A multi-priority service function chain deployment method for virtual TSN networks

Abstract

The present invention discloses a multi-priority service function chain deployment method for a virtual TSN network, and relates to the field of communications. Specifically, the present invention combines the characteristics of the underlying physical network and the characteristics of the TSN gateway to propose a service function chain deployment method for multi-priority tasks and a TSN multi-queue injection mechanism. First, the SFC constraints and TSN characteristics are jointly considered to construct an augmented logical topology, which is suitable for common routing algorithms and greatly reduces the algorithm complexity brought by the SFC constraints. Then, the multi-priority of tasks and the multi-queue properties of TSN are considered to replicate the constructed augmented logical topology and divide the queues driven by task priorities, and finally realize lightweight end-to-end deterministic delivery of dynamic heterogeneous tasks in the virtual TSN network. The present invention solves the problem of high complexity in the existing service function chain deployment, fills the gap in related content in the virtual TSN network, and better utilizes limited communication resources to provide differentiated deterministic communication services for industrial tasks.

Description

Technical Field

[0001] The present invention relates to the field of communications, and in particular to a multi-priority service function chain deployment method for a virtual TSN network. Background Art

[0002] Time Sensitive Networking (TSN) is a series of communication protocols for deterministic communication proposed by the IEEE802.1 working group in 2016. It proposes mechanisms such as time-aware Shaper (TAS), gate control list (GCL), per-stream filtering and policing (PSFP), and parameters such as IPV (Internal Priority Value). It implements multi-priority low-latency and low-jitter communication of industrial heterogeneous data through a unified communication framework, and guarantees the upper bounds of latency and jitter with strict time windows. It is very suitable for large-scale industrial communication scenarios, but its centralized computing scheduling method brings high computational complexity and cannot handle dynamic production needs well, and it is in urgent need of flexible transformation.

[0003] Faced with the contradiction between the small batch and customized production requirements proposed by the future Industry 4.0 and the limited industrial production and communication resources, the traditional "one application-one network" model can no longer adapt to the dynamic requirements issued by the upper layer, which greatly hinders the transformation and development of the manufacturing industry and the reform and innovation of communication technology. It is urgent to transform the production communication mode. To this end, the academic and industrial circles have introduced Software-Defined Networking (SDN) and Network Function Virtualization (NFV) technologies. By abstracting and virtualizing the underlying communication network to varying degrees, the flexible deployment of service functions can be achieved without considering the specific technical implementation details of the underlying layer, thereby achieving the purpose of resource reuse and solving the bottleneck of resource shortage for industrial production transformation. Therefore, using network virtualization to model the physical network can break the underlying technical barriers and provide a flexible and scalable communication solution to better meet dynamic production needs.

[0004] In a virtual network, application services usually arrive at the network in the form of a service function chain (SFC) containing a series of ordered virtual network functions (VNF). Mapping all VNFs into the physical network in order implements the service of the SFC. Since the location, function, and connection relationship of different network modules are different, the network functions they can provide are also different. Therefore, it is necessary to consider both the requirements of SFC and the capabilities of the underlying network to map VNF, which greatly increases the complexity of the problem. In addition, the design of SFC deployment in a virtual TSN network is an almost blank area. How to deploy SFC in a virtual TSN network based on the characteristics of the TSN standard itself remains to be studied.

[0005] After searching the existing literature, it is found that the most similar implementation scheme is the Chinese patent application number: 201810541014.2, named: A TSN scheduling method for real-time application requirements, and its specific content is: Based on the software-defined network technology, a TSSDN architecture is proposed to allocate time slots to time-triggered flows and route them to maximize resource utilization. This method also considers the TSN network data flow routing problem at the resource level, but it does not perform service mapping according to the SFC mode, and its centralized scheduling algorithm has a high complexity and cannot adapt to dynamic production needs. The patent application number is: 202080014340.3, named: 5G system support for virtual TSN bridge management, QoS mapping and TSN Qbv scheduling, and the specific content is: The cellular network is connected to the TSN network through a virtual TSN bridge, and the specific TSN interface and its corresponding functions for QoS flow mapping in 5G and TSN are given, but no specific instructions are given on how to calculate the virtual TSN network routing machine path. The patent application number is: 201710536013.4, and the name is: A method for optimizing the deployment of service function chains in 5G mobile networks. The specific content is: By considering the deployment methods of VNF merging strategy, virtual reuse strategy and temporary link strategy during deployment, three mapping schemes with different optimization objectives are proposed, but the complexity of the mapping algorithm is very high and it is not suitable for dynamic application scenarios. The patent application number is: 202011359654.5, and the name is: A method for reliability deployment of service function chains based on deep reinforcement learning. The specific content is: By dynamically perceiving the priority of the service function chain, the backup scheme of the virtual network function on the service function chain is selected, and deep reinforcement learning is used to simultaneously determine the deployment nodes and backup nodes of the virtual network function in the service function chain. The routing scheme of this method is deployed in advance and cannot reflect the dynamic changes of resources. The patent application number is: 202111133810.0, and the name is: A joint optimization method for service function chains based on software-defined networks. The specific content is: using the logical function diagram to obtain the deployment plan of the service function chain, by deleting the nodes and links that do not meet the node resource capacity constraints and link bandwidth resource capacity constraints. However, this method needs to judge whether the constraints are met every time a routing path is obtained. The computational complexity is very high and there is a large application bottleneck.

