Service function chain mapping method applied to elastic optical network

Abstract

The invention relates to the technical field of elastic optical networks, and discloses a service function chain mapping method applied to an elastic optical network, which comprises the following steps of: calculating a virtual link time delay sensitivity and a virtual link time delay sensitivity, and according to a descending order of the sensitivities, calculating a virtual link time delay sensitivity; sequentially mapping the virtual links in the virtual link mapping set on the basis of the principle that the virtual link with the highest virtual link delay sensitivity in the virtual link mapping set is mapped to the physical link with the highest physical link delay sensitivity in the physical link mapping set; and meanwhile, a network chromatography analysis method is introduced, factors influencing the end-to-end delay of the service function chain are comprehensively considered, and weights of different influence factors are calculated, so that the delay sensitivity of a physical link is optimized, the mapping blocking rate of the service function chain is further reduced, and the mapping cost of the service function chain is reduced.

Description

Technical Field

[0001] This invention relates to the field of flexible optical network technology, and in particular to a service function chain mapping method applied to flexible optical networks. Background Technology

[0002] With the widespread adoption of 5G and other networks, the Industrial Internet of Things (IIoT) is experiencing revolutionary development. The number of 5G devices, ranging from everyday consumer electronics to industrial production, is growing exponentially, making application scenarios more complex. Traditional network architectures rely on dedicated hardware or equipment, which is not only expensive but also vendor-specific and geographically fixed. This means that network service providers will need to invest heavily in equipment and human resources to dynamically expand functionality or add new features to existing hardware. The high cost hinders the expansion of the IIoT, and the rigid network paradigm significantly reduces the flexibility of network systems. Scholars believe that Network Functions Virtualization (NFV) technology is an effective way to eliminate network rigidity.

[0003] Benefiting from network function virtualization (NFV), user-requested network services consist of abstract virtual network functions and virtual links connecting them. In this paradigm, virtual network functions are connected in a specific functional order, linking to the request initiator and receiver to form a service function chain. Data flows sequentially through the entire service function chain to complete the service. Service function chain mapping is a crucial aspect of service function chain research. In a static scenario, for a given set of service function chain requests, the key to improving the profitability of service function chain mapping lies in designing a reasonable mapping scheme that meets the resource requirements of this set of service function chains, reduces the service function chain mapping blocking rate, and increases the number of successfully mapped service function chains, thereby increasing mapping revenue. Simultaneously, by rationally aggregating functions, mapping costs can be reduced, thus improving the profitability of service function chain mapping.

[0004] As network systems evolve towards ultra-dense and diversified architectures, traditional wavelength division multiplexing (WDM) optical networks suffer from low spectrum resource utilization and cannot adapt to dynamic bandwidth changes in service function chain (SFC) requests. Elastic optical networks (OOR) achieve finer-grained spectrum allocation through optical orthogonal frequency division multiplexing (OFDM), dynamically allocating spectrum resources based on different SFC requests. Flexible spectrum management enhances the data transmission capabilities of OOR networks, providing robust support for deploying SFC requests with varying bandwidth requirements, especially high-bandwidth SFC requests. To reduce the virtual network function (VNF) instantiation costs associated with SFC mapping, VNFs of the same type in different SFC requests can be mapped to the same physical node in the physical network, achieving aggregation of VNFs of the same type. However, this introduces new problems and challenges, such as the ranking of candidate mapping physical nodes and physical links. In OOR networks, the added constraints on VNF mapping, VNF instantiation, and end-to-end latency lead to a more complex network environment.

[0005] In summary, existing service function chain mapping methods cannot comprehensively quantify the impact of physical node computing resources, physical network topology, and physical link spectrum resources on the end-to-end latency of the service function chain, making it difficult to accurately control latency. Furthermore, existing solutions struggle to balance mapping costs and benefits; either unoptimized aggregation increases instantiation costs, or ignoring resource constraints increases blocking rates, resulting in low profitability for service function chain mapping and a degraded user service experience. Summary of the Invention

[0006] Therefore, the technical problem to be solved by the present invention is to overcome the problem that the existing technology cannot comprehensively map the influencing factors, making it difficult to balance the mapping cost and benefit, resulting in increased blocking rate and reduced mapping profit.

[0007] To address the aforementioned technical problems, this invention provides a service function chain mapping method for elastic optical networks, comprising: Obtain the request initiator, request receiver, and end-to-end latency requirements of the service function chain to be mapped; the virtual network function sequence and the function type, computing resource requirements, and instantiation resource requirements of each virtual network function; and the virtual link set and the bandwidth and spectrum resource requirements of each virtual link. Based on the bandwidth resource requirements of the virtual link, as well as the computational and instantiation resource requirements of the virtual network functions at both ends of the virtual link, the latency sensitivity of the virtual link is calculated; all virtual links in the virtual link set are sorted in descending order of virtual link latency sensitivity to obtain the virtual link mapping set. Based on the total computing resources, remaining available computing resources, instantiated virtual network function types, remaining available bandwidth resources, remaining available spectrum resources, and physical distance of each physical node in the elastic optical network, a mapping auxiliary graph is constructed. Based on the computational resources of the physical nodes at both ends of the physical link in the mapping auxiliary graph and the spectrum resources and physical distance on their shortest paths, the physical link latency sensitivity is calculated; all physical links in the elastic optical network are sorted in descending order of physical link latency sensitivity to obtain the physical link mapping set; Based on the principle of mapping the virtual links with the highest latency sensitivity in the virtual link mapping set to the physical links with the highest latency sensitivity in the physical link mapping set, the virtual links in the virtual link mapping set are mapped sequentially.

[0008] Preferably, the regulation factor is calculated using the hierarchical analysis of networks, including: The elements are the physical topology, available spectrum resources, time delay, and total computing resources of the physical nodes at both ends of each physical link in the elastic optical network. Each element group is used as a criterion, and based on the scaling method, the element groups are compared pairwise to obtain multiple judgment matrices; For each judgment matrix, the initial weight of each row element in the judgment matrix is ​​calculated using the geometric mean method and then normalized to obtain the normalized weight vector of each row element in the judgment matrix. Based on the eigenvectors and the largest eigenvalue of the judgment matrix, a consistency check is performed on the judgment matrix. If the consistency check passes, the eigenvectors of all judgment matrices are placed in order to form a weighted matrix. Based on the normalized weight vectors between each pair of elements to be compared, a supermatrix is ​​constructed by filling in the corresponding elements according to their positions. Multiply the weighted matrix by the values ​​at corresponding positions in the hypermatrix to obtain the weighted hypermatrix; After performing self-multiplication iteration on the weighted hypermatrix, the limiting hypermatrix is ​​obtained; Obtain the column vectors at preset positions in the limiting hypermatrix, normalize them, and obtain three adjustment factors.

[0009] Preferably, after obtaining the physical link mapping set, the method further includes: Based on the virtual network function types that each physical node in the elastic optical network has already instantiated, as well as the request initiator and request receiver, and prioritizing physical nodes with instantiation experience, candidate physical paths are selected from the physical link mapping set to construct a candidate physical path mapping set. Based on the principle of mapping the virtual links with the highest latency sensitivity in the virtual link mapping set to the physical links with the highest latency sensitivity in the candidate physical path mapping set, the virtual links in the virtual link mapping set are mapped sequentially.

[0010] Preferably, based on the principle of mapping the virtual link with the highest latency sensitivity in the virtual link mapping set to the physical link with the highest latency sensitivity in the physical link mapping set, the virtual links in the virtual link mapping set are mapped sequentially, including: If the virtual link mapping set is the The bandwidth resource requirement of the virtual link is greater than that of the first virtual link in the physical link mapping set. The remaining available bandwidth resources of the physical links then let The next physical link in the physical link mapping set is determined until the remaining available bandwidth resources meet the requirements of the first step. The physical link that meets the bandwidth resource requirements of the virtual link is used as the target physical link; Obtain the target physical link Find the shortest path and initialize it. Using the first hit method to determine the first Do the remaining available spectrum resources on the shortest path satisfy the condition of the first shortest path? Spectrum resource requirements for each virtual link: If satisfied, then use the first... Find the shortest path and use it as the target path; If not satisfied, then let Repeatedly use the first hit method to determine the first hit. Does the remaining available spectrum resource on the shortest path satisfy the condition of the first shortest path? The spectrum resource requirements of each virtual link are determined until the first one is obtained. The shortest path for the spectrum resource requirements of each virtual link is taken as the target path; Based on the target path, for the first The virtual link will allocate spectrum resources to the first one. Each virtual link is mapped to the target physical link; make , for the Map each virtual link until... Complete the mapping of the service function chain to be mapped; This indicates the total number of virtual links in the virtual link set.

