtain the optimal inter-domain path from the source node to the destination node, that is, the service function chain with the shortest total inter-domain communication delay, according to the inter-domain connectivity and the inter-domain communication delay, by traversing the feasible paths, and obtain the network domains through which the service function chain passes.

[0104] In summary, the service function chain deployment method provided by the embodiments of the present application can generate a chain splitting scheme of a service function chain by a main control network, send the chain splitting scheme to each domain network, obtain a target service function sub-chain of a target domain network from the chain splitting scheme of the service function chain by an in-domain controller of any target domain network in the multi-domain network, generate a plurality of candidate paths corresponding to the target service function sub-chain under the constraint of a first constraint condition constructed by considering the target domain network physical nodes and the resource residual quantity of each physical node, then generate an optimal VNF scheme corresponding to each candidate path by taking the minimization of the intra-domain communication delay and the resource occupation rate standard deviation as an optimization target and by means of a particle swarm optimization algorithm, determine the optimal target path of the target domain network and the optimal VNF scheme corresponding to the target path from the optimal VNF schemes, and deploy the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path. Through the above method, reasonable candidate paths can be generated by fully considering the actual physical node resource status of the target domain network, and deployment schemes that are not feasible due to resource limitations can be avoided. When determining the VNF instantiation scheme, both the intra-domain communication delay and the resource occupation rate standard deviation are taken into account, and a scheme with better performance is determined by means of the particle swarm optimization algorithm, so that the network can ensure the timeliness of data transmission and improve the allocation effect of the intra-domain link during operation, achieve the balanced allocation of VNFs on the physical nodes, and improve the stability and efficiency of the network as a whole.

[0105] Based on Figure 1 With Figure 2The illustrated embodiment, Figure 4 A flow chart of a method for deploying a service function chain is shown, which can be performed by a master controller of a multi-domain network and an intra-domain controller of each domain network, as shown in Figure 4 The method can include the following steps.

[0106] Step 401, receiving a service function chain deployment request, the service function chain deployment request containing the number and execution order of VNFs, the source node and the destination node of the service function chain, and the resource requirement of each VNF.

[0107] Step 402, obtaining the inter-domain communication delay between each domain network and the resource remaining amount of each domain network.

[0108] Step 403, based on the inter-domain communication delay between each domain network, the resource remaining amount of each domain network and the resource requirement of each VNF, determining a service function chain from the source node to the destination node of the service function chain; the total inter-domain communication delay of the service function chain is the shortest.

[0109] Step 404, generating a chain cutting scheme of the service function chain; the chain cutting scheme is used to indicate that the service function chain is cut into a plurality of service function sub-chains corresponding to the multi-domain network.

[0110] Wherein, the chain cutting scheme of the service function chain is generated with load balancing as a constraint, and the number of service function sub-chains is the same as the number of domain networks in the multi-domain network.

[0111] Step 405, sending the chain cutting scheme to the intra-domain controller of each domain network, and correspondingly, each intra-domain controller receives the chain cutting scheme of the service function chain.

[0112] The related content of steps 401 to 405 can refer to the description of the corresponding steps of the Figure 2 The illustrated embodiment, which will not be described here.

[0113] For each domain network, after receiving the chain cutting scheme of the service function chain, the corresponding service function sub-chain is deployed; taking the deployment process of the service function sub-chain of any one target network as an example, the service function sub-chain deployment process of each domain network can be implemented as:

[0114] Step 406, obtaining the target service function sub-chain of the target domain network based on the chain cutting scheme of the service function chain; the target service function sub-chain contains the resource requirement of each virtual network function VNF allocated to the target domain network.

[0115] Step 407, under the constraint of the first constraint condition, a plurality of candidate paths corresponding to the target service function sub-chain are generated; the first constraint condition is constructed based on physical nodes in the target domain network and resource remaining amounts of respective physical nodes.

[0116] The first constraint condition can include:

[0117] VNFs are indivisible;

[0118] At most one VNF is deployed on each physical node;

[0119] The candidate path passes through a target physical node, which is a physical node for providing VNF services for the target service function sub-chain;

[0120] A total resource consumption amount of the candidate path is less than or equal to a resource remaining amount of the target domain network.

[0121] That is, in the process of in-domain service function sub-chain deployment, VNFs are not allowed to be divided, that is, each VNF is placed on a physical node, and is not allowed to be prevented from being placed on two or more physical nodes, and the constraint can be represented as:

[0122]

[0123] wherein, denotes an i-th VNF in a service function sub-chain , is a binary variable, indicating whether the VNF is placed on a physical node , denotes a set of VNFs of the service function sub-chain .

[0124] In addition, it is also required that each physical node deploys at most one VNF for each SFC, and the constraint can be represented as:

[0125]

[0126] Considering the virtual link mapping of the service function chain, the transmission path needs to pass through a physical node for providing VNF services for the service function sub-chain, that is, a target physical node, and the constraint can be represented as:

[0127]

[0128] wherein, denotes a transmission path of a service function sub-chain , denotes a placement manner of VNFs in the service function sub-chain, and the above constraint indicates that the transmission path needs to pass through a physical node for providing VNF services for the service function sub-chain​ a physical node providing a VNF service.