[0006] Most existing TSN routing problems use a centralized approach to calculate the gating list of all TSN gateways to achieve deterministic end-to-end scheduling of data flows through the TSN network. The scheduling method is offline and cannot adapt to dynamic production scenarios. There are few existing works on virtual TSN network routing problems. The few works only use software-defined network technology to allocate and schedule data streams in a centralized manner. The scheduling method is difficult to apply to large-scale, highly dynamic heterogeneous industrial networks. Most existing SFC deployment works use iterative trial methods to simultaneously meet SFC constraints and underlying resource constraints. The computational intensity is high and the time complexity is high. Most of them do not consider the dynamic characteristics of resources for scheduling mechanism design, which is not suitable for future industrial scenarios. Most existing VNF routing works are based on SFC constraints to select the optimal path. There is little work on modifying the underlying topology based on the SFC properties. Finding the optimal path that meets SFC on an unmodified topology is a very difficult problem.

[0007] Therefore, technicians in this field are committed to developing a multi-priority service function chain deployment method for virtual TSN networks. This method solves the problem of high complexity in existing service function chain deployment and fills the gap in related mechanisms in service function chain deployment in virtual TSN networks, thereby better utilizing limited resources to provide deterministic communication services for future Industry 4.0 needs. Summary of the invention

[0008] In view of the above-mentioned defects of the prior art, the technical problem to be solved by the present invention is how to build a virtual TSN network based on software-defined network technology and network function virtualization technology, and realize dynamic resource scheduling in combination with the characteristics of the TSN standard; how to characterize the network dynamics (dynamic resource consumption, dynamic node birth and death, dynamic topology construction); how to build a more universal routing topology based on the underlying physical network and SFC constraints to facilitate the implementation of common routing algorithms; how to combine the advantages of TSN multi-queues to provide differentiated transmission services for multi-priority data in the virtual TSN network.

[0009] To achieve the above object, the present invention provides a multi-priority service function chain deployment method for a virtual TSN network, which is characterized by comprising the following steps:

[0010] Step 1: According to the SFC you have reached For each physical node n j The supported VNF set is extracted and streamlined, leaving only the VNFs that reach the SFC as the new supported VNF set for each physical node, and then VNF0 is added to the front end of the SFC;

[0011] Step 2: Adopt the augmented logical topology construction method considering SFC constraints, add VNF0 to the front end of the VNF set supported by each physical node, abstract each VNF into an independent virtual node, and form a virtual node set belonging to the physical node. Traverse and execute the connection within the physical node range, the connection between the physical node range and the topology simplification operation to obtain the initial logical augmented topology graph G' o ;

[0012] Step 3: Adopt the virtual queue augmentation mechanism based on TSN multi-queue characteristics, and augment the logical topology G' according to the number of TSN queues. o Each virtual node in is further augmented into several virtual nodes representing TSN queues, and an augmented logical topology G' is obtained that jointly considers SFC constraints and TSN characteristics. f .

[0013] Furthermore, the VNF set that each physical node can support and the SFC reached in step 1 are both a subset of the total VNF set, and may be the same or different.

[0014] Furthermore, the embedding of VNFs in the SFC in step 1 is accompanied by the consumption of computing resources and cache resources of the underlying physical nodes, and since each VNF is responsible for different functions, there are certain mapping order constraints between them.

[0015] Furthermore, in step 2, VNF0 is added to the front end of the VNF set supported by each physical node. After the VNF is extracted, the original link will not disappear, thereby avoiding changing the connection relationship of the underlying physical nodes.

[0016] Furthermore, the step 2 comprises the following steps:

[0017] Step 2.1: For the same physical node n j Virtual nodes in the virtual node set If the VNFs represented by two of the nodes are adjacent VNFs in the SFC, the two nodes are connected within the physical node range according to the SFC order to form an AC chain, and the link weight in the AC chain is 0;

[0018] Step 2.2: Execute the operation in step 2.1 for the virtual node sets of all physical nodes to obtain the AC chain in the virtual node set of each physical node and the corresponding highest VNF ​​index in the AC chain;

[0019] For virtual nodes in virtual node sets in different physical nodes, according to the highest VNF ​​index in their respective AC chains, according to the connection relationship of the underlying directed non-fully connected topology, if two adjacent nodes n i and n jThere is a directed connection between i →n j , determine node n i and n j The highest VNF ​​index of the AC chain and Whether the conditions are met: or If the condition is satisfied, the virtual nodes corresponding to the two highest indexes that meet the condition are connected according to the underlying topology; if the condition is not satisfied, the virtual nodes corresponding to the two highest indexes that meet the condition are connected according to the underlying topology; Find n i The highest VNF ​​index that meets the conditions in the AC chain Connect the virtual nodes corresponding to the two highest indexes that meet the conditions according to the underlying topology;

[0020] Step 2.3: Execute the step 2.2 operation on the virtual node sets of all physical nodes according to the physical topology until all physical nodes are traversed to establish a preliminary logical augmented topology graph G' o ;

[0021] Step 2.4: The initial augmented topological graph G' obtained by the operation in step 2.3 o , delete the nodes that only enter but not exit or only exit but not enter except the starting and ending points, to reduce the topological dimension;

[0022] Step 2.5: For node pairs that do not have virtual connections after link deletion, if the underlying network has physical connections, the node pairs re-execute the step 2.2 operation to build new links, ensuring that the logical augmented topology can fully reflect the underlying topology connection relationship and the underlying device VNF support, and establish the logical augmented topology graph G' considering SFC constraints o .