[0011] Compared with the prior art, the above-described technical solution of the present invention has the following advantages:

[0012] The service function chain mapping method for elastic optical networks described in this invention calculates the latency sensitivity of a virtual link based on its bandwidth resource requirements and the computational and instantiation resource requirements of the virtual network functions at both ends of the virtual link. It also calculates the latency sensitivity of a physical link based on the computational resources of the physical nodes at both ends of the physical link and the spectrum resources and physical distance along their shortest path. By prioritizing mapping virtual links with higher latency sensitivity to physical links with higher physical latency sensitivity, the method increases the probability of successful service function chain mapping, reduces the mapping blocking rate, and improves the benefits of service function chain mapping. This effectively reduces the end-to-end latency of service function chain mapping and improves mapping efficiency.

[0013] This invention introduces network hierarchy analysis (HMA) to construct a judgment matrix, a hypermatrix, and a weighted matrix. Through consistency checks and iterative calculations, it quantifies the impact weights of physical node computing resources, physical network topology, and physical link spectrum resources on the end-to-end latency of the service function chain, thereby determining the adjustment factor. This balances the impact weights of multiple resources, avoids excessive occupation or idleness of a single resource, improves the utilization rate of physical node computing resources and link spectrum resources, and further reduces the resource consumption cost in the mapping process.

[0014] This invention constructs a candidate physical link mapping set, which prioritizes mapping virtual network functions to physical nodes that have already instantiated the same function type, and then considers physical nodes that have not instantiated the same virtual network function type. This avoids the resource consumption caused by repeatedly instantiating virtual network functions, reduces instantiation costs, and lowers the cost of service function chain mapping.

[0015] During mapping, the present invention pre-configures the shortest working path for all physical node pairs. The sum of the distances of this working path is used as the weight of the corresponding physical link in the physical optical network mapping auxiliary graph. In each mapping, the shortest path that meets the requirements is selected first for mapping, thereby reducing the occupation of physical link spectrum resources and physical node computing resources, avoiding excessive waste of resources, and improving resource utilization. Attached Figure Description

[0016] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart of the steps of the service function chain mapping method for elastic optical networks according to the present invention; Figure 2 This is a flowchart of the service function chain cost-benefit optimization mapping method; Figure 3 It is a diagram of a network hierarchical analysis model used to calculate the weight coefficients of influencing factors; Figure 4 This is a flowchart of the weighting coefficient calculation process; Figure 5 This is a service function chain mapping system structure diagram; Figure 6 This is a diagram of the service function chain; Figure 7 This is a topology diagram of a 6-node, 8-link elastic optical network; Figure 8 It is a physical optical network mapping auxiliary graph; Figure 9 This is a schematic diagram of the service function chain mapping results. Detailed Implementation

[0017] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0018] Reference Figure 1 The flowchart shown illustrates the steps of the service function chain mapping method for elastic optical networks according to the present invention. The specific steps include: S101: Obtain the request initiator, request receiver, and end-to-end latency requirements of the service function chain to be mapped; the virtual network function sequence and the function type, computing resource requirements, and instantiation resource requirements of each virtual network function; and the virtual link set and the bandwidth and spectrum resource requirements of each virtual link. S102: Based on the bandwidth resource requirements of the virtual link, as well as the computational resource requirements and instantiation resource requirements of the virtual network functions at both ends of the virtual link, calculate the latency sensitivity of the virtual link; sort all virtual links in the virtual link set in descending order of virtual link latency sensitivity to obtain the virtual link mapping set; Virtual link latency sensitivity is expressed as: ; in, Indicates virtual network function and Virtual link latency sensitivity of virtual links between them Indicates the first A set of virtual links in a service function chain; and These represent the virtual network functions at both ends of the virtual link. and The computing resource requirements; and These represent the virtual network functions at both ends of the virtual link. and The instantiation resource requirements; Indicates the first Virtual links on each service function chain bandwidth resource requirements; S103: Construct a mapping auxiliary graph based on the total computing resources, remaining available computing resources, instantiated virtual network function types, remaining available bandwidth resources, remaining available spectrum resources, and physical distance of each physical node in the elastic optical network. S104: Calculate the physical link latency sensitivity based on the computational resources of the physical nodes at both ends of the physical link in the mapping auxiliary graph and the spectrum resources and physical distance on their shortest paths; sort all physical links in the elastic optical network in descending order of physical link latency sensitivity to obtain the physical link mapping set; S105: Based on the principle of mapping the virtual link with the highest latency sensitivity in the virtual link mapping set to the physical link with the highest latency sensitivity in the physical link mapping set, the virtual links in the virtual link mapping set are mapped sequentially.

[0019] The service function chain mapping method for elastic optical networks described in this invention calculates the latency sensitivity of a virtual link based on its bandwidth resource requirements and the computational and instantiation resource requirements of the virtual network functions at both ends of the virtual link. It also calculates the latency sensitivity of a physical link based on the computational resources of the physical nodes at both ends of the physical link and the spectrum resources and physical distance along their shortest path. By prioritizing mapping virtual links with higher latency sensitivity to physical links with higher physical latency sensitivity, the method increases the probability of successful service function chain mapping, reduces the mapping blocking rate, and improves the benefits of service function chain mapping. This effectively reduces the end-to-end latency of service function chain mapping and improves mapping efficiency.

[0020] Specifically, in step S103, the physical link latency sensitivity is calculated, including: S103-1: Based on physical nodes and Calculate the impact of spectrum resources on the shortest path between them. , is represented as: ; S103-2: Based on physical nodes and Connectivity and physical nodes and Calculate the physical distance of the shortest path between them, and calculate the physical topology impact. , is represented as: ; S103-3: Based on physical nodes and The remaining available computing resources, and the impact of computing resources. , is represented as: ; S103-4: Based on adjustment factors , and The impact of spectrum resources, physical topology, and computing resources is weighted and summed to obtain the physical node. and Physical link latency sensitivity between physical links , is represented as: ; in, Indicates from physical node To physical node The number of spectral slots that are idle on the shortest path between them. and These represent the maximum and minimum values ​​of the number of idle spectrum slots on the shortest path between all pairs of physical nodes, respectively. For smoothing parameters; Represents physical nodes To physical node The physical distance of the shortest path between them. and These represent the maximum and minimum physical distances of the shortest paths between all pairs of physical nodes, respectively. and Representing physical nodes Connectivity and physical nodes connectivity, , physical nodes connectivity , Indicates direct connection between physical nodes in a physical network A set of connected physical nodes. Represents the set of all physical nodes in a physical network. Indicates direct connection to physical nodes Connected physical nodes The degree of connectivity; As a binary variable, when compared with a physical node and When directly connected, The value is 1 if it is not 1, otherwise it is 0. and These represent the maximum and minimum connectivity values ​​among all physical nodes in the physical network, respectively. and Representing physical nodes With physical nodes The remaining available computing resources; and These represent the maximum and minimum values ​​of the remaining available computing resources for all physical nodes in the physical network, respectively. and Representing physical nodes With physical nodes Total computing resources; and These represent the maximum and minimum values ​​of the total computing resources of all physical nodes in the physical network, respectively; adjustment factor. , and ,satisfy .

[0021] In this embodiment of the invention, the adjustment factor used to calculate the physical link delay sensitivity is obtained by calculating using the network analytic hierarchy process (AHP), specifically including: The elements are the physical topology, available spectrum resources, time delay, and total computing resources of the physical nodes at both ends of each physical link in the elastic optical network. Each element group is used as a criterion, and based on the scaling method, the element groups are compared pairwise to obtain multiple judgment matrices; For each judgment matrix, the initial weight of each row element in the judgment matrix is ​​calculated using the geometric mean method and then normalized to obtain the normalized weight vector of each row element in the judgment matrix. Based on the eigenvectors and the largest eigenvalue of the judgment matrix, a consistency check is performed on the judgment matrix. If the consistency check passes, the eigenvectors of all judgment matrices are placed in order to form a weighted matrix. Based on the normalized weight vectors between each pair of elements to be compared, a supermatrix is ​​constructed by filling in the corresponding elements according to their positions. Multiply the weighted matrix by the values ​​at corresponding positions in the hypermatrix to obtain the weighted hypermatrix; After performing self-multiplication iteration on the weighted hypermatrix, the limiting hypermatrix is ​​obtained; Obtain the column vectors at preset positions in the limiting hypermatrix, normalize them, and obtain three adjustment factors.