[0129] Finally, a binary variable is defined to represent whether a virtual link between a physical node and a physical node is mapped, where denotes a virtual link between a physical node and a physical node , each virtual link in a service function chain has the same bandwidth resource requirement, and the service function chain makes the physical node consume a certain amount of computing resource and the physical link

[0130] consume a certain amount of bandwidth resource, respectively, which are represented as:

[0131]

[0132] wherein denotes the computing resource requirement of the th VNF in the service function chain , and denotes the bandwidth resource requirement of the service function chain .

[0133] denotes the remaining computing resource on the physical node , and denotes the remaining bandwidth resource on the link , assuming that the computing resource requirement of a VNF is constant within a service function chain . Therefore, for a deployed VNF physical node and physical link, the capacity limit of computing resource and bandwidth resource can be represented as:

[0134]

[0135]

[0136] When deploying a service function chain in a single-domain network, under the constraint of the above first constraint condition, the process of generating multiple candidate paths corresponding to the target service function chain by the in-domain controller can be implemented as:

[0137] obtaining a source node and a destination node of the target service function chain; the source node is a routing node connected with a previous domain network by the target service function chain; the destination node is a routing node connected with a next domain network by the target service function chain;

[0138] Based on the source node and the destination node of the target service function sub-chain, the KSP algorithm is used to generate the shortest path under the constraint of the first constraint condition, and a plurality of candidate paths corresponding to the target service function sub-chain are obtained.

[0139] In the context of deploying a service function sub-chain in a single-domain network, the source node is a routing node connected to the target service function sub-chain and the previous domain network. It is the entry point of the network traffic entering the target service function sub-chain processing flow in the current target domain network from other domain networks. This node plays a role in the entire multi-domain network service function chain system, which is to receive data from other domain networks and guide it into the specific service function sub-chain of the current domain network. Corresponding to the source node, the destination node is a routing node connected to the target service function sub-chain and the next domain network. It marks the exit location of the network traffic after being processed by the target service function sub-chain in the current target domain network, which is about to leave the current domain network and enter the next domain network. It is a transfer point for data transmission after completing the service function sub-chain processing flow in the current domain network. By clearly defining these two key nodes, the starting and ending key positions for generating reasonable candidate paths in a single-domain network are determined, so that path generation can be planned around the traffic entrance and exit of the service function sub-chain.

[0140] Since the deployment of the service function sub-chain is an NP-hard problem, this problem has many constraints and high computational complexity. Therefore, when generating a plurality of candidate paths of the service function sub-chain, the in-domain controller can calculate K paths satisfying the first constraint condition as a plurality of candidate paths of the service function sub-chain based on the KSP (K-Shortest Paths, K shortest paths) algorithm, K being a positive integer. The KSP algorithm is an algorithm for finding a plurality of shortest paths (the first K shortest paths) from a source node to a destination node in a network. It is based on the shortest path algorithm idea in graph theory and can calculate a plurality of different shortest paths by considering network topology, link weight (such as link bandwidth, delay, etc. can be used as weight), etc. In this scenario, these paths are candidate paths corresponding to the service function sub-chain. By finding a plurality of candidate paths, more selection space can be provided for further screening and optimization.

[0141] In the process of executing the KSP algorithm, the intra-domain controller can regard each physical node in the network topology as a vertex in a graph, and the connection link between the physical nodes as an edge in the graph, and assign a corresponding weight to the edge (the weight can comprehensively consider the bandwidth, delay, and other factors of the link, while satisfying the limit on the remaining amount of resources of the physical node in the first constraint condition). Through the calculation process of the algorithm, starting from the source node, according to the established rules and weight calculation manner, multiple paths to the destination node are gradually explored, and these paths are the multiple candidate paths corresponding to the service function sub-chain that satisfy the first constraint condition.

[0142] In a possible implementation, when the KSP algorithm is applied, the embodiment of the present application can be provided with two different storage lists, one of which is used to place the shortest path fitted based on the KSP algorithm and the corresponding offset path, and the first storage list is "shortest_paths", and the other is used to place the shortest path that satisfies the first constraint condition determined after screening by the first constraint condition, and the second storage list is "ksp_paths". Figure 5 A flowchart of the process of optimal path selection based on the KSP algorithm provided by an example embodiment of the present application is shown in FIG. 5. Figure 5 As shown in FIG. 5, the process of path selection based on the KSP algorithm by the intra-domain controller can include the following steps.

[0143] S501, add the first shortest path to the first storage list "shortest_paths";

[0144] S502, calculate the offset path of the last path in "shortest_paths";

[0145] S503, add the shortest offset path to "shortest_paths";

[0146] S504, determine whether the shortest offset path satisfies the first constraint condition, if yes, execute S505, otherwise, return to S502;

[0147] S505, add the shortest offset path to the second storage list "ksp_paths";

[0148] S506, determine whether the number of paths in "ksp_paths" is less than K, if yes, return to S502, otherwise, execute S507;

[0149] S507, determine and output K paths in "ksp_paths" as the multiple candidate paths of the service function sub-chain, and end the path selection process.

