Active Migration Method of Network Service Function Chain in Core Network

Abstract

The present invention discloses a method for active migration of a network service function chain in a core network. The scheme of the present invention models the network service function chain migration problem as an integer linear programming model in view of the dynamic changes in VNF's resource demand in a 5G network, and proposes an active migration strategy. This method strategy can effectively reduce the imbalance of physical resource distribution. At the same time, the migration strategy of this scheme comprehensively considers the constraints of bandwidth, latency, computing resources, and transmission paths that need to be considered when migrating the network service function chain; the optimization goal of the integer linear programming model is to minimize the standard deviation of resource distribution and the additional overhead caused by migration.

Description

Technical Field

[0001] The present invention relates to the field of network communication technology, and in particular to a method for actively migrating a network service function chain in a core network. Background Art

[0002] As technology develops, a range of new vertical use cases have emerged, such as online education, virtual reality, and autonomous manufacturing. However, today's mobile communication networks adopt a one-size-fits-all approach to provide services without considering the different needs of vertical businesses. Therefore, it is necessary to update the network architecture to meet different application requirements.

[0003] With the development of Network Function Virtualization (NFV) and Software Defined Network (SDN) technologies, network service function chain (SFC) has been proposed as a key architectural technology to provide flexible and adaptable services on a single physical network. A service function chain (SFC) consists of a set of virtual nodes and virtual links, where virtual nodes can be implemented through NFV and virtual links can be composed of one or more physical links. When a service function chain request is received, the management and orchestration (MANO) layer of the network system will allocate resources and configure the service function chain. The MANO layer is also responsible for resource recycling after the service function chain life cycle ends. In addition, in order to avoid Quality of Service (QoS) and Service Level Agreement (SLA) violations caused by traffic load changes in the service function chain, MANO should dynamically adjust the resource allocation of each service function chain. However, the above process may lead to uncertain resource distribution of service function chain requests in the service queue, resulting in uneven distribution of physical resources. In order to eliminate the negative impact of unbalanced physical resource allocation on subsequent service function chain requests and network operators (such as low service function chain request acceptance rate and low long-term profit), a reasonable service function chain request migration strategy is crucial. Summary of the invention

[0004] In view of this, the purpose of the present invention is to propose a method for active migration of network service function chains in a core network. This method strategy can solve the problem of uneven allocation of existing physical resources and the negative impact on subsequent service function chain requests and operators.

[0005] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0006] A method for active migration of a network service function chain in a core network, comprising:

[0007] S1. Establishing a corresponding network model, migration digital model, resource distribution model and / or migration cost model in view of the dynamic changes of network migration resources;

[0008] S2. Determine the service function chain that needs to be migrated;

[0009] S3. Reclaim resources for the service function chain that needs to be migrated according to preset conditions;

[0010] S4. Formulate a resource allocation plan for the service function chain that needs to be migrated according to preset conditions;

[0011] S5. Migrate the service function chain that needs to be migrated according to the resource allocation plan, and then remap resources and output the migrated service function chain to complete the migration.

[0012] As a possible implementation mode, further, in the present solution S1, establishing the network model includes:

[0013] The underlying network is represented as a weighted directed graph, denoted as G = (N, E), where N represents the set of physical nodes and E represents the set of physical links. Since the underlying network nodes are nodes N with VNF functions, V and normal node N F Therefore, the set of physical nodes can be expressed as N = N v ∪N F , where nodes with VNF functions can provide different types of VNF functions, which at least include network address translation and firewall, and ordinary nodes are only used to provide data packet forwarding functions.

[0014] As a preferred implementation method, preferably, the present solution further includes: defining the network service function chain r in the network as a linear chain, the specific structure of which is: r = (s r ,d r ,F r ,τ r ,c r ,b r ), where s r , d r represents the starting and destination nodes of the service function chain r, F r represents the VNF demand of service function chain r, τ r , c r , b r They are respectively expressed as delay requirement, CPU resource requirement of VNF in the service function chain, and bandwidth resource requirement.

