CN107666412A - The virtual network function dispositions method of service function chain - Google Patents

Abstract

The invention discloses a virtual network function deployment method of a service function chain, which includes sequentially selecting undeployed VNFs in the service function chain; when the current VNF is not the last VNF, calculating the distance between each vertex in the vertex set and the source point s Reliability value; when the non-destination vertex meets the preset constraints, the reliability value from the vertex to the source point s is used to update the vertex preset reliability value; select the non-destination vertex corresponding to the largest vertex preset reliability value to deploy the current VNF; when the current VNF is the last VNF in the service function chain, calculate the first calculated value; when the vertex satisfies the set condition, use the ratio of the first calculated value of the current vertex to the attribute reliability value of the current vertex to update The set threshold of the current vertex; when all the vertices in the vertex set have been traversed, select the vertex corresponding to the largest set threshold among all vertices to deploy the last VNF, and output the deployment plan.

Description

Virtual network function deployment method for service function chain

technical field

The invention relates to a virtual network function deployment technology in a network function virtualization network topology, in particular to a virtual network function deployment method of a service function chain.

Background technique

In recent years, as user demands become more and more variable and diverse, communication service providers (TSPs) are increasingly required to find a more flexible and lower-cost way to deploy network services. In this regard, the emergence of NFV makes it possible to flexibly and effectively deploy service function chains dynamically without changing special equipment; since the European Telecommunications Standards Institute (ETSI) proposed the standard description structure of NFV, the architecture of NFV has become a hot topic of many scholars. subject of research.

Figure 1 describes a simple NFV architecture. In Figure 3, the virtualization layer includes all virtual machines, and the physical layer includes all underlying nodes. Both virtual machines and underlying nodes have certain computing, storage, Network and other resources to serve users. On top of the Network Functions Virtualization Infrastructure (NFVI) is the aforementioned network service, called a Service Function Chain (SFC), which consists of a chain of different Virtual Network Functions (VNFs) connected by virtual links , a VNF can represent an actual network function.

The reliability of NFV is the key factor and prerequisite for the successful implementation of SFC. In this regard, the current research on NFV includes: NFV deployment algorithm Guaranteeing Reliability with Enhanced Protection (GREP) on reliability assurance and framework research on NFV deployment reliability assessment Minimum Total Failure Removal (MTFR). Although the first method can guarantee the reliability of users, the backup protection mechanism it proposes will consume a lot of resources to provide and maintain backup VNFs; the second algorithm is to ensure the reliability of NFV after it is deployed. Evaluation, and deployment algorithms do not provide reliability guarantees.

In this regard, how to improve the reliability of network services to meet user needs, and how to map these user requests to the underlying network needs to be resolved urgently.

Contents of the invention

Aiming at the above-mentioned shortcomings in the prior art, the present invention provides a virtual network function deployment method of a service function chain that can improve the reliability of network services.

In order to achieve the above-mentioned purpose of the invention, the technical scheme adopted in the present invention is:

A virtual network function deployment method of a service function chain is provided, which includes:

Obtain the topology diagram of the service function chain and the underlying network;

Deploy the source and destination vertices of the service function chain on the corresponding nodes of the underlying network, and sequentially select the undeployed virtual network functions in the service function chain;

When the current virtual network function is not the last virtual network function in the service function chain, calculate the reliability value from each vertex in the vertex set to the source point of the service function chain;

Traversing the non-purpose vertices in the vertex set, when the non-purpose vertices meet the preset constraints, use the reliability value of the vertex to the source point of the service function chain to update its vertex preset reliability value;

When all non-purpose vertices in the vertex set have been traversed, select the non-purpose vertex corresponding to the largest vertex preset reliability value among all non-purpose vertices to deploy the current virtual network function, and delete the largest vertex preset in the vertex set The non-purpose vertices corresponding to the reliability value;

When the current virtual network function is the last virtual network function in the service function chain, calculate the product of the reliability value from the current vertex to the destination vertex in the vertex set and the reliability value from the current vertex to the source point of the service function chain as the current vertex The first calculated value of ;

Traversing all the vertices in the vertex set, when the vertex satisfies the set condition, the ratio of the first calculated value of the current vertex to the attribute reliability value of the current vertex is used to update the set threshold of the current vertex;

When all vertices in the vertex set have been traversed, select the vertex corresponding to the maximum set threshold among all vertices to deploy the last virtual network function, and output the vertex set including the deployment of all virtual network functions and the path of virtual link deployment The deployment scheme of forwarding nodes on the set and path.