[0023] Furthermore, the step 3 comprises the following steps:

[0024] Step 3.1: For virtual nodes representing TSN queues belonging to different physical nodes, if the TSN queue priorities IPVs they represent are the same, then the link weights are also the same, that is, the link weights of all virtual TSN queue nodes with the same IPVs are the same;

[0025] Step 3.2: The resources of the physical node are evenly divided by the virtualized TSN queue nodes. If a physical node has more than one virtual node with the same queue IPV, these nodes compete for 1 / z resources inherited by the IPV, where z is the number of IPVs, i.e., the number of TSN queues.

[0026] Step 3.3: The link weight depends on the output queue IPv of the data and has nothing to do with the queue IPv to be injected.

[0027] Step 3.4: Once the data is injected into the TSN gateway, the IPV inside the gateway cannot be changed. However, the IPVs of the queues injected into the same gateway are independent of each other and do not affect each other. The final logical augmented topology G' is obtained by considering the SFC constraints and TSN multi-queue characteristics. f .

[0028] Furthermore, the number of IPv in step 3.2 is the same as the number of queues of the TSN gateway itself, which is an integer between 1 and 8.

[0029] Furthermore, the QoS requirements of industrial field data are different, and the two data flows with different priorities, real-time traffic (RTT) and best-effort traffic (BET), are topologically segmented in a differentiated manner; the following steps are included:

[0030] Step 4.1: Divide the virtual TSN queue node into two parts G' according to IPV be and G' rt ,The specific split ratio is set according to the total number of queues z. The number of queues and IPV of RTT flow are smaller than BET flow;

[0031] Step 4.2: For the logical augmented topology graph G' o The two virtual nodes connected in G' be and G' rt Fully connect the virtual TSN queue nodes contained in each of them;

[0032] Step 4.3: For virtual nodes whose physical resources are insufficient to accommodate new VNFs, in the final logical augmented topology G′ f Delete on;

[0033] Step 4.4, respectively in G' be and G' rt The shortest path algorithm can be used to obtain the full-path queue IPV setting whose RTT and BET meet the SFC constraints;

[0034] Step 4.5: Set G' be and G' rt The shortest paths obtained on the two sub-topologies are mapped to the original physical topology to obtain the final virtual TSN network queue scheduling result.

[0035] Furthermore, the GCLs of each TSN gateway are designed through a centralized SDN controller to achieve multi-queue dynamic scheduling.

[0036] Furthermore, the virtual topology can be segmented in a fine-grained manner according to data flow or application QoS requirements.

[0037] In a preferred embodiment of the present invention, in response to the defects of the prior art and the need for improvement, the present invention provides a service function chain deployment and multi-priority routing method for TSN virtual networks, which aims to solve the problem of high complexity of existing service function chain deployment problems and fill the gaps in service function chain deployment related mechanisms in virtual TSN networks, thereby better utilizing limited resources to provide deterministic communication services for future industrial needs.

[0038] Specifically, the present invention combines the characteristics of the underlying physical network and the TSN standard characteristics to propose a multi-priority service function chain deployment and multi-priority routing method for virtual TSN virtual networks, which includes two stages: ① an augmented logical topology construction method that jointly considers SFC constraints and TSN characteristics is used to reduce scheduling complexity; ② based on the augmented topology constructed in the previous stage, a differentiated topology segmentation mechanism that considers multi-priority data and TSN multi-queue properties is used to provide transmission services for data with different QoS, ultimately achieving end-to-end deterministic delivery of dynamic heterogeneous tasks.

[0039] To achieve the above objectives, the present invention adopts the following technical solutions:

[0040] 1. An augmented logic topology construction method that jointly considers SFC constraints and TSN characteristics includes the following steps:

[0041] S1, according to the SFC of arrival For each physical node n j The supported VNF set is extracted and streamlined, leaving only the VNFs that reach the SFC as the new supported VNF set for each physical node, and then VNF0 is added at the front end of the SFC.

[0042] S2. Add VNF0 to the front end of the VNF set supported by each physical node, abstract each VNF into an independent virtual node, and form a virtual node set belonging to the physical node. And do the following:

[0043] S2-1, for the same physical node n j Virtual nodes in the virtual node set If the VNFs represented by two of the nodes are adjacent VNFs in the SFC, the two nodes are connected within the range of the physical node in the SFC order to form an AC chain, and the link weight in the AC chain is 0.

[0044] S2-2. Perform S2-1 operation on the virtual node sets of all physical nodes to obtain the AC chain in the virtual node set of each physical node and the corresponding VNF highest index in the AC chain.