[0015] As a preferred implementation method, preferably, the scheme for establishing the migration digital model includes: i (π) indicates whether node i is capable of processing VNFπ, Expressed as Is it mapped on node i? Represents the mth VNF of service function chain r.

[0016] As a preferred implementation method, preferably, the VNF described in this solution has the following constraints:

[0017] The requirement is to ensure that each VNF is placed in an N V Satisfaction

[0018] It is required that VNF ​​can only be mapped on N V Satisfaction

[0019] Requires each N V The maximum value is F r Provide a VNF to meet

[0020] Among them, the virtual link mapping constraints of the network service function chain are considered to include:

[0021] It is required that path splitting is not allowed;

[0022] The total delay of the transmission link is required to be less than the delay required by the service function chain.

[0023] The transmission path must pass through the physical nodes that provide VNF function services for the service function chain.

[0024] Among them, P r is the transmission path of service function chain r, M r Make the set of mapping nodes corresponding to the service function chain r;

[0025] The constraints on the upper bounds of resources in the network include:

[0026] Require node resource constraints to satisfy

[0027] Among them, C i represents the total computing resources of node i, R C (i) Computational resources consumed on node i C r Represents the computing resources of a single VNF in service function chain r;

[0028] Link resource constraints are required to satisfy

[0029] Among them, R B (i,j) represents the bandwidth resources consumed by physical link (i,j) B i,j Represented as the total bandwidth resource of the physical link (i, j); where, is a binary variable, indicating the virtual link Is it mapped on the physical link (i, j), Represents a virtual node arrive The virtual link between r Represents the bandwidth resources required by the service function chain r.

[0030] As a preferred implementation method, preferably, the method for establishing the resource distribution model in this solution includes:

[0031] The standard deviation formula is used to measure the distribution of physical resources, which can be calculated as:

[0032]

[0033] Among them, NUM(N V ) represents the number of available VNF nodes, NUM(E) represents the number of physical links, Represents node N V The average resource consumption is Indicates the average resource consumption of the physical link.

[0034] As a preferred implementation mode, preferably, the method for establishing the migration cost model in this solution includes:

[0035] The migration cost of a single service function chain is specifically expressed as

[0036] in Represents the transmission path before migration, Represented as the transmission path after migration, represents the difference between the paths before and after migration, u r (t) represents the flow of service function chain r at time t, t r is the time before migration, is the time when the migration ends;

[0037] The total migration cost of all service function chains is specifically expressed as

[0038] As a preferred implementation method, preferably, in this solution S2, the service function chain to be migrated is determined as follows:

[0039] All service function chains in the network are chained by formula Sort by r Represents the computing resources required by the total service function chain r, select y r The smallest one is the service function chain that needs to be migrated;

[0040] In step S3, the resources of the service function chain that needs to be migrated are recycled specifically as follows:

[0041] The resources previously allocated to the service function chain r are reclaimed, which include computing resources in the node and bandwidth resources in the physical link.

[0042] As a preferred implementation mode, preferably, step S4 of this solution includes:

[0043] The migration of the service function chain is divided into the migration of VNF and the migration of virtual links;

[0044] VNF migration includes:

[0045] Greedily placing VNFs on the physical nodes with the most remaining resources;

[0046] Migration of virtual links includes:

[0047] Remove the physical links in the network that cannot meet the bandwidth requirements of the service function chain r, and obtain a new network;

[0048] Find all paths from the starting node to the ending node of the service function chain r in the new network by using the DFS algorithm;

[0049] Select the path that meets the virtual link constraints to form a set of feasible solutions P r ;

[0050] Based on the resource distribution model formula, the feasible solution set P of the path is obtained. r The standard deviation of the remaining bandwidth resources in;

[0051] In the path feasible solution set P r Find the path p with the smallest standard deviation of the remaining bandwidth resources, and place the service function chain on the path p.