[0150] Step 508: With minimizing intra-domain communication latency and the standard deviation of resource utilization as optimization objectives, the optimal VNF instantiation scheme corresponding to each candidate path is generated by the particle swarm optimization algorithm; wherein, the VNF instantiation scheme is used to indicate the VNF deployment status of each physical node on the candidate path.

[0151] Taking the determination of the optimal VNF instantiation scheme for any candidate path among the candidate paths as an example, the process of generating the optimal VNF instantiation scheme for each candidate path using the particle swarm optimization algorithm can be implemented as follows:

[0152] The particle swarm optimization algorithm is initialized based on the target candidate path; the particle population in the particle swarm optimization algorithm is the set of each VNF instantiation scheme on the corresponding candidate path, and one particle in the particle population represents a VNF instantiation scheme; the target candidate path is any one of the candidate paths;

[0153] Based on the particle velocity update mechanism in the particle swarm optimization algorithm, the particle position of each particle is iteratively adjusted until the optimization objective is met, and the optimal VNF instantiation scheme corresponding to the target candidate path is obtained.

[0154] In particle swarm optimization (PSO) algorithms, the particle population is a collection of multiple particles. In this embodiment, the particle population refers to the collection of VNF instantiation schemes on the corresponding candidate path, where each particle represents a possible VNF instantiation scheme, and the entire particle population covers many different VNF ​​deployment possibilities. The size of the particle population can be set based on factors such as computing resources, desired search accuracy, and algorithm convergence speed. For example, if computing resources are sufficient and a more comprehensive search for the optimal solution is desired, the particle population size can be appropriately increased; conversely, the size needs to be appropriately reduced to ensure algorithm efficiency. This application does not limit the size of the particle population. If the particle population size is set to N, the particle population can be represented as:

[0155]

[0156] in, This represents the number of iterations in the particle swarm. Indicates the transmission path of the service function subchain. This indicates the particle population corresponding to the instantiation scheme under this path. Represents the first in the particle swarm Each particle represents a VNF instantiation scheme. The specific details are as follows:

[0157]

[0158] in, This indicates that the deployment of the first VNF ​​instantiation scheme has been completed. M represents the number of VNF nodes in the subchain of this service function.

[0159] Since the instantiation scheme itself does not have the concept of position and velocity, in this embodiment, the nodes deployed in the VNF in the instantiation scheme can be assigned position and particle velocity to make them move. By trying different deployment schemes and judging the advantages and disadvantages of different deployment schemes, the optimal VNF instantiation scheme can be obtained.

[0160] definition VNF The speed indicates the movement method of the VNF in the next iteration, that is, the deployment node of the m-th VNF will change, either forward or backward along the transmission path of the service function subchain.

[0161] particle speed It consists of three parts: the particle's original velocity, the particle's adjustment of velocity based on its individual optimal position, and the particle's adjustment of velocity based on its population optimal position. This particle velocity update mechanism can be represented as:

[0162]

[0163] in, Represents particles The velocity value of the m-th VNF. Represents particles The optimal position of the m-th individual in the VNF. This represents the population-optimal position of the m-th VNF in the entire particle swarm. and It is a random number between 0 and 1, in order to expand the search space for particles; It is the inertial weight, which represents the degree of importance that the particle attaches to the original velocity when it iterates; For individual learning factors, The population learning factor represents the degree of importance attached to individual and group experiences.

[0164] Based on the position and velocity of the VNF, the position update formula can be obtained:

[0165]

[0166] During iterative adjustments, certain constraints must be met to ensure that the generated VNF instantiation scheme satisfies the path requirements. The process of iteratively adjusting the particle positions of each particle based on the particle velocity update mechanism in the particle swarm optimization algorithm can be implemented as follows:

[0167] Under the constraint of the second constraint condition, the particle position of each particle is iteratively adjusted based on a particle velocity updating mechanism in the particle swarm optimization algorithm.

[0168] The second constraint condition includes:

[0169] The physical node of the VNF deployment is within the path range of the target candidate path.

[0170] The execution order between each VNF remains unchanged.

[0171] That is, under the constraint of the second condition, it can be ensured that the new position of the VNF in the iteration process does not exceed the path range, and the movement of the VNF does not change the execution order of each VNF.

[0172] The optimization target when iteratively adjusting in the service function sub-chain is to minimize the resource occupation rate standard deviation and minimize the intra-domain communication delay. In a possible implementation manner, a fitness function can be designed based on the above optimization target to quantitatively evaluate the above optimization target. In this case, in each iteration adjustment process, the fitness value of each particle is calculated by the fitness function, and the fitness value is a comprehensive value of the intra-domain communication delay and the resource occupation rate standard deviation corresponding to the particle.

[0173] The VNF instantiation scheme represented by the particle with the minimum fitness value in the iteration adjustment is determined as the optimal VNF instantiation scheme corresponding to the target candidate path.