[0045] For virtual nodes in virtual node sets in different physical nodes, according to the highest VNF ​​index in their respective AC chains, according to the connection relationship of the underlying directed and non-fully connected topology, if two adjacent nodes n i and n j There is a directed connection between i →n j , determine node n i and n j The highest VNF ​​index of the AC chain and Whether the conditions are met: or If the condition is satisfied, the virtual nodes corresponding to the two highest indexes that meet the condition are connected according to the underlying topology; if the condition is not satisfied, the virtual nodes corresponding to the two highest indexes that meet the condition are connected according to the underlying topology; Find n i The highest VNF ​​index that meets the conditions in the AC chain Connect the virtual nodes corresponding to the two highest indexes that meet the conditions according to the underlying topology.

[0046] S2-3, according to the physical topology, perform S2-2 operation on the virtual node set of all physical nodes until all physical nodes are traversed, and then establish a preliminary logical augmented topology graph G' o .

[0047] S2-4. The initial augmented topological graph G′ obtained by the S2-3 operation o , delete the nodes that only enter but not exit or only exit but not enter except the starting and ending points, reduce the topological dimension and improve routing efficiency.

[0048] S2-5. For the node pair that does not have a virtual connection after the link is deleted, if there is a physical connection in the underlying network, re-execute S2-2 operation on the node pair to build a new link to ensure that the logical augmented topology can fully reflect the underlying topology connection relationship and the underlying device VNF support. So far, the logical augmented topology graph G' considering SFC constraints is established o .

[0049] S3. Considering the multi-queue characteristics of the TSN standard, each virtual node in the logical augmented topology graph G' is further augmented into several virtual nodes representing TSN queues according to the number of TSN queues. The specific augmentation rules are as follows:

[0050] S3-1. For virtual nodes representing TSN queues belonging to different physical nodes, if the TSN queue priorities IPVs they represent are the same, their link weights are also the same, that is, the link weights of all virtual TSN queue nodes with the same IPVs are the same.

[0051] S3-2. The resources of a physical node are evenly divided among multiple TSN queue nodes virtualized by it. If a physical node has more than one virtual node with the same queue IPV, these nodes compete for the 1 / z resources inherited by the IPV, where z is the number of IPVs, that is, the number of TSN queues.

[0052] S3-3. The link weight depends on the output queue IPv of the data and has nothing to do with the queue IPv into which it is to be injected.

[0053] S3-4. Once the data is injected into the TSN gateway, its IPV inside the gateway cannot be changed. However, the IPVs of the queues injected into the same gateway are independent of each other and do not affect each other. Thus, the final logical augmented topology G' can be obtained by jointly considering the SFC constraints and the TSN multi-queue characteristics. f .

[0054] 2. A differentiated topology segmentation mechanism that considers multi-priority data and TSN multi-queue properties, including the following parts:

[0055] S4. Considering the different QoS requirements of industrial field data, here we consider real-time traffic (RTT) and best-effort traffic (BET) with different priorities to perform differentiated topology segmentation. The specific operations are as follows:

[0056] S4-1. Divide the virtual TSN queue node into two parts G' according to IPV be and G' rt ,The specific split ratio is set according to the total number of queues z. The number of queues and IPV of RTT flow are smaller than those of BET flow.

[0057] S4-2, for the logical augmented topology graph G' o The two virtual nodes connected in G' be and G' rt The virtual TSN queue nodes contained in each of the two are fully connected.

[0058] S4-3, for virtual nodes whose physical resources are insufficient to support the new VNF, in the final logical augmented topology G' f on to delete it.

[0059] S4-4, respectively in G' beand G' rt The commonly used shortest path algorithm can be used to obtain the full-path queue IPV setting whose RTT and BET meet the SFC constraints.

[0060] S4-5, will be in G' be and G' rt The shortest paths obtained on the two sub-topologies are mapped to the original physical topology to obtain the final virtual TSN network queue scheduling result.

[0061] In another preferred embodiment of the present invention, the link cost is modeled and characterized according to the transmission cost.

[0062] Compared with the prior art, the present invention has the following obvious substantial features and significant advantages:

[0063] 1. Based on software-defined network technology and network function virtualization technology, a virtual TSN network is proposed, and the characteristics of the TSN standard are combined in the proposed network to perform data path scheduling and routing to achieve dynamic resource scheduling that meets the TSN standard.

[0064] 2. Dynamically characterize the resource changes of devices in the network, update the available information of network nodes in real time according to the remaining resources, and realize dynamic network connection and dynamic resource utilization for dynamic tasks.

[0065] 3. According to the characteristics of the underlying physical network and the content reaching the SFC, the set of VNFs supported by the device is extracted and refined. Then, according to the device capabilities, a logical routing topology that meets the SFC constraints is constructed based on the underlying physical topology, so that any routing algorithm can be applied to obtain a routing path that meets the SFC constraints.

[0066] 4. Considering the different priorities of TSN multiple queues, the logical topology is replicated, connected, and resource allocated according to the TSN queue segmentation. Data with different priorities are allocated to the topologies corresponding to different TSN queues for service, so as to realize differentiated transmission services based on the virtual TSN network.