[0052] In this solution, the network may be a 5G network. Aiming at the dynamic changes in resource demands of VNFs in the 5G network, the network service function chain migration problem is modeled as an integer linear programming model, and an active migration strategy is proposed.

[0053] Based on the above, the present invention also provides a computer-readable storage medium, in which at least one instruction, at least one program, code set or instruction set is stored, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the above-mentioned method for active migration of network service function chains in the core network.

[0054] By adopting the above-mentioned technical scheme, the present invention has the following beneficial effects compared with the prior art: the scheme of the present invention models the network service function chain migration problem as an integer linear programming model in view of the dynamic changes in the resource demand of VNF in the 5G network, and proposes an active migration strategy. This method strategy can effectively reduce the imbalance of physical resource distribution. At the same time, the migration strategy of this scheme comprehensively considers the constraints of bandwidth, latency, computing resources, and transmission path that need to be considered when migrating the network service function chain; the optimization goal of the integer linear programming model is to minimize the standard deviation of resource distribution and the additional overhead caused by migration. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 It is the algorithm flow chart of the present invention;

[0057] Figure 2 is a brief schematic diagram corresponding to the network service function chain in the embodiment of the present invention;

[0058] Figure 3 is an example network diagram of the present invention;

[0059] Figure 4 A comparison diagram of the standard deviation before network migration, the standard deviation after migration using a conservative migration algorithm (comparison algorithm), and the standard deviation after migration using the algorithm of the present invention in one example of the present invention;

[0060] Figure 5 The figure is a comparison chart of the standard deviation before network migration, the standard deviation after migration using the conservative migration algorithm, and the standard deviation after migration using the algorithm of the present invention under different examples of the present invention. DETAILED DESCRIPTION

[0061] The present invention will be further described in detail below in conjunction with the accompanying drawings and examples. It is particularly noted that the following examples are only used to illustrate the present invention, but are not intended to limit the scope of the present invention. Similarly, the following examples are only partial embodiments of the present invention rather than all embodiments, and all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0062] This embodiment provides a method for active migration of a network service function chain in a core network, which includes:

[0063] S1. Establishing a corresponding network model, migration digital model, resource distribution model and / or migration cost model in view of the dynamic changes of network migration resources;

[0064] S2. Determine the service function chain that needs to be migrated;

[0065] S3. Reclaim resources for the service function chain that needs to be migrated according to preset conditions;

[0066] S4. Formulate a resource allocation plan for the service function chain that needs to be migrated according to preset conditions;

[0067] S5. Migrate the service function chain that needs to be migrated according to the resource allocation plan, and then remap resources and output the migrated service function chain to complete the migration.

[0068] Among them, in this scheme S1, establishing a network model includes:

[0069] The underlying network is represented as a weighted directed graph, denoted as G = (N, E), where N represents the set of physical nodes and E represents the set of physical links. Since the underlying network nodes are nodes N with VNF functions, V and normal node N F Therefore, the set of physical nodes can be expressed as N = N v ∪N F , where nodes with VNF functions can provide different types of VNF functions, which at least include network address translation and firewall, and ordinary nodes are only used to provide data packet forwarding functions.

[0070] In addition, this solution also includes: defining the network service function chain r in the network as a linear chain, and its specific structure is: r = (s r ,d r ,F r ,τ r ,c r ,b r ), where s r , d r represents the starting and destination nodes of the service function chain r, Fr represents the VNF demand of service function chain r, τ r , c r , b r They are respectively expressed as delay requirement, CPU resource requirement of VNF in the service function chain, and bandwidth resource requirement; for example, the following formula:

[0071] r3=(s2,d6,{VNF1,VNF2,VNF3},τ3,c3,b3)

[0072] In this solution S1, establishing the migration digital model includes: modeling the network model and defining binary variables; i (π) indicates whether node i is capable of processing VNFπ, Expressed as Is it mapped on node i? Represents the mth VNF of service function chain r.