[0022] The consistency check of the judgment matrix includes: Based on the judgment matrix The value of each element in Calculate the first judgment matrix. Initial weight of the row , is represented as: , , Indicates the order of the judgment matrix; For the The initial weights of the rows are normalized to obtain the first row. normalized weight vector of rows , is represented as: ; Based on the judgment matrix and the first The normalized weight vector of the rows is used to obtain the largest eigenvalue of the judgment matrix. , is represented as: ; Based on the maximum eigenvalue of the judgment matrix With order Calculate the consistency index of the judgment matrix. , is represented as: ; Based on the consistency indices of judgment matrices of different orders, a random consistency index of the judgment matrix is ​​calculated. , is represented as: ; The ratio of the consistency index to the random consistency index is calculated and used as a test factor. , is represented as: ; The verification factor is compared with a preset threshold. If the verification factor is less than the preset threshold, the matrix is ​​judged to pass the consistency check.

[0023] This invention introduces network hierarchy analysis (HMA) to construct a judgment matrix, a hypermatrix, and a weighted matrix. Through consistency checks and iterative calculations, it quantifies the impact weights of physical node computing resources, physical network topology, and physical link spectrum resources on the end-to-end latency of the service function chain, thereby determining the adjustment factor. This balances the impact weights of multiple resources, avoids excessive occupation or idleness of a single resource, improves the utilization rate of physical node computing resources and link spectrum resources, and further reduces the resource consumption cost in the mapping process.

[0024] Based on the above embodiments, in this embodiment, after obtaining the physical link mapping set, the method further includes: selecting candidate physical paths from the physical link mapping set based on the virtual network function types that each physical node in the elastic optical network has already instantiated, as well as the request initiator and request receiver, prioritizing physical nodes with instantiation experience, and constructing a candidate physical path mapping set; and mapping the virtual links in the virtual link mapping set sequentially based on the principle of mapping the virtual link with the highest virtual link latency sensitivity in the virtual link mapping set to the physical link with the highest physical link latency sensitivity in the candidate physical path mapping set.

[0025] This embodiment addresses different situations by providing different methods for constructing candidate physical path mapping sets, specifically including: ① If the source node of the virtual link is the request initiator: In the physical link mapping set, select all physical links with the request initiator as the source node to form the first link candidate set; sort the physical links in the first link candidate set whose destination node is the virtual network function type corresponding to the instantiated virtual link destination node in descending order according to the physical link latency sensitivity to form the candidate physical path mapping set; sort the remaining physical paths in the first link candidate set in descending order according to the physical link latency sensitivity and connect them sequentially to the candidate physical path mapping set. ② If the destination node of the virtual link is the request receiver: Select all physical links in the physical link mapping set that have the request receiver as the destination node to form the first link candidate set; sort the physical links in the first link candidate set whose source node is the virtual network function type corresponding to the virtual link source node that has been instantiated, in descending order of physical link latency sensitivity to form the candidate physical path mapping set; sort the remaining physical paths in the first link candidate set in descending order of physical link latency sensitivity and connect them sequentially to the candidate physical path mapping set. ③ If the virtual network function type corresponding to the source node of the virtual link has been instantiated on the physical node: obtain all physical nodes that have instantiated the virtual network function type corresponding to the source node of the virtual link, and form a first node set; in the physical link mapping set, obtain all physical links whose source node is a physical node in the first node set, sort them in descending order of physical link latency sensitivity, and form a candidate physical path mapping set; sort the remaining physical paths in the physical link mapping set in descending order of physical link latency sensitivity, and connect them sequentially to the candidate physical path mapping set; ④ If the virtual network function type corresponding to the destination node of the virtual link has been instantiated on the physical node: obtain all physical nodes that have instantiated the virtual network function type corresponding to the destination node of the virtual link and form a first node set; in the physical link mapping set, obtain all physical links whose destination nodes are physical nodes in the first node set, sort them in descending order of physical link latency sensitivity, and form a candidate physical path mapping set; sort the remaining physical paths in the physical link mapping set in descending order of physical link latency sensitivity and connect them sequentially to the candidate physical path mapping set; ⑤ If the virtual network function types corresponding to the source and destination nodes of the virtual link have been instantiated on the physical nodes, obtain all physical nodes that have instantiated the virtual network function types corresponding to the source and destination nodes of the virtual link, and form a first node set; in the physical link mapping set, obtain all physical links whose source and destination nodes are physical nodes in the first node set, sort them in descending order of physical link latency sensitivity, and form a candidate physical path mapping set; sort the remaining physical paths in the physical link mapping set in descending order of physical link latency sensitivity, and connect them sequentially to the candidate physical path mapping set; ⑥ If the virtual network function types corresponding to the source node and the destination node of the virtual link have not been instantiated on the physical node, the physical link mapping set is used as the candidate physical path mapping set.

[0026] This invention constructs a candidate physical link mapping set, which prioritizes mapping virtual network functions to physical nodes that have already instantiated the same function type, and then considers physical nodes that have not instantiated the same virtual network function type. This avoids the resource consumption caused by repeatedly instantiating virtual network functions, reduces instantiation costs, and lowers the cost of service function chain mapping.

[0027] In step S105 of this embodiment of the invention, based on the principle of mapping the virtual link with the highest virtual link latency sensitivity in the virtual link mapping set to the physical link with the highest physical link latency sensitivity in the physical link mapping set, the virtual links in the virtual link mapping set are mapped sequentially, including: If the virtual link mapping set is the The bandwidth resource requirement of the virtual link is greater than that of the first virtual link in the physical link mapping set. The remaining available bandwidth resources of the physical links then let The next physical link in the physical link mapping set is determined until the remaining available bandwidth resources meet the requirements of the first step. The physical link that meets the bandwidth resource requirements of the virtual link is used as the target physical link; Obtain the target physical link Find the shortest path and initialize it. Using the first hit method to determine the first Do the remaining available spectrum resources on the shortest path satisfy the condition of the first shortest path? Spectrum resource requirements for each virtual link: If satisfied, then use the first... Find the shortest path and use it as the target path; If not satisfied, then let Repeatedly use the first hit method to determine the first hit. Does the remaining available spectrum resource on the shortest path satisfy the condition of the first shortest path? The spectrum resource requirements of each virtual link are determined until the first one is obtained. The shortest path for the spectrum resource requirements of each virtual link is taken as the target path; Based on the target path, for the first The virtual link will allocate spectrum resources to the first one. Each virtual link is mapped to the target physical link; make , for the Map each virtual link until... Complete the mapping of the service function chain to be mapped; This indicates the total number of virtual links in the virtual link set.

[0028] This invention pre-configures the shortest working path for all physical node pairs during mapping. The sum of the distances of these working paths serves as the weight of the corresponding physical link in the physical optical network mapping auxiliary graph. During each mapping, the shortest path that meets the requirements is prioritized for mapping, reducing the occupation of physical link spectrum resources and physical node computing resources, avoiding excessive resource waste, and improving resource utilization. This invention finds available spectrum resources that satisfy spectrum continuity, spectrum consistency, and end-to-end latency constraints on the selected shortest working path, and maps this virtual link to the selected working path, realizing a service function chain cost-benefit optimization mapping method based on link latency sensitivity.

[0029] In this embodiment, based on the target path, for the first After allocating spectrum resources to each virtual link, the following is also included: Get the The propagation delay generated by mapping a virtual link to the target physical link; Obtain the processing latency incurred from mapping virtual network functions to physical nodes; Get the End-to-end latency of a virtual link; If the sum of propagation delay, processing delay, and end-to-end delay meets the end-to-end delay requirement, then the [missing information] will be [missing information]. A virtual link is mapped to the target physical link.