[0174] Illustratively, the fitness function can be represented as:

[0175]

[0176] wherein, f n represents the fitness function of the nth particle; σ represents the resource occupation rate standard deviation of the entire network, T n represents the normalized intra-domain communication delay of the service function sub-chain through the candidate path Since the optimization target in the service function sub-chain deployment scenario is to minimize the intra-domain communication delay and the resource occupation rate standard deviation, the fitness value is calculated based on these two key indicators (intra-domain communication delay and resource occupation rate standard deviation), therefore, the smaller the fitness value, the better the performance of the VNF instantiation scheme on the two optimization targets.

[0177] The calculation method of the intra-domain communication delay of the service function sub-chain may be represented as:

[0178] ​

[0179] wherein, denotes the link communication delay between physical node i and physical node j, i.e., the link communication delay of link

[0180] Resource occupation standard deviation The calculation method can be represented as:

[0181]

[0182] The resource occupation standard deviation is the sum of the node resource occupation standard deviation and the link resource occupation standard deviation, The smaller the value, the more balanced the deployment of the service function sub-chain in the network, and the better the network state. Wherein, denotes the resource occupation of physical node , denotes the average value of all node resource occupations, denotes the resource occupation of link , denotes the average value of all link resource occupations.

[0183] Then, define denotes the individual optimal position of the particle , define as the fitness value of the individual optimal position of the nth particle; define denotes the group optimal position of the particle swarm, is the fitness value of the population optimal position, i.e., the optimal deployment method on a path; define denotes the optimal deployment scheme of the service function sub-chain, is the fitness value of the optimal deployment scheme of the service function sub-chain.

[0184] Figure 6 The flowchart for generating the optimal VNF scheme of the candidate path is shown in an example embodiment of the present application, as shown in Figure 6 , the process can be implemented as:

[0185] S601, initialize the particle population of the candidate path;

[0186] S602, initialize the individual optimal position and the population optimal position;

[0187] S603, update the particle position and the particle velocity;

[0188] S604, update the individual optimal position;

[0189] S605, update the population optimal position;

[0190] S606, determining whether a termination condition is met; if yes, ending, otherwise, returning to S603.

[0191] wherein, in the case of meeting the termination condition, it is indicated that the current VNF instantiation scheme reaches the optimization target, and the current VNF instantiation scheme is determined as the optimal VNF instantiation scheme corresponding to the candidate path.

[0192] That is, for each particle (i.e., each VNF instantiation scheme) in the current particle swarm, first, according to the network topology of the target candidate path, the resource status of each physical node, and the VNF deployment represented by the current particle, the intra-domain communication delay and the resource occupation rate standard deviation corresponding to the scheme are calculated, and then the fitness value is calculated through the pre-set fitness function, which can comprehensively reflect the performance of the VNF instantiation scheme in the optimization target.

[0193] For each particle, the current fitness value is compared with the fitness value corresponding to its individual optimal position; if the current fitness value is better, the current position is updated to the individual optimal position, and the better VNF instantiation scheme is recorded; then, the fitness values of all particles in the entire particle swarm are compared to find the particle with the best fitness value, and if the fitness value is better than the fitness value corresponding to the current global optimal position, the position of the particle is updated to the global optimal position, which means that the VNF instantiation scheme with the best performance in the optimization target in the entire swarm is found.

[0194] According to the particle speed updating mechanism of the particle swarm optimization algorithm, the speed of each particle is updated, and according to the updated speed, the position of the particle is updated according to the position updating formula.

[0195] The above iteration process is repeated to continuously update the speed, position, individual optimal position and global optimal position of the particle, continuously calculate the fitness value and compare the optimization, until the pre-set optimization target is met, and the optimal VNF instantiation scheme corresponding to the candidate path is obtained.

[0196] The above process is repeated to obtain the optimal VNF instantiation scheme of each candidate path.

[0197] Step 409, based on the optimal VNF instantiation scheme corresponding to each candidate path, the target path and the optimal VNF instantiation scheme of the target path are determined.

[0198] After the path selection is completed, each service function sub-chain has K candidate paths that can be selected for deployment, and the intra-domain controller can select an optimal path and an optimal VNF instantiation scheme from the K candidate paths. In this process, the intra-domain controller can compare the fitness values of the K candidate paths, and determine the candidate path with the minimum fitness value as the target path, thereby obtaining the optimal VNF instantiation scheme of the target path.

[0199] At step 410, the target service function sub-chain is deployed in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path.

[0200] In summary, the service function chain deployment method provided by the embodiments of the present application can obtain the target service function sub-chain of a target domain network from the chain cutting scheme of the service function chain by the intra-domain controller of any target domain network in a multi-domain network, generate a plurality of candidate paths corresponding to the target service function sub-chain under the constraint of the first constraint condition constructed by comprehensively considering the target domain network physical nodes and the resource remaining amount of each physical node, then generate the optimal VNF scheme corresponding to each candidate path by means of the particle swarm optimization algorithm with the minimization of the intra-domain communication delay and the resource occupation rate standard deviation as the optimization target, determine the target path and the optimal VNF scheme corresponding to the target path from the optimal VNF scheme, and deploy the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path. Through the above method, the actual physical node resource status of the target domain network can be fully combined to generate reasonable candidate paths, and deployment schemes that are not feasible due to resource limitations can be avoided. When determining the VNF instantiation scheme, both the intra-domain communication delay and the resource occupation rate standard deviation are taken into account, and the particle swarm optimization algorithm is used to determine a scheme with better performance, so that the network can ensure the timeliness of data transmission and improve the allocation effect of the intra-domain link during operation, achieve balanced allocation of VNF on each physical node, and improve the stability and efficiency of the network as a whole.