Further, the preset constraint is that the computing capacity of the non-destination vertex is greater than or equal to the computing capacity required by the network virtual function, and the reliability value from the non-destination vertex to the source point of the service function chain is greater than the preset reliability value of the vertex.

Further, the setting condition is that the calculation capacity of the vertex is greater than or equal to the calculation capacity required by the network virtual function, and the ratio of the first calculation value to the attribute reliability value of the vertex is greater than or equal to the set threshold.

Further, the calculation method for calculating the reliability value from each vertex in the vertex set to the source point of the service function chain includes:

Get the vertex set and the current source point, and make the vertex set to be updated equal to the vertex set;

Traverse all outgoing edges of the current source point, and judge whether the remaining bandwidth resource of the current outgoing edge and the destination vertex of the outgoing edge satisfy the first preset constraint condition;

If it is satisfied, calculate the product of the reliability value from the current source point to the source point of the service function chain, the reliability value of the current out-degree edge, and the attribute reliability value of the destination vertex of the out-degree edge as the second calculation value;

When the second calculated value is greater than the reliability value from the destination vertex of the current out-degree edge to the source point of the service function chain and the reliability value expected by the user, the second calculated value is used to update the destination vertex of the current out-degree edge to the service The reliability value of the source point of the function chain;

When all out-degree edges of the current source point have been traversed, update the current source point with the vertices satisfying the second preset constraint condition among the destination vertices of all out-degree edges of the current source point, and delete the current source point in the vertex set to be updated ;

When the vertex set to be updated is a non-empty set, continue to traverse all out-degree edges of the current source point until the vertex set to be updated is an empty set, and then obtain the reliability value from each vertex in the vertex set to the source point of the service function chain .

Further, the first preset constraint condition is that the remaining bandwidth resource of the current out-degree edge is greater than or equal to the bandwidth resource required by the virtual link to be deployed on the out-degree edge and the destination vertex of the out-degree edge of the current source belongs to the vertex to be updated gather.

Further, the second preset constraint condition is that the reliability value is the largest.

Further, the second preset constraint condition is the minimum load factor.

Further, the calculation formula of the load factor is:

Among them, δ is the load factor; is the remaining computing resource capacity of vertex v _i ; is the outgoing edge of vertex v _i ; is the set of out-degree edges of vertex v _i ; is the remaining bandwidth resource of an outgoing edge of vertex v _i ; is all the bandwidth resource overhead on the path from vertex v _i to the source; V _P is the set of vertices in the underlying network.

Further, the virtual network function deployment method of the service function chain also includes an optimized deployment scheme:

Obtain the link with the smallest bandwidth in the service function chain in the deployment plan, and calculate the total number of links from the link with the smallest bandwidth to the destination vertex;

When the number of links is greater than zero, obtain all virtual network functions between the destination virtual network function of the link with the smallest bandwidth and the destination vertex of the service function chain to form a function set;

Determine whether there is a forwarding node between the destination vertex of the service function chain and the last virtual network function and between two adjacent virtual network functions in the function set;

If it exists, when the available computing resources of the forwarding node are greater than or equal to the computing resources required by the virtual network function adjacent to the destination vertex of the service function chain, and the virtual network function adjacent to the destination vertex of the service function chain moves, its deployed The remaining bandwidth resources of all links between the location of the service function chain and its previous virtual network function deployment location must be greater than or equal to the bandwidth of the virtual link between the virtual network function adjacent to the destination vertex of the service function chain and its previous virtual network function When required, deploy the virtual network function adjacent to the destination vertex of the service function chain on the forwarding node adjacent to the destination vertex of the service function chain, and reduce the number of links by one;

When the number of links is equal to zero, the optimized output includes the vertex set for deploying all virtual network functions, the path set for virtual link deployment, and the deployment scheme of forwarding nodes on the path.