[0030] To improve the success rate and profitability of service function chain (SFC) mapping, a cost-benefit optimized SFC mapping method can be used. During SFC mapping, end-to-end latency constraints of SFC requests need to be considered. Considering the end-to-end latency and the number of successfully mapped SFCs, this invention combines network hierarchy analysis (NHP) with traditional mapping methods. It comprehensively considers the impact of computing resources of nodes on the physical network, physical network topology, and spectrum resources on physical links on the end-to-end latency of the SFC. A hypermatrix and a weighted matrix are designed, resulting in a weighted hypermatrix. This weighted hypermatrix is ​​then iteratively multiplied to obtain the limiting hypermatrix, thus yielding the weight coefficients of each factor affecting end-to-end latency. Based on the characteristics of SFCs, this invention introduces the concept of link latency sensitivity and proposes a cost-benefit optimized SFC mapping method based on NHP. This method reduces the end-to-end latency of SFCs, thereby ensuring the end-to-end latency requirements while reducing the SFC mapping blocking rate and increasing SFC mapping profitability. Simultaneously, by effectively aggregating virtual network functions of the same type, mapping costs are reduced, thereby maximizing the profit generated by SFC mapping.

[0031] This invention primarily addresses the cost-benefit optimization problem in service function chain mapping, considering the computational resources of physical nodes on the physical network, the physical network topology, and the spectrum resources on physical links. First, a mapping auxiliary graph (AG) is constructed based on the shortest path algorithm and the latency sensitivity of physical links to simplify the virtual link mapping of service function chains. Second, virtual links are sorted in descending order of latency sensitivity to obtain a priority set for virtual link mapping. Finally, when the spectrum resources required by a virtual link on the service function chain are less than or equal to the available spectrum resources of the physical links on the physical network, and the computational and instantiation resources required by the virtual network functions at both ends of the virtual link are less than or equal to the available computational resources of the physical nodes at both ends of the physical link, the virtual link is mapped to that physical link. Based on the principle of mapping the virtual link with the highest latency sensitivity in the service function chain to the physical link with the highest latency sensitivity in the AG, all virtual links are mapped to physical links.

[0032] Based on the above embodiments, in this embodiment of the invention, the service function chain mapping method for elastic optical networks provided by the present invention is used for mapping, referring to... Figure 2 The diagram shown is a flowchart of the service function chain cost-benefit optimization mapping method, which specifically includes: S201: Based on the network analytic hierarchy process, design the supermatrix and weighted matrix. After consistency verification, multiply the corresponding matrices to obtain the weighted supermatrix. After iterating the weighted supermatrix by multiplication, obtain the limit supermatrix, and then obtain the weight coefficients of influencing factors such as the computing resources of physical nodes, physical network topology, and spectrum resources on physical links. In constructing the weighted hypermatrix, this embodiment establishes a network hierarchy analysis method for the service function chain mapping problem based on the relationships between various influencing factors; referring to... Figure 3 The diagram shown is a network hierarchical analysis model used to calculate the weight coefficients of influencing factors; in Figure 3 In the diagram, arrows indicate the relationships between these elements. (See reference...) Figure 4 The diagram shown is a flowchart for calculating the weighting coefficients; the first step is based on... Figure 3 The relationships between all individual elements in each element group are determined by comparing the importance of each pair of elements to obtain a comparison matrix and calculating eigenvectors. Arranging all eigenvectors in order yields the hypermatrix. Next, the criterion layer is entered, where importance comparisons are performed pairwise on an element group basis to obtain a comparison matrix and calculate eigenvectors. Arranging these eigenvectors in order forms the weighted matrix. Multiplying the weighted matrix by the corresponding regions of the hypermatrix yields the weighted hypermatrix. The weighted hypermatrix is ​​then multiplied and iterated to obtain the limiting hypermatrix. At this point, the values ​​in each row of the limiting hypermatrix have converged; normalization is then performed to obtain the weight coefficients of each influencing factor.

[0033] S202: Input a set of service function chains ; Indicates the first A service function chain, in which the virtual network function set Virtual Link Set End-to-end latency requirement set End-to-end delay is initialized to 0. Calculate the Time Dealy Sensitivity of Virtual Link (VTDS) and sort them in descending order to obtain the virtual link mapping set. ; Indicates the first A virtual link, The initial value is 1; The virtual link latency sensitivity is expressed as: ; and Represents the virtual network functions at both ends of the virtual link. and The required computing resources; and Represents the virtual network functions at both ends of the virtual link. and The required instantiation resources; Indicates the first Virtual links on each service function chain The required bandwidth resources; If one end of the virtual link is connected to either the request initiator or the request receiver, then the numerator of the first term only calculates the required computational resources and instantiation resources for a single virtual network function.

[0034] S203: Compare the serial numbers of virtual network functions The total number of virtual links in the virtual link set : like Then from the virtual link mapping set Extracting virtual links Proceed to step S204; Otherwise, mark the first Service Function Chain Mapping successful; update physical network resources. S204: Construct an auxiliary graph (AG) for service function chain mapping based on the shortest path algorithm and physical link latency sensitivity. S204-1: Based on the Time Delay Sensitivity of Physical Link (PTDS) of each physical link, sort the physical link mapping sets in descending order to obtain the physical link mapping set. ; This represents the total number of physical nodes; The physical link latency sensitivity is expressed as: ; The impact of spectrum resources on PTDS along the shortest path from physical node s to physical node d on the physical network is expressed as: ; This represents the number of idle spectrum slots on the shortest path from physical node s to physical node d on the physical network. and These represent the maximum and minimum values ​​of the number of idle spectrum slots on the shortest path between all pairs of physical nodes, respectively. The value is 0.001, used to prevent the denominator from being 0; The effect of physical network topology on PTDS is expressed as: ; This represents the physical distance of the shortest path from physical node s to physical node d on a physical network. and These represent the maximum and minimum physical distances of the shortest paths between all pairs of physical nodes, respectively. and Representing physical paths Zhongyuan physical node Connectivity and host physical node connectivity, Data center nodes connectivity , Indicates the physical network directly connected to the data center node. A collection of connected data center nodes. This represents the set of data center nodes of all physical networks in a resilient optical network oriented towards a data center. Indicates direct connection to data center nodes Connected data center nodes The degree of connectivity; As a binary variable, when interacting with data center nodes With data center nodes When directly connected, The value is 1 if it is not 1, otherwise it is 0. and These represent the maximum and minimum connectivity values ​​among all physical nodes in the physical network, respectively. The expression representing the impact of the computing resources of the physical nodes at both ends of the physical path on PTDS is: ; and Representing physical paths Zhongyuan physical node With the host physical node The remaining available computing resources; and These represent the maximum and minimum values ​​of the remaining available computing resources for all physical nodes in the physical network, respectively. and Representing physical paths Zhongyuan physical node With the host physical node Total computing resources; and These represent the maximum and minimum values ​​of the total computing resources of all physical nodes in the physical network, respectively. , and It is a regulating factor, satisfying ; S204-2: Query the functional types that have been instantiated for each physical node in turn, and construct the set of instantiated functional types for each node. , , jointly build A collection of already instanced functional types; next, starting from the virtual link mapping set... Extract the largest virtual link from the VRUML and construct a candidate physical path mapping set. ; The extracted virtual link may fall into one of the following six categories: ① Virtual Link The source node is the request initiator: In this scenario, based on the service function type of the destination node's virtual network function, all physical nodes that have instantiated that function are queried, and the IDs of these nodes are added to the CPN set. Extract all physical paths that use the physical node number of the request initiator as the source node and the physical node number in the CPN as the destination node, and sort them in descending order according to the PTDS value of each physical path to obtain the physical path mapping set. Next, from Extract all physical paths that use the physical node number of the request initiator as the source node and the physical node number of the physical node that has not instantiated this function as the destination node. Sort each physical path in descending order of its physical link latency sensitivity to obtain the physical path mapping set. .Will Sequential access Finally, the physical paths from the two sets are sequentially fed into the candidate physical path mapping set. .

[0035] ② Virtual Link The destination node is the request receiver: In this scenario, based on the service function type of the source node's virtual network function, all physical nodes that have instantiated that function are queried, and the IDs of these nodes are added to the set CSN. Extract all physical paths that use the physical node number in the CSN as the source node and the physical node number of the requesting receiver as the destination node, and sort them in descending order according to the PTDS value of each physical path to obtain the physical path mapping set. Next, from Extract all physical paths that use the physical node number of the physical node where the function has not been instantiated as the source node and the physical node number of the request receiving end as the destination node. Sort the physical paths in descending order according to their PTDS values ​​to obtain the physical path mapping set. .Will Sequential access Finally, the physical paths from the two sets are sequentially fed into the candidate physical path mapping set. .