[0073] Preferably, the VNF described in this solution has the following constraints:

[0074] The requirement is to ensure that each VNF is placed in an N V Satisfaction

[0075] It is required that VNF ​​can only be mapped on N V Satisfaction

[0076] Requires each N V The maximum value is F r Provide a VNF to meet

[0077] Among them, the virtual link mapping constraints of the network service function chain are considered to include:

[0078] It is required that path splitting is not allowed;

[0079] The total delay of the transmission link is required to be less than the delay required by the service function chain.

[0080] The transmission path must pass through the physical nodes that provide VNF function services for the service function chain.

[0081] Among them, P r is the transmission path of service function chain r, M r Make the set of mapping nodes corresponding to the service function chain r;

[0082] The constraints on the upper bounds of resources in the network include:

[0083] Require node resource constraints to satisfy

[0084] Among them, C i represents the total computing resources of node i, R C (i) Computational resources consumed on node i C r Represents the computing resources of a single VNF in service function chain r;

[0085] Link resource constraints are required to satisfy

[0086] Among them, R B (i,j) represents the bandwidth resources consumed by physical link (i,j) B i,j Represented as the total bandwidth resource of the physical link (i, j); where, is a binary variable, indicating the virtual link Is it mapped on the physical link (i, j), Represents a virtual node arrive The virtual link between r Represents the bandwidth resources required by the service function chain r.

[0087] In this solution S1, the method for establishing the resource distribution model includes:

[0088] The standard deviation formula is used to measure the distribution of physical resources, which can be calculated as:

[0089]

[0090] Among them, NUM(N V ) represents the number of available VNF nodes, NUM(E) represents the number of physical links, Represents node N V The average resource consumption is Indicates the average resource consumption of the physical link.

[0091] In this solution S1, the method for establishing the migration cost model includes:

[0092] Assuming that the migration cost is related to the time required for service function chain migration, the placement of VNFs and the change of transmission paths, and the traffic in the service function chain, the migration cost of a single service function chain is specifically expressed as

[0093] in Represents the transmission path before migration, Represented as the transmission path after migration, represents the difference between the paths before and after migration, u r(t) represents the flow of service function chain r at time t, t r is the time before migration, is the time when the migration ends;

[0094] To this end, the total migration cost of all service function chains is specifically expressed as

[0095] In this solution S2, the service function chain to be migrated is determined as follows:

[0096] All service function chains in the network are chained by formula Sort by r Represents the computing resources required by the total service function chain r, select y r The smallest one is the service function chain that needs to be migrated;

[0097] In step S3 of this solution, the resources of the service function chain that needs to be migrated are recycled specifically as follows:

[0098] The resources previously allocated to the service function chain r are reclaimed, which include computing resources in the node and bandwidth resources in the physical link.

[0099] The pseudo code of the algorithm of step S3 is represented by the following algorithm 1:

[0100]

[0101] Step S4 of this scheme includes:

[0102] The migration of the service function chain is divided into the migration of VNF and the migration of virtual links;

[0103] VNF migration includes:

[0104] In order to improve resource utilization and reduce the resource load of network nodes, VNFs are greedily placed on the physical nodes with the most remaining resources, which is expressed as where i mos t represents the physical node with the most available resources. This process is mainly reflected in lines 5 to 8 of the following Algorithm 2;

[0105] Migration of virtual links includes:

[0106] According to the algorithm 1 described in step 3 above, the physical links that cannot meet the bandwidth requirements of service function chain r are removed from the network. Get a new network;

[0107] Find all paths from the starting node to the ending node of the service function chain r in the new network by using the DFS algorithm;

[0108] Selecting a path that meets the virtual link constraint satisfies the constraint. Form a set of feasible solutions P r ,This process is mainly reflected in lines 9 to 16 in the following algorithm 2;

[0109] Based on the resource distribution model formula, the feasible solution set P of the path is obtained. r The standard deviation of the remaining bandwidth resources in is specifically expressed as

[0110] In the path feasible solution set P r Find the path p with the smallest standard deviation of the remaining bandwidth resources, and place the service function chain on the path p.