The beneficial effects of the present invention are: since the present invention mainly considers nodes and links with relatively high reliability to deploy virtual functions and links during the deployment process, and the reliability metrics in the underlying network are all positive numbers less than 1, Therefore, when looking for a suitable node to deploy a virtual function, the found node will be as close as possible to the reference source node position, which will greatly save the search path time when deploying a virtual function, thereby reducing the deployment time of the entire service request and improving deployment efficiency.

By comparing a large amount of experimental data, the present invention finds that because the method of this scheme is outstanding in terms of request acceptance rate and network load balancing, in the process of dynamic arrival of requested services, the time of blocking or being rejected will be longer than other similar Algorithms came late.

When deploying virtual network functions in this solution, due to the comprehensive consideration of the load in the network, each network service request is to ensure the load balance of the underlying network as much as possible while ensuring that the reliability requirements are met during the deployment process. Therefore, the underlying network has more capacity to accommodate subsequent service requests, so the blocking rate is smaller.

During deployment, due to consideration of load balancing, the underlying network can receive more request services, so resource utilization naturally increases. With the increase of resource utilization, the operator's cost will be reduced for the same underlying network, so that the price that users need to pay for the same service will also be reduced accordingly.

Description of drawings

Fig. 1 is a flowchart of an embodiment of a method for deploying a virtual network function of a service function chain.

Fig. 2 is a flow chart of calculating the reliability value from each vertex in the vertex set to the source point of the service function chain.

Fig. 3 is a simple NFV architecture described in the background art.

Figure 4 is an example of VNF mapping.

Detailed ways

The specific embodiments of the present invention are described below so that those skilled in the art can understand the present invention, but it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art, as long as various changes Within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concept of the present invention are included in the protection list.

Referring to FIG. 1 , FIG. 1 shows a flowchart of an embodiment of a method for deploying a virtual network function of a service function chain; as shown in FIG. 1 , the method includes steps 101 to 108 .

In step 101, the topology diagram of the service function chain and the underlying network is obtained; a in Figure 4 is the service function chain, b is the underlying network topology, as shown in a in Figure 4, a service function chain SFC request includes a series A virtual network function VNF connected in series by links, a source node s, a destination node t, and a virtual link connecting these nodes and the VNF.

The present invention adopts SR=(N _S , L _S , s, t) to represent an SFC request, wherein denotes the set of network features, and | _NS | denotes the number of features in the request. Indicates the link set of SFC, and | _LS | indicates the number of links in the requested service. The subscripts s and t in the request SR represent the source node and the destination node of the request respectively, and they represent two different nodes in the underlying network, and these two nodes do not have any resource requirements.

The underlying network consists of different underlying nodes distributed in multiple geographical locations and the physical links connecting these nodes. The underlying network contains routers, switches, servers and physical links. Each underlying node has a series of service functions, these nodes contain corresponding resource attributes, and each physical link has corresponding bandwidth resource capacity, etc. In the present invention, the underlying network diagram is represented by a mathematical expression of G _P = (V _P , E _P ), wherein the underlying network model is randomly generated by WaxMan2 of the existing GT-ITM, wherein Indicates the set of underlying network nodes, |V _P |indicates the number of underlying network nodes; Indicates the set of edges (out-degree edges) in the underlying network, and |E _P | indicates the number of physical links.

In step 102, the source and destination vertices of the service function chain are deployed on the corresponding nodes of the underlying network, and the undeployed virtual network functions in the service function chain are sequentially selected; the service function chain shown in a in Figure 4, in When the service function chain is deployed for the first time, except for the source and destination vertices of the service function chain, VNF1 to VNF3 are undeployed virtual network functions.

In step 103, when the current virtual network function is not the last virtual network function in the service function chain, calculate the reliability value from each vertex in the vertex set to the source point of the service function chain.

In one embodiment of the present invention, the calculation method 200 for calculating the reliability value from each vertex in the vertex set to the source point of the service function chain includes:

In step 201, obtain the vertex set and the current source point, and make the vertex set to be updated equal to the vertex set; since the source point s and destination node t of the service function chain are deployed on the corresponding nodes of the underlying network, the virtual network function When deployed, its initial current source is the source of the service function chain.