[0036] ③ Virtual Link The service function type of the source node virtual network function is instantiated on the physical node in the physical network: In this scenario, based on the service function type of the source node's virtual network function, all physical nodes that have instantiated that function are queried, and the IDs of these nodes are added to the set CSN. Extract all physical paths with physical node numbers in the CSN as source nodes, and sort them in descending order according to their PTDS values ​​to obtain the physical path mapping set. . The remaining physical paths are sorted in descending order of their PTDS values ​​to obtain the physical path mapping set. .Will Sequential access Finally, the physical paths from the two sets are sequentially fed into the candidate physical path mapping set. .

[0037] ④ Virtual Link The service function types of the virtual network functionality of the destination node are instantiated on the physical nodes in the physical network: In this scenario, based on the service function type of the destination node's virtual network function, all physical nodes that have instantiated that function are queried, and the IDs of these nodes are added to the set CSN. Extract all physical paths whose destination nodes are physical node numbers in the CSN, and sort them in descending order according to their PTDS values ​​to obtain the physical path mapping set. . The remaining physical paths are sorted in descending order of their PTDS values ​​to obtain the physical path mapping set. .Will Sequential access Finally, the physical paths from the two sets are sequentially fed into the candidate physical path mapping set. .

[0038] ⑤ Virtual Link The function types of the virtual network functions of both the source and destination nodes have been instantiated on the physical nodes in the physical network: In this scenario, based on the service function type of the source node's virtual network function, query all physical nodes that have instantiated that function, and add their node IDs to the set CSN. Similarly, based on the service function type of the destination node's virtual network function, query all physical nodes that have instantiated that function, and add their node IDs to the set CSN. From... Extract all physical paths that use physical node numbers in the CSN as source nodes and destination nodes, and sort them in descending order according to their PTDS values ​​to obtain the physical path mapping set. . The remaining physical paths are sorted in descending order of their PTDS values ​​to obtain the physical path mapping set. .Will Sequential access Finally, the physical paths from the two sets are sequentially fed into the candidate physical path mapping set. .

[0039] ⑥ Virtual Link The virtual network functionality types of both the source and destination nodes have not been instantiated on the physical nodes in the physical network. directly The physical paths in the data are sequentially fed into the candidate physical path mapping set. .

[0040] Based on the above six scenarios, construct a set of candidate mapping physical paths. ,in, Indicates the first A physical path, and The initial value is 1.

[0041] S205: Compare the available bandwidth resources of the physical link with the bandwidth resource requests of the virtual link; If the bandwidth resource request of the virtual link is greater than the available bandwidth resource of the physical link, then and check Is it empty? like If empty, then mark the first Service Function Chain block; like If it is not empty, the virtual link mapping falls into the following three categories: The virtual network functions of the source and destination nodes on the virtual link are not mapped: In this case, the virtual network functions with high computing resource requirements are mapped to physical nodes with more available computing resources.

[0042] If only one virtual network function is not mapped at either end of the virtual link: In this case, start the search from the physical node mapped by the virtual network function, find an unmapped physical node, and ensure that the PTDS value of the physical path between the two nodes is the highest.

[0043] The virtual network functions of the source and destination nodes on the virtual link have been mapped. In this case, it is determined whether the spectrum resources provided by the physical link can meet the bandwidth resource requirements of the virtual link. If they can, the physical link is selected as the physical link for mapping the virtual link on the auxiliary graph; if they cannot, it is marked. block.

[0044] S206: Calculate physical links using Dijkstra's method There are K shortest paths, where k represents the k-th path, and the initial value of k is 1; S207: Compare the current path number k with the total number of shortest paths K: If k > K, then let Proceed to step S206; Conversely, proceed to step S209; S208: Check the spectrum resource occupancy of each link on the k-th path. Use the first hit method. Once an available spectrum resource segment is found, allocate the spectrum resource segment to the virtual link. If no available spectrum resource segment is found, let k+1 and go to step S206; otherwise, go to step S209. S209: Calculate the propagation delay generated by mapping the current virtual link to the physical link, and the processing delay generated by mapping the virtual network function to the physical node. Add the processing delay and the propagation delay to the end-to-end delay; if the end-to-end delay is less than or equal to... Proceed to step S210; otherwise, mark. block; S210: Will Mapped to From the virtual link mapping set Delete ,examine Is it empty? If it is empty, mark it. Mapping successful, proceed to step S201; otherwise, let... Proceed to step S202.

[0045] This invention combines network hierarchy analysis (HMA) with traditional mapping methods. During the service function chain (SFC) mapping process, it considers both the bandwidth resource requirements of virtual links, the instantiation resources of virtual network functions, and their computational resource requirements, as well as the end-to-end latency constraints of the SFC. First, it considers the physical link propagation latency and the processing latency of physical nodes that affect end-to-end latency. Second, to balance the factors influencing end-to-end latency, it introduces α, β, and γ weight coefficients, calculated using HMA. This effectively increases the number of successfully mapped SFCs while meeting end-to-end latency requirements, thereby improving the mapping revenue of SFCs. Third, it divides the candidate mapping physical node set for virtual network functions into two parts, prioritizing physical nodes that have already instantiated the virtual network function type, thus effectively reducing the mapping cost of SFCs. This invention exhibits significantly superior performance in SFC mapping, effectively improving the mapping success rate and mapping profit.

[0046] Based on the above embodiments, in this embodiment of the invention, a system based on the above-described service function chain mapping method applied to elastic optical networks is provided, referring to... Figure 5 The diagram shown is a system architecture diagram of the service function chain mapping system. Specific modules include: ① Module for calculating the weight coefficients of influencing factors using network tomography: Based on the network analytic hierarchy process, a supermatrix and a weighted matrix are designed. After passing the consistency test, the corresponding supermatrix is ​​multiplied to obtain the weighted supermatrix. The weighted supermatrix is ​​then iterated by self-multiplication to obtain the limiting supermatrix, which in turn yields the weight coefficients of influencing factors such as the computing resources of physical nodes, the physical network topology, and the spectrum resources on physical links. ② Elastic Optical Network Initialization Module: In elastic optical networks In the process, the topology information of the elastic optical network is read, and the optical network connection status, the number of node computing resources, the virtual network function types that the nodes have been instantiated, the number of network switching nodes, the number of fiber links, the number of spectrum slots for each fiber link, and the bandwidth of each spectrum slot are initialized. ③ Service function chain generation module: Generate a set of service function chains Configure information such as the number of service function chains, the number of virtual network functions, the type of virtual network function, the number of virtual links, the service function chain topology, the computing and instantiation resources required for virtual network functions, and the bandwidth resources required for virtual links; ④ Virtual link latency sensitivity calculation module: Calculate the latency sensitivity of virtual links in the service function chain, and sort them in descending order of latency sensitivity to determine the mapping order of virtual links; ⑤ Physical link latency sensitivity calculation module: Calculate the physical link latency sensitivity and determine the mapping order of physical links by sorting them in descending order of physical link latency sensitivity; ⑥ Elastic optical network mapping auxiliary graph construction module: Based on the target requirements of the service function chain, simplify the complexity of mapping the service function chain to the data-oriented elastic optical network, and pre-configure working paths for all physical node pairs in the elastic optical network. Configure the shortest distance between each node pair as the weight of the corresponding node pair mapping link in the optical network auxiliary graph; ⑦ Service Function Chain Virtual Link Mapping Module: In the elastic optical network mapping auxiliary diagram for data centers, all virtual links are mapped to physical links according to the principle of mapping the virtual links with the highest latency sensitivity in the service function chain to the physical links with the highest latency sensitivity on the AG. During the virtual link mapping process, it is necessary to meet the spectrum allocation and end-to-end latency constraints, and at the same time allocate the required spectrum resources for the virtual link. If resource allocation is successful, the virtual link is mapped to the physical link row; otherwise, the mapping fails. ⑧ Virtual Network Function Mapping Module: Based on the already instantiated virtual network functions and the principle of mapping virtual network functions with high computing resource requirements to physical nodes with high available computing resources, the virtual network functions are mapped to physical nodes. If the virtual network function's computing and instantiation resource requirements are met, the mapping will succeed; otherwise, the mapping will fail. ⑨ Network status monitoring module: The main functions completed include calculating the weight coefficients of influencing factors using network tomography, initializing the elastic optical network, generating service function chains, calculating the latency sensitivity of virtual links and physical links, constructing the auxiliary mapping graph for the elastic optical network, and monitoring the status of virtual links and virtual network function mapping, in order to achieve the goal of optimizing the cost-benefit of service function chain mapping. ⑩ Judgment and Early Warning Module: It performs coordination functions between various modules, as well as judgment and early warning functions for whether each module has been successfully established, in order to achieve the goal of optimizing the profit of the service function chain mapping.