[0067] The concept, specific structure and technical effects of the present invention will be further described below in conjunction with the accompanying drawings to fully understand the purpose, characteristics and effects of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 This is a virtual TSN network application architecture diagram of a preferred embodiment of the present invention;

[0069] Figure 2 It is a physical topology modeling and VNF extraction diagram of a preferred embodiment of the present invention;

[0070] Figure 3It is a logical routing topology construction and optimization diagram of a preferred embodiment of the present invention;

[0071] Figure 4 It is a multi-priority data topology construction diagram of a preferred embodiment of the present invention;

[0072] Figure 5 It is a preferred embodiment of the present invention that the routing path is remapped to the physical topology diagram;

[0073] Figure 6 It is a general flow chart of a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0074] The following describes several preferred embodiments of the present invention with reference to the drawings in the specification, so that the technical content is clearer and easier to understand. The present invention can be embodied in many different forms of embodiments, and the protection scope of the present invention is not limited to the embodiments mentioned in the text.

[0075] In the drawings, components with the same structure are indicated by the same numerical reference numerals, and components with similar structures or functions are indicated by similar numerical reference numerals. The size and thickness of each component shown in the drawings are arbitrarily shown, and the present invention does not limit the size and thickness of each component. In order to make the illustration clearer, the thickness of the components is appropriately exaggerated in some places in the drawings.

[0076] like Figure 1 As shown in the figure, an augmented logic topology construction method that jointly considers SFC constraints and TSN characteristics and a differentiated topology segmentation mechanism that considers multi-priority data and TSN multi-queue properties are both based on the edge computing architecture. The TSN gateway is deployed at the edge computing layer. There are various QoS applications in the industrial field, including mode control, perception information monitoring, and signal communication. The SDN controller implements physical-level scheduling of the virtual TSN network by centrally designing GCLs. Figure 1 The TSN gateway in the middle edge layer is abstracted Figure 2 The virtual nodes shown and the set of VNFs they can support.

[0077] 1. An augmented logic topology construction method that jointly considers SFC constraints and TSN characteristics includes the following steps:

[0078] S1, such as Figure 2 As shown, according to the SFC that arrives For each physical node n j The VNF set that can be supported by (j∈[2,7]) is extracted and streamlined, leaving only the VNFs that reach the SFC as the new VNF set that can be supported by each physical node, and then VNF0 is added to the front end of the SFC. At this time, the SFC

[0079] S2, such as Figure 3 As shown in the figure, VNF0 is added to the front end of the VNF set supported by each physical node, and each VNF in it is abstracted into an independent virtual node to form a virtual node set belonging to the physical node. And do the following:

[0080] S2-1, taking node n4 as an example, for the virtual nodes in the virtual node set of the same physical node n4, because and Satisfy the order of reaching SFC, that is, its order in SFC The two are adjacent VNFs, so these three virtual nodes are linked to form an AC chain. The link weight within the chain is 0, and the corresponding highest VNF ​​index in the AC chain is 4.

[0081] S2-2, perform S2-1 operation on the virtual node sets of all physical nodes to obtain the AC chain in the virtual node set of each physical node and the corresponding VNF highest index in the AC chain;

[0082] For virtual nodes in the virtual node sets in different physical nodes, according to the highest VNF ​​index in their respective AC chains, according to the connection relationship of the underlying directed and non-fully connected topology, taking nodes n4 and n6 as an example, since there is a directed connection n4→n6 between them, the highest VNF ​​index of their respective AC chains is and Judgment conditions or If it is not satisfied, then go back and find the maximum value that satisfies the condition. Last connection

[0083] S2-3. Perform S2-2 on the virtual node sets of all physical nodes according to the physical topology until all physical nodes are traversed. At this time, a preliminary logical augmented topology graph G′ is established. o like Figure 3 Shown on the left.

[0084] S2-4. The initial augmented topological graph G′ obtained by the S2-3 operation o , delete the nodes that only enter but not exit or only exit but not enter except the starting and ending points, reduce the topological dimension and improve routing efficiency.

[0085] S2-5. For the node pair that does not have a virtual connection after the link is deleted, if there is a physical connection in the underlying network, re-execute S2-2 operation on the node pair to build a new link to ensure that the logical augmented topology can fully reflect the underlying topology connection relationship and the VNF support of the underlying device. At this point, the logical augmented topology graph G′ considering the SFC constraint is updated. o like Figure 3 As shown on the right.

[0086] S3, such as Figure 5 As shown, considering the multi-queue characteristics of the TSN standard, each virtual node in the logical augmented topology graph G′ is further augmented into several virtual nodes representing the TSN queues according to the number of TSN queues. The specific augmentation rules are as follows:

[0087] S3-1. For virtual nodes representing TSN queues belonging to different physical nodes, if the TSN queue priorities IPVs they represent are the same, their link weights are also the same, that is, the link weights of all virtual TSN queue nodes with the same IPVs are the same.

[0088] S3-2. The resources of a physical node are evenly divided among multiple TSN queue nodes virtualized by it. If a physical node has more than one virtual node with the same queue IPV, these nodes compete for the 1 / z resources inherited by the IPV, where z is the number of IPVs, that is, the number of TSN queues.