In step 202, traverse all out-degree edges of the current source point, and judge whether the remaining bandwidth resource of the current out-degree edge and the destination vertex of the out-degree edge satisfy the first preset constraint condition; as shown in b in Figure 4, node A The outgoing edges of are AF, AE, and AB, and the nodes F, E, and B are the destination vertices of the outgoing edges AF, AE, and AB.

During implementation, the preferred first preset constraint condition of this solution is that the remaining bandwidth resource of the current out-degree edge is greater than or equal to the bandwidth resource required by the virtual link to be deployed on the out-degree edge and the destination vertex of the out-degree edge of the current source point belongs to The set of vertices to be updated.

In step 203, if it is satisfied, the product of the reliability value of the current source point to the source point of the service function chain, the reliability value of the current out-degree edge, and the attribute reliability value of the destination vertex of the out-degree edge is calculated as the first 2. Calculated value.

When the current source point is the source point of the service function chain, the reliability value from the current source point to the source point of the service function chain is the attribute reliability value of the source point of the service function chain itself, which is inherent in each node A reliability value of , which will not change as the network is updated.

In step 204, when the second calculated value is greater than the reliability value from the destination vertex of the current out-degree edge to the source point of the service function chain and the reliability value expected by the user, update the current out-degree edge with the second calculated value The reliability value from the destination vertex to the source point of the service function chain.

When the second calculated value does not meet the above conditions, the reliability value from the destination vertex of the corresponding outgoing edge to the source point of the service function chain is not updated.

In step 205, when all out-degree edges of the current source point have been traversed, update the current source point with the vertices satisfying the second preset constraint condition among the destination vertices of all out-degree edges of the current source point, and delete the set of vertices to be updated The current source point in .

During implementation, the second preset constraint condition of this solution may be the maximum reliability value, or the minimum load factor.

The calculation formula for the load factor introduced during the deployment of virtual network functions is:

Among them, δ is the load factor; is the remaining computing resource capacity of vertex v _i ; is the outgoing edge of vertex v _i ; is the set of out-degree edges of vertex v _i ; is the remaining bandwidth resource of an outgoing edge of vertex v _i ; is all the bandwidth resource overhead on the path from vertex v _i to the source; V _P is the set of vertices in the underlying network.

The effect of this scheme on the two situations of the second preset constraint condition is described below:

Assume that when deploying a service function chain, the reliability required by the user is 0.90. Referring to diagram b in Figure 4, if the vertex with the highest reliability value among the vertices is selected to update the current source point in steps 101 to 108, it will eventually form The deployment path of the service function link is the service function forwarding path 2 (AECH) in b in Figure 4. The reliability of this deployment scheme is 0.97, and the resource overhead is 232.

If in step 101 to step 108, the vertex with the smallest load factor among the destination vertices of all outgoing edges of the current source point is selected to update the current source point, the deployment path of the service function link finally formed is Service function forwarding path 1 (AFGEDH), the reliability of this deployment scheme is 0.94, and the resource overhead is 202.

It can be seen from this that using the vertex with the highest reliability value among the vertices to update the current source point to deploy the virtual network function will make the network smoother for users and provide a better user experience. For network operators TSPs In other words, it will cost higher CAPEX (capital expenditure) and OPEX (operating expenditure).

Use the vertex with the smallest load factor among the destination vertices of all out-degree edges of the current source point to update the current source point. This method is to introduce the load factor under the premise of satisfying the reliability value expected by the user. The minimum load factor can make each vertex The load of the network is more balanced, and the reliability of network services can be appropriately reduced under the premise of ensuring user reliability requirements, which will make the underlying network load lighter, and the underlying network will be able to handle more requests, which will greatly reduce network operators CAPEX (capital expenditure) and OPEX (operating expenditure) of TSPs to improve resource utilization.