[0047] Reference Figure 6 The diagram shown is a service function chain diagram; refer to Figure 7 The diagram shows the topology of a 6-node, 8-link elastic optical network. Virtual network functions are represented by regular hexagons. A, B, and E represent the types of virtual network functions, respectively. The numbers within the squares represent the number of instantiated resources required for that virtual network function, and the numbers in the rectangles below the squares represent the number of computing resources required. Dashed arrows between virtual network functions represent virtual links, with the numbers on the arrows representing the bandwidth resources required for that virtual link. The direction of the arrows indicates the flow of data within the service function chain. The circle at the left end of the service function chain represents the initiator of the service function chain request, and the circle at the right end represents the receiver of the service function chain request. The numbers within the circles represent the physical nodes where the service function chain request initiator and receiver are located. Assume the end-to-end latency requirement for the service function chain is 1 second.

[0048] Based on the above embodiments, in this embodiment of the invention, the service function chain mapping method for elastic optical networks provided by the present invention is used to map such... Figure 6 The service function chain shown is mapped to, as follows: Figure 7 In the 6-node, 8-link elastic optical network shown, the specific steps include: S301: Establish a network hierarchical analysis model for the service function chain mapping problem and calculate and obtain the adjustment factor; Network Hierarchy Process (HP) is based on comparison matrices, which are also called judgment matrices and are represented as follows: ; The elements in the matrix represent the importance of pairwise element groups. Since the goal of the decision is to obtain a physical path that is more sensitive to time delay, time delay... The first element is the most important of the four elements, so the fourth row of the matrix consists entirely of positive integers. The second most important element is the computing resources of the data center nodes at both ends of the physical link. Because processing latency in the time delay is affected by the remaining available computing resources of data center nodes in the physical network, only the fourth column of the third row of the matrix is ​​a fraction; the rest are positive integers. The third most important factor is the physical topology environment. This is because the propagation delay in time is affected by the physical path distance. The reason is... The importance ranking is The reason for this is that the physical links in resilient optical networks for data centers are entirely laid with optical fibers, which propagate at the speed of light, resulting in propagation latency lower than processing latency. The fourth important factor is the available spectrum resources along the physical path. This constructs the entire pairwise comparison matrix. Next, a consistency check needs to be performed on the constructed pairwise comparison matrix to calculate the Consistency Index (CI), expressed as: ; The largest eigenvalue of the judgment matrix is... The order is In order to obtain First, the weight vector needs to be calculated. Then, the weight vector is normalized. Finally, the largest eigenvalue of the judgment matrix is ​​calculated, including: Weight vector , , Indicates the order of the judgment matrix; Normalized weight vector ; Determine the largest eigenvalue of the matrix: ; Consistency Indicators The larger the value of , the lower the consistency of the judgment matrix. Conversely, the lower the value of , the higher the consistency of the judgment matrix. When When the value of is 0, the judgment matrices are completely identical. This is for quantification. The magnitude of this value introduces a randomness index (RI), which is expressed as: ; in, It depends on the order of the judgment matrix. Generally, It increases with the matrix order. In network hierarchical analysis, because... It may be affected by random factors, therefore, when verifying whether the judgment matrix has high consistency, it is also necessary to consider... The impact of this. Therefore, the Consistency Ratio (CR) was introduced, calculated as follows: ; when If the condition is met, the judgment matrix passes the consistency check; otherwise, the judgment matrix fails to meet the consistency requirements. In this embodiment, the eigenvectors of the judgment matrix are (0.128, 0.056, 0.270, 0.546). T The maximum eigenvalue is 4.118, the consistency index is 0.039, and the consistency ratio is 0.04, thus passing the consistency test.

[0049] Available spectrum resources on the physical path Construct a judgment matrix based on the criteria: ; Available spectrum resources on the physical path Mainly affected by physical topology environment The influence of physical path distance, therefore Compare The importance is high, and the calculated eigenvector is (0.667, 0.333)T.

[0050] Due to physical topology environment Unaffected by F, P, C, and T, no judgment matrix needs to be constructed. The computing resources of the data center nodes at both ends of the physical link are considered. The judgment matrix constructed based on the criteria is expressed as follows: ; Computing resources of data center nodes at both ends of the physical link This is a special case because it is affected by its own elements. The remaining available computing resources of a data center node. Subject to the total computing resources of this data center node Impact, therefore Compare The importance is high, and the calculated eigenvector is (0.333, 0.667)T.

[0051] With physical link time delay Construct a judgment matrix based on the criteria. ; Physical link time delay This includes processing latency and propagation latency. Processing latency is affected by the computing resources of the data center nodes at both ends of the physical link. Remaining available computing resources The propagation delay is affected by the physical topology environment. The impact of physical path distance. Because the physical links in a resilient optical network for data centers are entirely laid with optical fibers, and the propagation speed is the speed of light, the resulting propagation delay is lower than the processing delay, therefore... In other words, Compare The importance is high, and the calculated eigenvector is (0.081, 0.188, 0.731)T.

[0052] The judgment matrices constructed based on the available spectrum resources on the physical path, the computing resources of the data center nodes at both ends of the physical link, and the time delay of the physical link all passed the consistency test. The calculated eigenvectors are arranged in order to form a weighted matrix (WM), which is represented as follows: ; Based on the pairwise relationships between the elements in the element group, the resulting hypermatrix is ​​shown below: ; The normalized eigenvectors of each pairwise comparison matrix are enclosed in boxes within the hypermatrix. These pairwise comparison matrices have all passed the consistency test. The positional relationships between the eigenvectors and 0 in the table correspond one-to-one with the positional relationships in the weighted matrix. Therefore, the eigenvector at the corresponding position can be multiplied by the value at the corresponding position in the weighted matrix to obtain the weighted hypermatrix, represented as: ; The sum of the numbers circled in the box in the weighted hypermatrix equals the value at the corresponding position in the element group of the weighted matrix. The essence of a weighted hypermatrix is ​​column normalization of the hypermatrix. After obtaining the weighted hypermatrix, iterative calculations are performed using self-multiplication until the values ​​in each row are equal, thus obtaining the limiting hypermatrix, represented as: ; After each self-multiplication of the weighted hypermatrix, column normalization is required; in this embodiment, 10,000 iterations were performed, and three decimal places were retained to obtain the limiting hypermatrix. Normalizing the column vectors enclosed in the box within the limit hypermatrix yields the weights of the three element groups that influence the physical link latency sensitivity. Adjustment factor , and The values ​​are 0.14, 0.28, and 0.58, respectively. The weight coefficients of the mapping method proposed in this invention change with the judgment index, and reasonable weight coefficient values ​​can improve the performance of service function chain mapping.

[0053] S302: Calculate the distance of the shortest path between all pairs of physical nodes and construct an auxiliary graph for the physical optical network mapping; Reference Figure 8The diagram shown is an auxiliary diagram for mapping physical optical networks. The numbers on the physical links between node pairs represent the available bandwidth resources corresponding to this node pair in the elastic optical network. The numbers in parentheses represent the distance of the physical links, that is, the distance of the shortest path between connected physical nodes. S303: Calculate the latency sensitivity of the virtual link based on the bandwidth requirements of the virtual link in the service function chain and the computational and instantiation resource requirements of the virtual network function.

[0054] The latency sensitivity of the virtual link between the request initiator and the virtual network function of function type A in the service function chain is 13. The latency sensitivity of the virtual link between the virtual network function of function type A and the virtual network function of function type B is 19. The latency sensitivity of the virtual link between the virtual network function of function type B and the virtual network function of function type E is 17. The latency sensitivity of the virtual link between the virtual network function of function type E and the request receiver is 8. Therefore, the mapping order of each virtual link in the service function chain can be determined.