[0201] Figure 7 A block diagram of a service function chain deployment apparatus provided by an example embodiment of the present application is shown, which is applied in an intra-domain controller of a target domain network to perform all or part of the steps performed by the intra-domain controller in the embodiments shown in Figure 1 or Figure 4 The target domain network is any domain network in a multi-domain network, as shown in Figure 7 The apparatus includes:

[0202] The link obtaining module 710 is configured to obtain a target service function sub-chain of the target domain network based on a chain cutting scheme of a service function chain; the chain cutting scheme is used to indicate that the service function chain is cut into a plurality of service function sub-chains corresponding to the multi-domain network; the target service function sub-chain includes resource requirements of each virtual network function (VNF) allocated to the target domain network;

[0203] The path generating module 720 is configured to generate a plurality of candidate paths corresponding to the target service function sub-chain under a constraint of a first constraint condition; the first constraint condition is constructed based on physical nodes in the target domain network and resource remaining amounts of each physical node.

[0204] The first scheme generating module 730 is configured to generate an optimal VNF instantiation scheme corresponding to each candidate path by using a particle swarm optimization algorithm, with minimization of in-domain communication delay and resource occupation rate standard deviation as an optimization target; the VNF instantiation scheme is used to indicate VNF deployment of each physical node on the candidate path.

[0205] The determining module 740 is configured to determine a target path and an optimal VNF instantiation scheme of the target path based on the optimal VNF instantiation scheme corresponding to each candidate path.

[0206] The deployment module 750 is configured to deploy the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path.

[0207] In a possible implementation, the first scheme generating module 730 includes:

[0208] The initialization submodule is configured to initialize the particle swarm optimization algorithm based on a target candidate path; a particle population in the particle swarm optimization algorithm is a set of VNF instantiation schemes corresponding to each VNF on the candidate path, and one particle in the particle population represents one VNF instantiation scheme; the target candidate path is any one of each candidate path.

[0209] The iterative adjustment submodule is configured to perform iterative adjustment of particle positions of each particle based on a particle speed updating mechanism in the particle swarm optimization algorithm, until the optimization target is met, and an optimal VNF instantiation scheme corresponding to the target candidate path is obtained.

[0210] In a possible implementation, the iterative adjustment submodule is configured to,

[0211] In each iterative adjustment process, the fitness value of each particle is calculated by using a fitness function, and the fitness value is a comprehensive value of the in-domain communication delay and the resource occupation rate standard deviation corresponding to the particle.

[0212] The VNF instantiation scheme characterized by the particle with the minimum fitness value in the iterative adjustment is determined as the optimal VNF instantiation scheme corresponding to the target candidate path.

[0213] In a possible implementation, the iterative adjustment submodule is configured to:

[0214] Under the constraint of the second constraint condition, perform particle position iterative adjustment on each particle based on a particle speed updating mechanism in the particle swarm optimization algorithm.

[0215] The second constraint condition includes:

[0216] The physical node where the VNF is deployed is within the path range of the target candidate path.

[0217] The execution order between each VNF remains unchanged.

[0218] In a possible implementation, the path generation module 720 includes:

[0219] A node acquisition submodule is configured to acquire a source node and a destination node of the target service function subchain; the source node is a routing node where the target service function subchain is connected to a previous domain network; and the destination node is a routing node where the target service function subchain is connected to a subsequent domain network.

[0220] A path generation submodule is configured to perform shortest path generation based on the source node and the destination node of the target service function subchain under the constraint of the first constraint condition by using a KSP algorithm, to obtain a plurality of candidate paths corresponding to the target service function subchain.

[0221] In a possible implementation, the first constraint condition includes:

[0222] The VNF is indivisible.

[0223] At most one VNF is deployed on each physical node.

[0224] The candidate path passes through a target physical node, which is a physical node for providing a VNF service for the target service function subchain.

[0225] The total resource consumption amount corresponding to the candidate path is less than or equal to the resource remaining amount of the target domain network.