In step 206, when the vertex set to be updated is a non-empty set, continue to traverse all the outgoing edges of the current source point until the vertex set to be updated is an empty set, and then obtain each vertex in the vertex set (except the service function chain) Except for the source point, each other vertex is the destination vertex of an out-degree edge. For details, refer to the reliability value of the source point of the service function chain in diagram b in Figure 4).

In step 104, the non-destination vertices in the vertex set are traversed, and when the non-destination vertices satisfy the preset constraints, the reliability values of the vertices to the source point of the service function chain are used to update the preset reliability values of the vertices.

Each vertex in this solution will have a vertex preset reliability value, and the vertex preset reliability value of each vertex may be different, and when this solution deploys each virtual network function, the vertex preset reliability value property values will be initialized.

During implementation, the preferred preset constraint of this scheme is that the computing capacity of the non-destination vertex is greater than or equal to the computing capacity required by the network virtual function, and the reliability value of the non-destination vertex to the source point of the service function chain is greater than the preset reliability of the vertex value.

In step 105, when all non-purpose vertices in the vertex set have been traversed, select the non-purpose vertex corresponding to the largest vertex preset reliability value among all non-purpose vertices to deploy the current virtual network function, and delete the vertex in the vertex set The non-destination vertex corresponding to the maximum vertex preset reliability value.

In step 106, when the current virtual network function is the last virtual network function in the service function chain, calculate the reliability value from the current vertex to the destination vertex in the vertex set and the reliability value from the current vertex to the source point of the service function chain The product of is used as the first computed value for the current vertex.

The reliability value from the current vertex to the destination vertex in the vertex set here is the product of all outgoing edges on the link between the current vertex and the destination vertex and the attribute reliability values of all vertices on the link.

In step 107, all vertices in the vertex set are traversed, and when the vertex satisfies the set condition, the ratio of the first calculated value of the current vertex to the attribute reliability value of the current vertex is used to update the set threshold of the current vertex.

The setting threshold here is roughly the same as the setting method of the vertex preset reliability value mentioned in step 104, that is, each vertex in this scheme will have a setting threshold, and the vertex setting threshold of each vertex may be There are differences, and in this solution, when each virtual network function is deployed, the vertex setting threshold will be initialized; however, the setting threshold and the initial value of the vertex preset reliability value can be set to the same parameters.

During implementation, this solution preferably sets the condition that the calculation capacity of the vertex is greater than or equal to the calculation capacity required by the network virtual function, and the ratio of the first calculation value to the attribute reliability value of the vertex is greater than or equal to the set threshold.

In step 108, when all vertices in the vertex set have been traversed, select the vertex corresponding to the maximum set threshold among all vertices to deploy the last virtual network function, and output the vertex set including deploying all virtual network functions, virtual The path collection of link deployment and the deployment scheme of forwarding nodes on the path.

The network deployment from step 101 to step 108 of this plan takes into account the reliability of network services and the CAPEX (capital expenditure) and OPEX (operational expenditure) of network operators' TSPs. However, in the formed deployment plan, some virtual chains There are some forwarding nodes on the road, and the existence of forwarding nodes will also affect the CAPEX (capital expenditure) and OPEX (operational expenditure) of network operators TSPs.

In order to further reduce the CAPEX (capital expenditure) and OPEX (operational expenditure) of the network operator TSPs, during implementation, this solution optimizes the virtual network function deployment method of the service function chain and also includes optimizing the deployment scheme formed in steps 101 to 108, Its specific optimization method 300 includes:

In step 301, obtain the link with the smallest bandwidth in the service function chain in the deployment plan, and calculate the total number of links from the link with the smallest bandwidth to the destination vertex;

In step 302, when the number of links is greater than zero, obtain a function set composed of all virtual network functions between the destination virtual network function of the link with the smallest bandwidth and the destination vertex of the service function chain;

In step 303, it is determined whether there is a forwarding node between the destination vertex of the service function chain and the last virtual network function and between two adjacent virtual network functions in the function set;

In step 304, if there is, then when the available computing resources of the forwarding node are greater than or equal to the required computing resources of the virtual network function adjacent to the destination vertex of the service function chain, and the virtual network function adjacent to the destination vertex of the service function chain moves After that, the remaining bandwidth resources of all links between the deployed location and the previous virtual network function deployment location must be greater than or equal to the distance between the virtual network function adjacent to the destination vertex of the service function chain and the previous virtual network function. When the bandwidth of the virtual link is required, the virtual network function adjacent to the destination vertex of the service function chain is deployed on the forwarding node adjacent to the destination vertex of the service function chain, and the number of links is reduced by one.