[0055] S304: The first step is to map the virtual link between virtual network functions of type A and type B. The virtual network function of type A has been instantiated on physical nodes 0 and 1, but since the service function chain's request originates from physical node 0, physical node 0 cannot be a candidate mapping physical node. The remaining available computing resources on physical node 1 can meet the computing resource requirements of the virtual network function of type A, therefore physical node 1 becomes a candidate mapping physical node for the virtual network function of type A. Since the virtual network function of type B has been instantiated on physical nodes 2 and 5, and both meet the computing resource requirements of the virtual network function of type B, physical nodes 2 and 5 both become candidate mapping physical nodes for the virtual network function of type B. This results in two candidate mapping links: physical nodes 1-2 and 1-5. The remaining available bandwidth resources of these two candidate mapping links can meet the bandwidth resource requirements of the virtual link between the virtual network functions of type A and type B. Based on the spectrum resource occupancy, distance, remaining available computing resources, total computing resources, and connectivity of physical links in the elastic optical network, the latency sensitivity of each physical link is calculated. The latency sensitivity of the physical link between physical nodes 1 and 2 is 0.693, and the latency sensitivity of the physical link between physical nodes 1 and 5 is 1.059. The propagation latency on the physical link between physical nodes 1 and 5 is 0.003 seconds. The processing latency generated by mapping the virtual network function of type A to physical node 1 is 0.333 seconds, and the processing latency generated by mapping the virtual network function of type B to physical node 5 is 0.133 seconds. The sum of these three latency values ​​is 0.47 seconds, which is less than the end-to-end latency requirement of this service function chain. Therefore, the physical link between physical nodes 1 and 5 will be selected to carry the virtual link between the virtual network function of type A and the virtual network function of type B. The physical link between physical nodes 1 and 5 will be put into a pending mapping state, and physical node 5 will become the pending mapping physical node for the virtual network function of type B.

[0056] S305: Next, map the virtual link between the B-type virtual network function and the E-type virtual network function. Since physical nodes 0, 1, 3, and 5 cannot be candidate mapping physical nodes, it is only necessary to determine whether physical nodes 2 and 4 meet the mapping constraints. Physical node 2 has not instantiated an E-type virtual network function, and its remaining available computing resources cannot meet the instantiation and computing resource requirements of the E-type virtual network function; therefore, physical node 2 cannot be a candidate mapping physical node. The E-type virtual network function has been instantiated on physical node 4, and its remaining available computing resources can meet the computing resource requirements of the E-type virtual network function; therefore, physical node 4 becomes a candidate mapping physical node for the E-type virtual network function. The remaining available bandwidth resources of the physical link between physical nodes 5 and 4 can meet the bandwidth resource requirements of the virtual link between the B-type and E-type virtual network functions. The propagation delay on the physical link between physical nodes 5 and 4 is 0.003 seconds, and the processing delay generated by mapping the E-type virtual network function to physical node 4 is 0.25 seconds. The sum of the two delays is 0.253 seconds, which, when added to the previous delay, results in 0.723 seconds, less than the end-to-end latency requirement of this service function chain. Therefore, the physical link between physical nodes 5 and 4 will be selected to carry the virtual link between the virtual network function of type B and the virtual network function of type E, and the physical link between physical nodes 5 and 4 will be put into a pending mapping state.

[0057] S306: Then, the virtual link between the request initiator and the virtual network function of function type A is mapped. Since the request initiator is located at physical node 0, the remaining available bandwidth resources on the physical link between physical nodes 0 and 1 meet the bandwidth resource requirements of this virtual link. The propagation delay on the physical link between physical nodes 0 and 1 is 0.0026 seconds, which, when summed with the previous delay, is 0.725 seconds, less than the end-to-end latency requirement of this service function chain. Therefore, the physical link between physical nodes 0 and 1 will be selected to carry the virtual link between the request initiator and the virtual network function of function type A, and the physical link between physical nodes 0 and 1 will be put into a pending mapping state.

[0058] S307: Finally, map the virtual link between the E-function type virtual network function and the request receiver. Since the request receiver is located at physical node 3, the remaining available bandwidth resources on the physical link between physical nodes 4 and 3 meet the bandwidth requirements of this virtual link. The propagation delay on the physical link between physical nodes 4 and 3 is 0.003 seconds, which, when summed with the previous delay, is 0.728 seconds, less than the end-to-end latency requirement of this service function chain. Therefore, the physical link between physical nodes 4 and 3 will be selected to carry the virtual link between the E-function type virtual network function and the request receiver, and the physical link between physical nodes 4 and 3 will be placed in a pending mapping state.

[0059] S308: At this point, all virtual links and virtual network functions in this service function chain have found the physical paths and physical nodes to be mapped, and the service function chain request is mapped to the underlying physical network. After updating the available resources in the elastic optical network, the next service function chain is traversed until all service function chain mapping requests are completed. (Refer to...) Figure 9 The diagram shown is a schematic representation of the service function chain mapping result.

[0060] S309: After processing all service function chain mapping requests, perform performance analysis on the service function chain mapping system.

[0061] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0062] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0063] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0064] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0065] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A service function chain mapping method applied to elastic optical networks, characterized in that, include: Obtain the request initiator, request receiver, and end-to-end latency requirements of the service function chain to be mapped; the virtual network function sequence and the function type, computing resource requirements, and instantiation resource requirements of each virtual network function; and the virtual link set and the bandwidth and spectrum resource requirements of each virtual link. Based on the bandwidth resource requirements of the virtual link, as well as the computational and instantiation resource requirements of the virtual network functions at both ends of the virtual link, the latency sensitivity of the virtual link is calculated; all virtual links in the virtual link set are sorted in descending order of virtual link latency sensitivity to obtain the virtual link mapping set. Based on the total computing resources, remaining available computing resources, instantiated virtual network function types, remaining available bandwidth resources, remaining available spectrum resources, and physical distance of each physical node in the elastic optical network, a mapping auxiliary graph is constructed. Based on the computational resources of the physical nodes at both ends of the physical link in the mapping auxiliary graph and the spectrum resources and physical distance on their shortest paths, the physical link latency sensitivity is calculated; all physical links in the elastic optical network are sorted in descending order of physical link latency sensitivity to obtain the physical link mapping set; Based on the principle of mapping the virtual links with the highest latency sensitivity in the virtual link mapping set to the physical links with the highest latency sensitivity in the physical link mapping set, the virtual links in the virtual link mapping set are mapped sequentially.

2. The service function chain mapping method for elastic optical networks according to claim 1, characterized in that, Based on the bandwidth resource requirements of the virtual link, and the computational and instantiation resource requirements of the virtual network functions at both ends of the virtual link, the latency sensitivity of the virtual link is calculated and expressed as follows: ； in, Indicates virtual network function and Virtual link latency sensitivity of virtual links between them Indicates the first A set of virtual links in a service function chain; and These represent the virtual network functions at both ends of the virtual link. and The computing resource requirements; and These represent the virtual network functions at both ends of the virtual link. and The instantiation resource requirements; Indicates the first Virtual links on each service function chain The bandwidth resource requirements.

3. The service function chain mapping method for elastic optical networks according to claim 1, characterized in that, Based on the computing resources of the physical nodes at both ends of the physical link and the spectrum resources and physical distance along its shortest path, the latency sensitivity of the physical link is calculated, including: Based on physical nodes and Calculate the impact of spectrum resources on the shortest path between them. , is represented as: ; Based on physical nodes and Connectivity and physical nodes and Calculate the physical distance of the shortest path between them, and calculate the physical topology impact. , is represented as: ； Based on physical nodes and The remaining available computing resources, and the impact of computing resources. , is represented as: ; Based on regulation factors , and The impact of spectrum resources, physical topology, and computing resources is weighted and summed to obtain the physical node. and Physical link latency sensitivity between physical links , is represented as: ; in, Indicates from physical node To physical node The number of spectral slots that are idle on the shortest path between them. and These represent the maximum and minimum values ​​of the number of idle spectrum slots on the shortest path between all pairs of physical nodes, respectively. For smoothing parameters; Represents physical nodes To physical node The physical distance of the shortest path between them. and These represent the maximum and minimum physical distances of the shortest paths between all pairs of physical nodes, respectively. and Representing physical nodes Connectivity and physical nodes connectivity, , physical nodes connectivity , Indicates direct connection between physical nodes in a physical network A set of connected physical nodes. Represents the set of all physical nodes in a physical network. Indicates direct connection to physical nodes Connected physical nodes The degree of connectivity; As a binary variable, when compared with a physical node and When directly connected, The value is 1 if it is not 1, otherwise it is 0. and These represent the maximum and minimum connectivity values ​​among all physical nodes in the physical network, respectively. and Representing physical nodes With physical nodes The remaining available computing resources; and These represent the maximum and minimum values ​​of the remaining available computing resources for all physical nodes in the physical network, respectively. and Representing physical nodes With physical nodes Total computing resources; and These represent the maximum and minimum values ​​of the total computing resources of all physical nodes in the physical network, respectively; adjustment factor. , and ,satisfy .