[0111] In step S5 of this solution, the resource remapping of the service function chain to be migrated is specifically as follows:

[0112] The resources are remapped according to the solution obtained in the above step S4.

[0113] The pseudo code of the active migration strategy of S5 is shown in Algorithm 2 below:

[0114]

[0115]

[0116] In order to further explain this scheme, this scheme combines the following examples to illustrate and compare the above steps:

[0117] Figure 1 This is a flowchart of the method for active migration of network service function chains in the core network described in this solution. Figure 1 As can be seen, the network is composed of different nodes and physical links, some of which can only have a simple forwarding function, while others have the ability to process VNF instances. In the figure, there are three network service function chains. Each service function chain is composed of a start node, a destination node, and a VNF. Assume that the three service function chains have been mapped in the underlying network and need to be migrated. Therefore, Figure 2 As shown in (a), the service function chain with the least required resources is first found and the computing resources of its nodes and the bandwidth resources of its links are recovered. Then, the VNF nodes in the service function chain are migrated. Figure 2 As shown in (b), the virtual link is finally migrated and resources are remapped as follows Figure 2 (c) as shown.

[0118] Figure 3 For an example diagram of the network of the present invention, refer to Figure 1It can be seen that there are 7 physical nodes supporting VNF in the underlying network, which can provide various VNF services. The resources of these nodes and the bandwidth resources of the physical links are set to 10000 units, and the delay of the physical link is set to 1ms. The network parameters in this example are shown in the following table:

[0119]

[0120] In this example, the subsequent service function chain request parameters are shown as follows:

[0121]

[0122] Considering that there are permanent and temporary service function chains in real world scenarios, the present invention designs a permanent service function chain, which will occupy a large amount of physical resources, but correspondingly, their number is scarce. The other two types of service function chains are temporary service function chains. They occupy fewer physical resources than permanent service function chains, but their number is very large. In addition, the bandwidth resource requirement of the service function chain R is related to its computing resource requirement. For this reason, the bandwidth requirement b r =β·c r Where β=1.

[0123] from Figure 4 , Figure 5 It can be seen from the above that the algorithm of the present invention makes the underlying network physical resources more balanced after migration, whether in a single instance or in multiple instances. Figure 4 It can be seen that in the case of multiple instances, the network load balancing of the algorithm of the present invention is reduced by an average of 22% compared with the comparison algorithm of conservative migration.

[0124] Based on the above, it can be seen that the load balancing rate of the network is significantly reduced after the network migration of the algorithm of the present invention.

[0125] The above descriptions are only some embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any equivalent device or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for active migration of a network service function chain in a core network, characterized in that: It includes: S1. Establish corresponding network models, migration digital models, resource distribution models and migration cost models according to the dynamic changes of network migration resources; S2. Determine the service function chain that needs to be migrated; S3. Reclaim resources for the service function chain that needs to be migrated according to preset conditions; S4, divide the migration of the service function chain into the migration of VNF and the migration of virtual links; VNF migration includes: Greedily placing VNFs on the physical nodes with the most remaining resources; Migration of virtual links includes: Remove the physical links in the network that cannot meet the bandwidth requirements of the service function chain r, and obtain a new network; Find all paths from the starting node to the ending node of the service function chain r in the new network by using the DFS algorithm; Select the path that meets the virtual link constraints to form a set of feasible solutions P r ; Based on the resource distribution model formula, the feasible solution set P of the path is obtained. r The standard deviation of the remaining bandwidth resources in; In the path feasible solution set P r Find the path p with the smallest standard deviation of the remaining bandwidth resources, and place the service function chain on the path p; S5. Migrate the service function chain that needs to be migrated according to the resource allocation plan, and then remap resources and output the migrated service function chain to complete the migration; The network service function chain in the network is defined as a linear chain, which includes the start and destination nodes of the service function chain, and one or more of the delay requirement, the CPU resource requirement of the VNF in the service function chain, and the bandwidth resource requirement; In S1, establishing the migration digital model includes: i (π) indicates whether node i is capable of processing VNFπ, Expressed as Is it mapped on node i? Represents the mth VNF of service function chain r, each VNF has path and / or resource constraints.