[0089] S3-3. The link weight depends on the output queue IPv of the data and has nothing to do with the queue IPv into which it is to be injected.

[0090] S3-4. Once the data is injected into the TSN gateway, its IPV inside the gateway cannot be changed. However, the IPVs of the queues injected into the same gateway are independent of each other and do not affect each other. Thus, the final logical augmented topology G′ can be obtained by jointly considering the SFC constraints and the TSN multi-queue characteristics. f .

[0091] 2. A differentiated topology segmentation mechanism that considers multi-priority data and TSN multi-queue properties Figure 4 As shown, it contains the following parts:

[0092] S4. Considering the different QoS requirements of industrial field data, here we consider real-time traffic (RTT) and best-effort traffic (BET) with different priorities to perform differentiated topology segmentation. The specific operations are as follows:

[0093] S4-1. Divide the virtual TSN queue node into two parts G′ according to IPV be and G′rt ,The specific split ratio is set according to the total number of queues z. The number of queues and IPV of RTT flow are smaller than those of BET flow.

[0094] S4-2. For the logical augmented topology graph G′ o The two virtual nodes connected in G belong to different physical nodes, respectively. be and G′ rt The virtual TSN queue nodes contained in each of the two are fully connected.

[0095] S4-3. For virtual nodes whose physical resources are insufficient to support the new VNF, in the final logical augmented topology G′ f on to delete it.

[0096] S4-4, respectively in G′ be and G′ rt The commonly used shortest path algorithm can be used to obtain the full-path queue IPV setting whose RTT and BET meet the SFC constraints.

[0097] S4-5, such as Figure 5 As shown, in G′ be and G′ rt The shortest paths obtained on the two sub-topologies are mapped to the original physical topology to obtain the final virtual TSN network queue scheduling result.

[0098] Figure 6 It is a general flow chart of a preferred embodiment of the present invention.

[0099] The augmented logic topology construction method considering SFC constraints and TSN characteristics jointly includes the following steps:

[0100] S1, according to the SFC of arrival For each physical node n j The supported VNF set is extracted and streamlined, leaving only the VNFs that reach the SFC as the new supported VNF set for each physical node, and then VNF0 is added at the front end of the SFC.

[0101] S2. Adopting the augmented logical topology construction method considering SFC constraints, VNF0 is added to the front end of the VNF set supported by each physical node, and each VNF is abstracted into an independent virtual node to form a virtual node set belonging to the physical node. Traverse and execute operations such as connection within the physical node range, connection between physical node ranges, and topology simplification to obtain the initial logical augmented topology graph G′ o .

[0102] S3, using the virtual queue augmentation mechanism based on TSN multi-queue characteristics, according to the number of TSN queues, the logical augmentation topology G′ o Each virtual node in is further augmented into several virtual nodes representing TSN queues, and the augmented logical topology G′ is obtained by jointly considering SFC constraints and TSN characteristics. f .

[0103] In the augmented logical topology construction method that jointly considers SFC constraints and TSN characteristics, the set of VNFs that can be supported by each physical node in step S1 and the SFC reached are both a subset of the total set of VNFs, and the two can be the same or different.

[0104] In the augmented logical topology construction method that jointly considers SFC constraints and TSN characteristics, the embedding of VNFs in SFC in step S1 is accompanied by the consumption of computing resources and cache resources of the underlying physical nodes, and since each VNF is responsible for different functions, there are certain mapping order constraints between them.

[0105] In the augmented logical topology construction method that jointly considers SFC constraints and TSN characteristics, the purpose of adding VNF0 at the front end of the VNF set supported by each physical node in step S2 is to ensure that the original link will not disappear after the VNF is extracted, thereby avoiding changing the connection relationship of the underlying physical nodes.

[0106] The augmented logic topology construction method considering SFC constraints, step S2 specifically includes the following sub-steps:

[0107] S2-1, for the same physical node n j Virtual nodes in the virtual node set If the VNFs represented by two of the nodes are adjacent VNFs in the SFC, the two nodes are connected within the range of the physical node in the SFC order to form an AC chain, and the link weight in the AC chain is 0.

[0108] S2-2. Perform S2-1 operation on the virtual node sets of all physical nodes to obtain the AC chain in the virtual node set of each physical node and the corresponding VNF highest index in the AC chain.

[0109] S2-2. For virtual nodes in virtual node sets in different physical nodes, according to the highest VNF ​​index in their respective AC chains, according to the connection relationship of the underlying directed non-fully connected topology, if two adjacent nodes n i and n j There is a directed connection between i →n j , determine node n i and n j The highest VNF ​​index of the AC chain and Whether the conditions are met: or If the condition is satisfied, the virtual nodes corresponding to the two highest indexes that meet the condition are connected according to the underlying topology; if the condition is not satisfied, the virtual nodes corresponding to the two highest indexes that meet the condition are connected according to the underlying topology; Find n i The highest VNF ​​index that meets the conditions in the AC chain Connect the virtual nodes corresponding to the two highest indexes that meet the conditions according to the underlying topology.

[0110] S2-3. Perform S2-1 and S2-2 operations on the virtual node sets of all physical nodes according to the physical topology until all physical nodes are traversed. At this time, a preliminary logical augmented topology graph G′ is established. o .