In step 305, when the number of links is equal to zero, output the optimized deployment plan including the vertex set for deploying all virtual network functions, the path set for virtual link deployment, and the forwarding nodes on the path.

Referring to Figure 4 below, combined with specific examples, the method for optimizing the deployment scheme is described in detail:

As shown in figure a of Figure 4, suppose the bandwidth on the virtual link from VNF1 to VNF2 is 5, the bandwidth on the virtual link from VNF2 to VNF3 is 15, and the bandwidth on the virtual link from VNF3 to destination t is 20 , the deployment scheme formed by adopting the deployment method from step 101 to step 108 is the service function forwarding path 1 (AFGEDH) in Figure 4b, VNF1 is deployed on node F, VNF2 is deployed on node G, and VNF3 is deployed on node E .

From the bandwidth on the virtual link, it can be concluded that the bandwidth on the virtual link from VNF1 to VNF2 is the smallest, and the link with the smallest bandwidth to the destination vertex contains a total of virtual links from VNF1 to VNF2, VNF2 to VNF3, and VNF3 to the destination vertex t , the number of links is 3.

All virtual network functions between the destination virtual network function of the link with the smallest bandwidth and the destination vertex of the service function chain form the function set function set VNF2 and VNF3. Step 303 is used to find that there is a forwarding node D between VNF3 and the destination vertex t. At this time, it is necessary to judge whether the available computing resources of forwarding node D are greater than or equal to the computing resources required by VNF3, and after VNF3 moves, the deployed Whether the remaining bandwidth resources of all links (GED) between location D and its previous virtual network function VNF2 deployment location are greater than or equal to the bandwidth requirements of the virtual link between VNF3 and its previous VNF2, if the above two conditions are equal If it is satisfied, deploy VNF3 on node D, and reduce the number of links by one.

For whether there is a forwarding node between two adjacent virtual network functions in the function set, the virtual network function is redeployed in the same way as above, or the last virtual network function can be deployed first, and then other virtual network functions are deployed.

Assuming that all forwarding nodes meet the two conditions in 304, then the final optimized deployment plan through steps 301 to 305 is: VNF1 is deployed on node F, VNF2 is deployed on node E, and VNF3 is deployed on node D.

After optimizing the deployment scheme in this way, the instance can save 15 bandwidths, which greatly reduces the CAPEX (capital expenditure) and OPEX (operational expenditure) of network operators TSPs.

The following describes some application scenarios and effects of the virtual network function deployment method of the service function chain involved in this solution:

The deployment method designed in this scheme can be deployed in the SDN network to realize the reliable deployment of network service requests and save the resource cost of the network operator. It separates the control function from the network switching device and moves it into a logically independent The control environment of the network - in the network control system, in which the SDN network transmits messages based on the OpenFlow protocol. The network control system can run on a common server, and any user can program the control function directly at any time. Therefore, the control function is no longer limited to the router, nor is it limited to only the manufacturer of the equipment can be programmed and defined.

After the deployment method designed in this scheme is deployed in the SDN network, SDN helps to realize the virtualization of the network, thereby realizing the integration of computing and storage resources of the network, and finally enables the realization of Control and management of the entire network. This is one of the many advantages of the SDN network, and it is also a key factor in determining whether it can be used to migrate multiple virtual machines between multiple data centers.

Network operators can also deploy the virtual network function deployment method provided by this solution on the control layer of the SDN control router. The SDN control router can schedule its own control and management functions to collect the information of the entire underlying network and obtain information about the network. The situation of all nodes and link resources, as well as the connection topology between nodes, through this centralized control mode, the router can obtain the supporting topology and corresponding resource information of the entire network.