[0226] In summary, the service function chain deployment apparatus provided by the embodiments of the present application, the domain controller of any one target domain network in the multi-domain network obtains a target service function sub-chain of the target domain network from a chain cutting scheme of the service function chain, generates a plurality of candidate paths corresponding to the target service function sub-chain under the constraint of a first constraint condition constructed by comprehensively considering the target domain network physical nodes and the resource remaining amount of each physical node, then takes the minimization of the domain communication time delay and the resource occupation rate standard deviation as the optimization target, generates an optimal VNF scheme corresponding to each candidate path by means of the particle swarm optimization algorithm, and then determines the optimal target path of the target domain network and the optimal VNF scheme corresponding thereto from the optimal VNF scheme, and deploys the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path. Through the above apparatus, reasonable candidate paths can be generated by fully combining the actual physical node resource status of the target domain network, and deployment schemes that are not feasible due to resource limitations can be avoided. When determining the VNF instantiation scheme, both the domain communication time delay and the resource occupation rate standard deviation, which are two key indicators, are taken into account, and the particle swarm optimization algorithm is used to determine a scheme with better performance, so that the network can ensure the timeliness of data transmission and improve the allocation effect of the intra-domain link during operation, and the balanced allocation of VNF on each physical node is achieved, thereby improving the stability and efficiency of the network as a whole.

[0227] Figure 8 A block diagram of a service function chain deployment apparatus provided by another example embodiment of the present application is shown, which is applied in a master controller of a multi-domain network to perform all or part of the steps performed by the master controller in the embodiments shown in Figure 2 or Figure 4 The multi-domain network includes at least two domain networks, as shown in Figure 8 The apparatus includes:

[0228] A second scheme generation module 810 generates a chain cutting scheme of the service function chain; the chain cutting scheme is used to indicate that the service function chain is cut into a plurality of service function sub-chains corresponding to the multi-domain network;

[0229] The scheme sending module 820 is configured to send the chain-cutting scheme to the in-domain controllers of each domain network, so that each in-domain controller obtains a corresponding service function sub-chain, and for the in-domain controller of any one target domain network, under the constraint of the first constraint condition, generates a plurality of candidate paths corresponding to the target service function sub-chain; takes minimizing the in-domain communication delay and the standard deviation of resource occupation rate as an optimization target, generates an optimal VNF instantiation scheme corresponding to each candidate path through a particle swarm optimization algorithm, determines a target path and an optimal VNF instantiation scheme of the target path based on the optimal VNF instantiation scheme corresponding to each candidate path, and determines a target path and an optimal VNF instantiation scheme of the target path based on the optimal VNF instantiation scheme corresponding to each candidate path; and deploys the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path.

[0230] In a possible implementation, the apparatus further includes:

[0231] The request receiving module is configured to receive a service function chain deployment request before the second scheme generating module generates the chain-cutting scheme of the service function chain, wherein the service function chain deployment request includes the number and execution order of VNFs, the source node and the destination node of the service function chain, and the resource requirement amount of each VNF.

[0232] The resource amount obtaining module is configured to obtain the inter-domain communication delay between each domain network and the resource remaining amount of each domain network.

[0233] The link determining module is configured to determine a service function chain from the source node to the destination node of the service function chain based on the inter-domain communication delay between each domain network, the resource remaining amount of each domain network, and the resource requirement amount of each VNF, wherein the total inter-domain communication delay of the service function chain is the shortest.

[0234] In summary, the service function chain deployment apparatus provided by the embodiments of the present application, after the main control network generates the chain cutting scheme of the service function chain, sends the chain cutting scheme to each domain network, the in-domain controller of any target domain network in the multi-domain network acquires the target service function sub-chain of the target domain network from the chain cutting scheme of the service function chain, generates a plurality of candidate paths corresponding to the target service function sub-chain under the constraint of the first constraint condition constructed by comprehensively considering the target domain network physical nodes and the resource residual amount of each physical node, then takes the minimization of the in-domain communication time delay and the resource occupation rate standard deviation as the optimization target, generates the optimal VNF scheme corresponding to each candidate path by means of the particle swarm optimization algorithm, and then determines the optimal target path of the target domain network and the optimal VNF scheme corresponding thereto from the optimal VNF scheme, and deploys the target service function sub-chain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path; by means of the above apparatus, the actual physical node resource condition of the target domain network can be fully combined to generate reasonable candidate paths, and the deployment scheme that is not feasible due to resource limitation can be avoided, when the VNF instantiation scheme is determined, the in-domain communication time delay and the resource occupation rate standard deviation, which are two key indexes, are taken into account, the performance of the scheme is determined by means of the particle swarm optimization algorithm, so that the network can guarantee the timeliness of data transmission and improve the allocation effect of the in-domain link in the running process, the balanced allocation of VNF on each physical node is realized, and the stability and efficiency of the network as a whole are improved.

[0235] Figure 9 A structural block diagram of a computer device 900 shown in an example embodiment of the present application is shown. The computer device can be implemented as the main controller or the in-domain controller in the above-mentioned schemes of the present application. The computer device 900 includes a central processing unit (CPU) 901, a system memory 904 including a random access memory (RAM) 902 and a read-only memory (ROM) 903, and a system bus 905 connecting the system memory 904 and the central processing unit 901. The computer device 900 further includes a mass storage device 906 for storing an operating system 909, application programs 910 and other program modules 911.

[0236] According to various embodiments of the present application, the computer device 900 can also run on a remote computer connected to a network such as the Internet. That is, the computer device 900 can be connected to a network 908 through a network interface unit 907 connected to the system bus 905, or can be connected to other types of networks or remote computer systems (not shown) using the network interface unit 907.