2. The method for actively migrating a network service function chain in a core network as claimed in claim 1, characterized in that: In S1, building a network model includes: The underlying network is represented as a weighted directed graph, denoted as G = (N, E), where N represents the set of physical nodes and E represents the set of physical links. Since the underlying network nodes are nodes N with VNF functions, V and normal node N F Therefore, the set of physical nodes can be expressed as N = N v ∪N F , where nodes with VNF functions can provide different types of VNF functions, which at least include network address translation and firewall, and ordinary nodes are only used to provide data packet forwarding functions.

3. The method for actively migrating a network service function chain in a core network as claimed in claim 2, characterized in that: It also includes: defining the network service function chain r in the network as a linear chain, whose specific structure is: r = (s r ,d r ,F r ,τ r ,c r ,b r ), where s r , d r represents the starting and destination nodes of the service function chain r, F r represents the VNF demand of service function chain r, τ r ,cr,b r They are respectively expressed as delay requirement, CPU resource requirement of VNF in the service function chain, and bandwidth resource requirement.

4. The method for actively migrating a network service function chain in a core network as claimed in claim 2, characterized in that: The VNF has the following constraints: The requirement is to ensure that each VNF is placed in an N V Satisfaction It is required that VNF ​​can only be mapped on N V Satisfaction Requires each N V The maximum value is F r Provide a VNF to meet Among them, the virtual link mapping constraints of the network service function chain are considered to include: It is required that path splitting is not allowed; The total delay of the transmission link is required to be less than the delay required by the service function chain. The transmission path must pass through the physical nodes that provide VNF function services for the service function chain. Among them, P r is the transmission path of service function chain r, M r Make the set of mapping nodes corresponding to the service function chain r; The constraints on the upper bounds of resources in the network include: Require node resource constraints to satisfy Among them, C i represents the total computing resources of node i, R C (i) Computational resources consumed on node i C r Represents the computing resources of a single VNF in service function chain r; Link resource constraints are required to satisfy Among them, R B (i,j) represents the bandwidth resources consumed by physical link (i,j) B i,j Represented as the total bandwidth resource of the physical link (i, j); where, is a binary variable, indicating the virtual link Is it mapped on the physical link (i, j), Represents a virtual node arrive The virtual link between r Represents the bandwidth resources required by the service function chain r.

5. The method for actively migrating a network service function chain in a core network as claimed in claim 4, characterized in that: In S1, establishing a resource distribution model includes: The standard deviation formula is used to measure the distribution of physical resources, which can be calculated as: Among them, NUM(N V ) represents the number of available VNF nodes, NUM(E) represents the number of physical links, Represents node N V The average resource consumption is Indicates the average resource consumption of the physical link.

6. The method for actively migrating a network service function chain in a core network as claimed in claim 5, characterized in that: In S1, establishing the migration cost model includes: The migration cost of a single service function chain is specifically expressed as in It is represented as the transmission path before migration, P r m Represented as the transmission path after migration, represents the difference between the paths before and after migration, u r (t) represents the flow of service function chain r at time t, t r is the time before migration, is the time when the migration ends; The total migration cost of all service function chains is specifically expressed as 7. The method for actively migrating a network service function chain in a core network as claimed in claim 6, characterized in that: In S2, the service function chain to be migrated is determined to be: All service function chains in the network are chained by formula Sort by r Represents the computing resources required by the total service function chain r, select y r The smallest one is the service function chain that needs to be migrated; In step S3, the resources of the service function chain that needs to be migrated are recycled specifically as follows: The resources previously allocated to the service function chain r are reclaimed, which include computing resources in the node and bandwidth resources in the physical link.

8. A computer-readable storage medium, characterized in that: The storage medium stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the method for active migration of the network service function chain in the core network as described in one of claims 1 to 7.

Images