[0111] S2-4. The initial augmented topological graph G′ obtained by the S2-3 operation o , delete the nodes that only enter but not exit or only exit but not enter except the starting and ending points, reduce the topological dimension and improve routing efficiency.

[0112] S2-5. For the node pair that does not have a virtual connection after the link is deleted, if there is a physical connection in the underlying network, re-execute S2-2 operation on the node pair to build a new link to ensure that the logical augmented topology can fully reflect the underlying topology connection relationship and the underlying device VNF support. So far, the logical augmented topology graph G′ considering SFC constraints is established. o .

[0113] Based on the virtual queue augmentation mechanism of TSN multi-queue characteristics, step S3 specifically includes the following sub-steps:

[0114] S3-1. For virtual nodes representing TSN queues belonging to different physical nodes, if the TSN queue priorities IPVs they represent are the same, their link weights are also the same, that is, the link weights of all virtual TSN queue nodes with the same IPVs are the same.

[0115] S3-2. The resources of a physical node are evenly divided among multiple TSN queue nodes virtualized by it. If a physical node has more than one virtual node with the same queue IPV, these nodes compete for the 1 / z resources inherited by the IPV, where z is the number of IPVs, that is, the number of TSN queues.

[0116] S3-3. The link weight depends on the output queue IPv of the data and has nothing to do with the queue IPv into which it is to be injected.

[0117] S3-4. Once the data is injected into the TSN gateway, its IPV inside the gateway cannot be changed. However, the IPVs of the queues injected into the same gateway are independent of each other and do not affect each other. Thus, the final logical augmented topology G′ can be obtained by jointly considering the SFC constraints and the TSN multi-queue characteristics. f .

[0118] Based on the virtual queue augmentation mechanism of TSN multi-queue characteristics, the number of IPv in step S3-2 is the same as the number of queues of the TSN gateway itself, which is mostly an integer between 1 and 8.

[0119] The differentiated topology segmentation mechanism considering multi-priority data and TSN multi-queue properties includes step S4: Considering the different QoS requirements of industrial field data, two data flows with different priorities, real-time traffic (RTT) and best-effort traffic (BET), are considered for differentiated topology segmentation. The specific operations are as follows:

[0120] S4-1. Divide the virtual TSN queue node into two parts G′ according to IPV be and G′ rt ,The specific split ratio is set according to the total number of queues z. The number of queues and IPV of RTT flow are smaller than those of BET flow.

[0121] S4-2. For the logical augmented topology graph G′ o The two virtual nodes connected in G belong to different physical nodes, respectively. be and G′ rt The virtual TSN queue nodes contained in each of the two are fully connected.

[0122] S4-3. For virtual nodes whose physical resources are insufficient to support the new VNF, in the final logical augmented topology G′ f on to delete it.

[0123] S4-4, respectively in G′ be and G′ rt The commonly used shortest path algorithm can be used to obtain the full-path queue IPV setting whose RTT and BET meet the SFC constraints.

[0124] S4-5, will be at G′ be and G′ rt The shortest paths obtained on the two sub-topologies are mapped to the original physical topology to obtain the final virtual TSN network queue scheduling result.

[0125] A differentiated topology segmentation mechanism for multi-priority data and TSN multi-queue properties is implemented by designing the GCLs of each TSN gateway through a centralized SDN controller to achieve multi-queue dynamic scheduling.

[0126] Considering the differentiated topology segmentation mechanism of multi-priority data and TSN multi-queue nature, the virtual topology can be segmented more finely based on the data flow / application QoS requirements. Here, only two QoS requirements are used as examples, and the actual QoS requirements may be more than two.