[0237] The memory further includes at least one instruction, at least one program, code set, or instruction set, which are stored in the memory. The central processing unit 901 executes the at least one instruction, at least one program, code set, or instruction set to implement all or part of the steps in the material recommendation method shown in the above embodiments.

[0238] Figure 10 A structural block diagram of a computer device 1000 illustrating another exemplary embodiment of this application is shown. The computer device 1000 can be implemented as the aforementioned master controller or domain controller.

[0239] Typically, computer device 1000 includes a processor 1001 and a memory 1002.

[0240] In some embodiments, the computer device 1000 may optionally include a peripheral device interface 1003 and at least one peripheral device. The processor 1001, memory 1002, and peripheral device interface 1003 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 1003 via a bus, signal line, or circuit board. Specifically, the peripheral device includes at least one of the following: a radio frequency circuit 1004, a display screen 1005, a camera assembly 1006, an audio circuit 1007, and a power supply 1008.

[0241] In some embodiments, the computer device 1000 further includes one or more sensors 1009. The one or more sensors 1009 include, but are not limited to, an accelerometer 1010, a gyroscope 1011, a pressure sensor 1012, an optical sensor 1013, and a proximity sensor 1014.

[0242] Those skilled in the art will understand that Figure 10 The structure shown does not constitute a limitation on the computer device 1000, and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0243] In one exemplary embodiment, a computer-readable storage medium is also provided, which stores at least one computer program that is loaded and executed by a processor to implement all or part of the steps in the deployment method of the service function chain described above. For example, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, or optical data storage device, etc.

[0244] In an exemplary embodiment, there is also provided a computer program product comprising at least one computer program which, when loaded into a processor, causes the processor to perform the above-described Figure 2 or Figure 3 all or part of the steps of the method of deploying a service function chain as shown in any of the embodiments.

[0245] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. It is intended that the specification and examples be considered as exemplary only, with the true scope and spirit of the application being indicated by the following claims.

[0246] It is to be understood that the application is not limited to the precise details of construction and the exemplary embodiments described above and illustrated in the drawings. The scope of the application is indicated by the appended claims rather than by the exemplary embodiments given above.

Claims

1. A method for deploying a service function chain, characterized in that, The method is executed by an intra-domain controller of a target domain network, which is any one of a multi-domain network. The method includes: The chain-slicing scheme based on the service function chain obtains the target service function sub-chain of the target domain network; the chain-slicing scheme is used to indicate that the service function chain is divided into multiple service function sub-chains corresponding to the multi-domain network; the target service function sub-chain contains the resource requirements of each virtual network function (VNF) allocated to the target domain network. Under the constraint of the first constraint, multiple candidate paths corresponding to the target service function subchain are generated; the first constraint is constructed based on the physical nodes in the target domain network and the remaining resources of each physical node. With the optimization objectives of minimizing intra-domain communication latency and the standard deviation of resource utilization, the optimal VNF instantiation scheme corresponding to each candidate path is generated by the particle swarm optimization algorithm; wherein, the VNF instantiation scheme is used to indicate the VNF deployment status of each physical node on the candidate path; Based on the optimal VNF instantiation scheme corresponding to each candidate path, the target path and the optimal VNF instantiation scheme of the target path are determined. Based on the target path and the optimal VNF instantiation scheme for the target path, the target service function subchain is deployed in the target domain network.

2. The method according to claim 1, characterized in that, The optimization objective is to minimize intra-domain communication latency and the standard deviation of resource occupancy. The optimal VNF instantiation scheme for each candidate path is generated using the particle swarm optimization algorithm, including: The particle swarm optimization algorithm is initialized based on the target candidate path; the particle population in the particle swarm optimization algorithm is the set of each VNF instantiation scheme on the corresponding candidate path, and one particle in the particle population represents a VNF instantiation scheme; the target candidate path is any one of the candidate paths; Based on the particle velocity update mechanism in the particle swarm optimization algorithm, the particle position of each particle is iteratively adjusted until the optimization objective is met, thereby obtaining the optimal VNF instantiation scheme corresponding to the target candidate path.

3. The method according to claim 2, characterized in that, The method further includes: In each iteration adjustment process, the fitness value of each particle is calculated by the fitness function. The fitness value is a comprehensive value of the intra-domain communication delay and the standard deviation of resource occupancy rate corresponding to the particle. The VNF instantiation scheme representing the particle with the smallest fitness value during iterative adjustment is determined as the optimal VNF instantiation scheme corresponding to the target candidate path.

4. The method according to claim 2, characterized in that, The iterative adjustment of particle positions based on the particle velocity update mechanism in the particle swarm optimization algorithm includes: Under the constraint of the second constraint, the particle position of each particle is iteratively adjusted based on the particle velocity update mechanism in the particle swarm optimization algorithm. The second constraint includes: The physical nodes deployed in the VNF are within the path range of the target candidate path; The execution order between each VNF remains unchanged.