4. The service function chain mapping method for elastic optical networks according to claim 3, characterized in that, The moderating factor is calculated using the analytic hierarchy process (AHP), including: The elements are the physical topology, available spectrum resources, time delay, and total computing resources of the physical nodes at both ends of each physical link in the elastic optical network. Each element group is used as a criterion, and based on the scaling method, the element groups are compared pairwise to obtain multiple judgment matrices; For each judgment matrix, the initial weight of each row element in the judgment matrix is ​​calculated using the geometric mean method and then normalized to obtain the normalized weight vector of each row element in the judgment matrix. Based on the eigenvectors and the largest eigenvalue of the judgment matrix, a consistency check is performed on the judgment matrix. If the consistency check passes, the eigenvectors of all judgment matrices are placed in order to form a weighted matrix. Based on the normalized weight vectors between each pair of elements to be compared, a supermatrix is ​​constructed by filling in the corresponding elements according to their positions. Multiply the weighted matrix by the values ​​at corresponding positions in the hypermatrix to obtain the weighted hypermatrix; After performing self-multiplication iteration on the weighted hypermatrix, the limiting hypermatrix is ​​obtained; Obtain the column vectors at preset positions in the limiting hypermatrix, normalize them, and obtain three adjustment factors.

5. The service function chain mapping method for elastic optical networks according to claim 4, characterized in that, The consistency check of the judgment matrix includes: Based on the judgment matrix The value of each element in Calculate the first judgment matrix. Initial weight of the row , is represented as: , , Indicates the order of the judgment matrix; For the The initial weights of the rows are normalized to obtain the first row. normalized weight vector of rows , is represented as: ; Based on the judgment matrix and the first The normalized weight vector of the rows is used to obtain the largest eigenvalue of the judgment matrix. , is represented as: ; Based on the maximum eigenvalue of the judgment matrix With order Calculate the consistency index of the judgment matrix. , is represented as: ; Based on the consistency indices of judgment matrices of different orders, a random consistency index of the judgment matrix is ​​calculated. , is represented as: ; The ratio of the consistency index to the random consistency index is calculated and used as a test factor. , is represented as: ; The verification factor is compared with a preset threshold. If the verification factor is less than the preset threshold, the matrix is ​​judged to pass the consistency check.

6. The service function chain mapping method applied to elastic optical networks according to claim 1, characterized in that, After obtaining the physical link mapping set, it also includes: Based on the virtual network function types that each physical node in the elastic optical network has already instantiated, as well as the request initiator and request receiver, and prioritizing physical nodes with instantiation experience, candidate physical paths are selected from the physical link mapping set to construct a candidate physical path mapping set. Based on the principle of mapping the virtual links with the highest latency sensitivity in the virtual link mapping set to the physical links with the highest latency sensitivity in the candidate physical path mapping set, the virtual links in the virtual link mapping set are mapped sequentially.

7. The service function chain mapping method for elastic optical networks according to claim 6, characterized in that, Construct a set of candidate physical path maps, including: If the source node of the virtual link is the request initiator: In the physical link mapping set, select all physical links with the request initiator as the source node to form the first link candidate set; sort the physical links in the first link candidate set whose destination node is the virtual network function type corresponding to the instantiated virtual link destination node in descending order according to the physical link latency sensitivity to form the candidate physical path mapping set; sort the remaining physical paths in the first link candidate set in descending order according to the physical link latency sensitivity and connect them sequentially to the candidate physical path mapping set. If the destination node of the virtual link is the request receiver: select all physical links in the physical link mapping set that have the request receiver as the destination node to form the first link candidate set; sort the physical links in the first link candidate set whose source node is the virtual network function type corresponding to the virtual link source node that has been instantiated, in descending order of physical link latency sensitivity to form the candidate physical path mapping set; sort the remaining physical paths in the first link candidate set in descending order of physical link latency sensitivity and connect them sequentially to the candidate physical path mapping set. If the virtual network function type corresponding to the source node of the virtual link has been instantiated on the physical node: obtain all physical nodes that have instantiated the virtual network function type corresponding to the source node of the virtual link, and form a first node set; in the physical link mapping set, obtain all physical links whose source node is a physical node in the first node set, sort them in descending order of physical link latency sensitivity, and form a candidate physical path mapping set; sort the remaining physical paths in the physical link mapping set in descending order of physical link latency sensitivity, and connect them sequentially to the candidate physical path mapping set; If the virtual network function type corresponding to the destination node of the virtual link has been instantiated on the physical node: obtain all physical nodes that have instantiated the virtual network function type corresponding to the destination node of the virtual link, and form a first node set; in the physical link mapping set, obtain all physical links whose destination nodes are physical nodes in the first node set, sort them in descending order according to the physical link latency sensitivity, and form a candidate physical path mapping set; sort the remaining physical paths in the physical link mapping set in descending order according to the physical link latency sensitivity, and connect them sequentially to the candidate physical path mapping set; If the virtual network function types corresponding to the source and destination nodes of a virtual link have both been instantiated on physical nodes, obtain all physical nodes that have instantiated the virtual network function types corresponding to the source and destination nodes of the virtual link, and form a first node set; in the physical link mapping set, obtain all physical links whose source and destination nodes are physical nodes in the first node set, sort them in descending order of physical link latency sensitivity, and form a candidate physical path mapping set; sort the remaining physical paths in the physical link mapping set in descending order of physical link latency sensitivity, and connect them sequentially to the candidate physical path mapping set; If neither the virtual network function type corresponding to the source node nor the destination node of the virtual link has been instantiated on the physical node, the physical link mapping set is used as the candidate physical path mapping set.

8. The service function chain mapping method applied to elastic optical networks according to claim 1, characterized in that, Based on the principle of mapping the virtual links with the highest latency sensitivity in the virtual link mapping set to the physical links with the highest latency sensitivity in the physical link mapping set, the virtual links in the virtual link mapping set are mapped sequentially, including: If the virtual link mapping set is the The bandwidth resource requirement of the virtual link is greater than that of the first virtual link in the physical link mapping set. The remaining available bandwidth resources of the physical links then let The next physical link in the physical link mapping set is determined until the remaining available bandwidth resources meet the requirements of the first step. The physical link that meets the bandwidth resource requirements of the virtual link is used as the target physical link; Obtain the target physical link Find the shortest path and initialize it. Using the first hit method to determine the first Do the remaining available spectrum resources on the shortest path satisfy the condition of the first shortest path? Spectrum resource requirements for each virtual link: If satisfied, then use the first... Find the shortest path and use it as the target path; If not satisfied, then let Repeatedly use the first hit method to determine the first hit. Does the remaining available spectrum resource on the shortest path satisfy the condition of the first shortest path? The spectrum resource requirements of each virtual link are determined until the first one is obtained. The shortest path for the spectrum resource requirements of each virtual link is taken as the target path; Based on the target path, for the first The virtual link will allocate spectrum resources to the first one. Each virtual link is mapped to the target physical link; make , for the Map each virtual link until... Complete the mapping of the service function chain to be mapped; This indicates the total number of virtual links in the virtual link set.

9. The service function chain mapping method for elastic optical networks according to claim 8, characterized in that, Based on the target path, for the first After allocating spectrum resources to each virtual link, the following is also included: Get the The propagation delay generated by mapping a virtual link to the target physical link; Obtain the processing latency incurred from mapping virtual network functions to physical nodes; Get the End-to-end latency of a virtual link; If the sum of propagation delay, processing delay, and end-to-end delay meets the end-to-end delay requirement, then the [missing information] will be [missing information]. A virtual link is mapped to the target physical link.

10. The service function chain mapping method applied to elastic optical networks according to claim 1, characterized in that, Situations where the service function chain to be mapped is blocked include: After traversing all physical links in the physical link mapping set, no remaining available bandwidth resources were found that satisfy the condition. If a physical link requires bandwidth resources from a virtual link, then the service function chain to be mapped is marked as blocked. Complete traversal After the shortest path, it is impossible to obtain the remaining available spectrum resources to satisfy the first... When the shortest path for the spectrum resource requirements of a virtual link is found, the service function chain to be mapped is marked as blocked; If the sum of propagation delay, processing delay, and end-to-end delay does not meet the end-to-end delay requirement, then the service function chain to be mapped is marked as blocked.

Images