[0127] The preferred specific embodiments of the present invention are described in detail above. It should be understood that ordinary technicians in the field can make many modifications and changes based on the concept of the present invention without creative work. Therefore, all technical solutions that can be obtained by technicians in the technical field based on the concept of the present invention through logical analysis, reasoning or limited experiments on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A multi-priority service function chain deployment method for a virtual TSN network, characterized in that: The following steps are involved: Step 1: According to the SFC you have reached For each physical node n j The supported VNF set is extracted and streamlined, leaving only the VNFs that reach the SFC as the new supported VNF set for each physical node, and then VNF 0 is added to the front end of the SFC; Step 2: Adopt the augmented logical topology construction method considering SFC constraints, add VNF 0 to the front end of the VNF set supported by each physical node, abstract each VNF into an independent virtual node, and form a virtual node set belonging to the physical node. Traverse and execute the connection within the physical node range, the connection between the physical node range and the topology simplification operation to obtain the initial logical augmented topology graph G' o ; Step 3: Adopt the virtual queue augmentation mechanism based on TSN multi-queue characteristics, and augment the logical topology G' according to the number of TSN queues. o Each virtual node in is further augmented into several virtual nodes representing TSN queues, and an augmented logical topology G' is obtained that jointly considers SFC constraints and TSN characteristics. f ; The step 2 comprises the following steps: Step 2.1: For the same physical node n j Virtual nodes in the virtual node set If the VNFs represented by two of the nodes are adjacent VNFs in the SFC, the two nodes are connected within the physical node range according to the SFC order to form an AC chain, and the link weight in the AC chain is 0; Step 2.2: Execute the operation in step 2.1 for the virtual node sets of all physical nodes to obtain the AC chain in the virtual node set of each physical node and the corresponding highest VNF ​​index in the AC chain; For virtual nodes in virtual node sets in different physical nodes, according to the highest VNF ​​index in their respective AC chains, according to the connection relationship of the underlying directed non-fully connected topology, if two adjacent nodes n i and n j There is a directed connection between i →n i , determine node n i and n i The highest VNF ​​index of the AC chain and Whether the conditions are met: or If the condition is satisfied, the virtual nodes corresponding to the two highest indexes that meet the condition are connected according to the underlying topology; if the condition is not satisfied, the virtual nodes corresponding to the two highest indexes that meet the condition are connected according to the underlying topology; Find n i The highest VNF ​​index that meets the conditions in the AC chain Connect the virtual nodes corresponding to the two highest indexes that meet the conditions according to the underlying topology; Step 2.3: Execute the step 2.2 operation on the virtual node sets of all physical nodes according to the physical topology until all physical nodes are traversed to establish a preliminary logical augmented topology graph G' o ; Step 2.4: The initial augmented topological graph G' obtained by the operation in step 2.3 o , delete the nodes that only enter but not exit or only exit but not enter except the starting and ending points, to reduce the topological dimension; Step 2.5: For node pairs that do not have virtual connections after link deletion, if the underlying network has physical connections, the node pairs re-execute the step 2.2 operation to build new links, ensuring that the logical augmented topology can fully reflect the underlying topology connection relationship and the underlying device VNF support, and establish the logical augmented topology graph G' considering SFC constraints o ; The step 3 comprises the following steps: Step 3.1: For virtual nodes representing TSN queues belonging to different physical nodes, if the TSN queue priorities IPVs they represent are the same, then the link weights are also the same, that is, the link weights of all virtual TSN queue nodes with the same IPVs are the same; Step 3.2: The resources of the physical node are evenly divided by the virtualized TSN queue nodes. If a physical node has more than one virtual node with the same queue IPV, these nodes compete for the use of 1 / z resources inherited by the IPV, where z is the number of IPVs, i.e., the number of TSN queues. Step 3.3: The link weight depends on the output queue IPv of the data and has nothing to do with the queue IPv to be injected. Step 3.4: Once the data is injected into the TSN gateway, the IPV inside the gateway cannot be changed. However, the IPVs of the queues injected into the same gateway are independent of each other and do not affect each other. The final logical augmented topology G' is obtained by jointly considering the SFC constraints and the TSN multi-queue characteristics. f .

2. The multi-priority service function chain deployment method for a virtual TSN network as claimed in claim 1, characterized in that: In step 1, the VNF set that each physical node can support and the SFC reached are both a subset of the total VNF set and can be the same or different.

3. The multi-priority service function chain deployment method for a virtual TSN network as claimed in claim 1, characterized in that: The embedding of VNFs in the SFC in step 1 is accompanied by the consumption of computing resources and cache resources of the underlying physical nodes, and because each VNF is responsible for different functions, there are certain mapping order constraints between them.

4. The multi-priority service function chain deployment method for a virtual TSN network as claimed in claim 1, characterized in that: In step 2, VNF 0 is added to the front end of the VNF set supported by each physical node. After the VNF is extracted, the original link will not disappear, thereby avoiding changing the connection relationship of the underlying physical nodes.

5. The multi-priority service function chain deployment method for a virtual TSN network as claimed in claim 1, characterized in that: The number of IPv in step 3.2 is the same as the number of queues of the TSN gateway itself, which is an integer between 1 and 8.

6. The multi-priority service function chain deployment method for a virtual TSN network as claimed in claim 1, characterized in that: Industrial field data QoS requirements are different. Differentiated topology segmentation is performed on two data flows with different priorities, real-time traffic (RTT) and best-effort traffic (BET). The following steps are included: Step 4.1: Divide the virtual TSN queue node into two parts G' according to IPV be and G' rt ,The specific split ratio is set according to the total number of queues z. The number of queues and IPV of RTT flow are smaller than BET flow; Step 4.2: For the logical augmented topology graph G' o The two virtual nodes connected in G' be and G' rt Fully connect the virtual TSN queue nodes contained in each of them; Step 4.3: For virtual nodes whose physical resources are insufficient to support the new VNF, in the final logical augmented topology G' f Delete on; Step 4.4, respectively in G' br and G' rt The shortest path algorithm can be used to obtain the full-path queue IPV setting whose RTT and BET meet the SFC constraints; Step 4.5: Set G' be and G' rt The shortest paths obtained on the two sub-topologies are mapped to the original physical topology to obtain the final virtual TSN network queue scheduling result.

7. The multi-priority service function chain deployment method for a virtual TSN network as claimed in claim 6, characterized in that: The gate control lists (GCLs) of each TSN gateway are designed through a centralized SDN controller to achieve dynamic scheduling of multiple queues.

8. The multi-priority service function chain deployment method for a virtual TSN network as claimed in claim 6, characterized in that: Virtual topology can be segmented in a fine-grained manner based on data flow or application QoS requirements.

Images