5. The method according to claim 1, characterized in that, The generation of multiple candidate paths corresponding to the target service function subchain under the first constraint includes: Obtain the source node and destination node of the target service function subchain; the source node is the routing node connecting the target service function subchain to the previous domain network; the destination node is the routing node connecting the target service function subchain to the next domain network. Based on the source node and destination node of the target service function subchain, and under the constraint of the first constraint, the shortest path is generated using the KSP algorithm to obtain multiple candidate paths corresponding to the target service function subchain.

6. The method according to claim 1, characterized in that, The first constraint includes: VNF is indivisible; At most one VNF can be deployed on each physical node; The candidate path passes through the target physical node, which is a physical node used to provide VNF services for the target service function subchain; The total resource consumption corresponding to the candidate path is less than or equal to the remaining resources of the target domain network.

7. A method for deploying a service function chain, characterized in that, The method is executed by the master controller of a multi-domain network, the multi-domain network comprising at least two domain networks, the method comprising: A chain-slicing scheme for generating service function chains; the chain-slicing scheme is used to indicate that the service function chain is divided into multiple service function sub-chains corresponding to the multi-domain network; The chain-cutting scheme is sent to the domain controllers of each domain network, enabling each domain controller to obtain the corresponding service function sub-chain. For any domain controller of a target domain network, under the constraint of the first constraint, multiple candidate paths corresponding to the target service function sub-chain are generated. With minimizing intra-domain communication latency and the standard deviation of resource occupancy as optimization objectives, the optimal VNF instantiation scheme corresponding to each candidate path is generated using the particle swarm optimization algorithm. Based on the optimal VNF instantiation scheme corresponding to each candidate path, the target path and its optimal VNF instantiation scheme are determined. Based on the optimal VNF instantiation scheme corresponding to each candidate path, the target path and its optimal VNF instantiation scheme are determined. Based on the target path and its optimal VNF instantiation scheme, the target service function sub-chain is deployed in the target domain network.

8. The method according to claim 7, characterized in that, Before generating the service function chain slicing scheme, the method further includes: Receive a service function chain deployment request, which includes the number and execution order of VNFs, the source node and destination node of the service function chain, and the resource requirements of each VNF. Obtain the inter-domain communication latency between various domain networks and the remaining resources of each domain network; Based on the inter-domain communication latency between various domain networks, the remaining resources of each domain network, and the resource requirements of each VNF, a service function chain from the source node to the destination node of the service function chain is determined; the total inter-domain communication latency of the service function chain is minimized.

9. A service function chain deployment apparatus, characterized in that, The device is used in an intra-domain controller of a target domain network, which is any one of multiple domain networks. The device includes: The link acquisition module is used to acquire the target service function sub-chain of the target domain network based on the service function chain slicing scheme; the slicing scheme is used to indicate that the service function chain is divided into multiple service function sub-chains corresponding to the multi-domain network; the target service function sub-chain contains the resource requirements of each virtual network function (VNF) allocated to the target domain network. The path generation module is used to generate multiple candidate paths corresponding to the target service function subchain under the constraints of a first constraint condition; the first constraint condition is constructed based on the physical nodes in the target domain network and the remaining resources of each physical node. The first scheme generation module is used to generate the optimal VNF instantiation scheme for each candidate path by using the particle swarm optimization algorithm, with the optimization objective of minimizing intra-domain communication latency and the standard deviation of resource utilization. The VNF instantiation scheme is used to indicate the VNF deployment status of each physical node on the candidate path. The determination module is used to determine the target path and the optimal VNF instantiation scheme of the target path based on the optimal VNF instantiation scheme corresponding to each candidate path. The deployment module is used to deploy the target service function subchain in the target domain network based on the target path and the optimal VNF instantiation scheme of the target path.

10. A deployment apparatus for a service function chain, characterized in that, The device is applied to the execution of a master controller in a multi-domain network, the multi-domain network comprising at least two domain networks, the device comprising: The second scheme generation module is used to generate a chain-splitting scheme for the service function chain; the chain-splitting scheme is used to indicate that the service function chain is divided into multiple service function sub-chains corresponding to the multi-domain network. The scheme sending module is used to send the chain-cutting scheme to the domain controllers of each domain network, so that each domain controller can obtain the corresponding service function sub-chain. For any domain controller of a target domain network, under the constraint of the first constraint, multiple candidate paths corresponding to the target service function sub-chain are generated. With minimizing the intra-domain communication latency and the standard deviation of resource utilization as the optimization objective, the optimal VNF instantiation scheme corresponding to each candidate path is generated by the particle swarm optimization algorithm. Based on the optimal VNF instantiation scheme corresponding to each candidate path, the target path and its optimal VNF instantiation scheme are determined. Based on the optimal VNF instantiation scheme corresponding to each candidate path, the target path and its optimal VNF instantiation scheme are determined. Based on the target path and its optimal VNF instantiation scheme, the target service function sub-chain is deployed in the target domain network.

11. A computer device, characterized in that, The computer device includes a processor and a memory, the memory storing at least one computer program, the at least one computer program being loaded and executed by the processor to implement the service function chain deployment method as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer program, which is loaded and executed by a processor to implement the deployment method of the service function chain as described in any one of claims 1 to 9.

Images