When multiple network service requests from users arrive at the same or different times, the SDN controller can schedule the resources deployed on the controller based on the network-wide information it has mastered, and use an effective reliability-aware deployment algorithm to calculate Key parameters such as the deployment time required for deployment, the response time of service requests in the network, the resource usage of service requests in the entire network, and the acceptance rate of service requests are fed back to network operators.

To sum up, the virtual network function deployment method of the service function chain provided by this solution has the advantages of low deployment blocking rate, high resource utilization rate, high deployment efficiency, late blocking time, and wide application scenarios.

Claims (9)

1. A virtual network function deployment method of a service function chain, characterized in that, comprising: Obtain the topology diagram of the service function chain and the underlying network; Deploy the source and destination vertices of the service function chain on the corresponding nodes of the underlying network, and sequentially select the undeployed virtual network functions in the service function chain; When the current virtual network function is not the last virtual network function in the service function chain, calculate the reliability value from each vertex in the vertex set to the source point of the service function chain; Traversing the non-purpose vertices in the vertex set, when the non-purpose vertices meet the preset constraints, use the reliability value of the vertex to the source point of the service function chain to update its vertex preset reliability value; When all non-purpose vertices in the vertex set have been traversed, select the non-purpose vertex corresponding to the largest vertex preset reliability value among all non-purpose vertices to deploy the current virtual network function, and delete the largest vertex preset in the vertex set The non-purpose vertices corresponding to the reliability value; When the current virtual network function is the last virtual network function in the service function chain, calculate the product of the reliability value from the current vertex to the destination vertex in the vertex set and the reliability value from the current vertex to the source point of the service function chain as the current vertex The first calculated value of ; Traversing all the vertices in the vertex set, when the vertex satisfies the set condition, the ratio of the first calculated value of the current vertex to the attribute reliability value of the current vertex is used to update the set threshold of the current vertex; When all vertices in the vertex set have been traversed, select the vertex corresponding to the maximum set threshold among all vertices to deploy the last virtual network function, and output the vertex set including the deployment of all virtual network functions and the path of virtual link deployment The deployment scheme of forwarding nodes on the set and path. 2. The virtual network function deployment method of a service function chain according to claim 1, wherein the preset constraint is that the computing capacity of the non-purpose vertices is greater than or equal to the computing capacity required by the network virtual function, and the non-purpose vertices The reliability value of the vertex to the source of the service function chain is greater than its vertex preset reliability value. 3. The virtual network function deployment method of a service function chain according to claim 1, wherein the setting condition is that the calculation capacity of the vertex is greater than or equal to the calculation capacity required by the network virtual function, and the first calculation value The ratio to the attribute reliability value of the vertex is greater than or equal to the set threshold. 4. The virtual network function deployment method of the service function chain according to claim 1, wherein the calculation method for calculating the reliability value from each vertex in the vertex set to the source point of the service function chain comprises: Get the vertex set and the current source point, and make the vertex set to be updated equal to the vertex set; Traverse all outgoing edges of the current source point, and judge whether the remaining bandwidth resource of the current outgoing edge and the destination vertex of the outgoing edge satisfy the first preset constraint condition; If it is satisfied, calculate the product of the reliability value from the current source point to the source point of the service function chain, the reliability value of the current out-degree edge, and the attribute reliability value of the destination vertex of the out-degree edge as the second calculation value; When the second calculation value is greater than the reliability value from the destination vertex of the current out-degree edge to the source point of the service function chain and the reliability value expected by the user, update the destination vertex of the current out-degree edge with the second calculation value The reliability value to the source point of the service function chain; When all out-degree edges of the current source point have been traversed, update the current source point with the vertices satisfying the second preset constraint condition among the destination vertices of all out-degree edges of the current source point, and delete the current source point in the vertex set to be updated ; When the vertex set to be updated is a non-empty set, continue to traverse all out-degree edges of the current source point until the vertex set to be updated is an empty set, and then obtain the reliability value from each vertex in the vertex set to the source point of the service function chain . 5. The virtual network function deployment method of a service function chain according to claim 4, wherein the first preset constraint condition is that the remaining bandwidth resources of the current outbound edge are greater than or equal to those to be deployed on the outbound edge The bandwidth resource required by the virtual link of , and the destination vertex of the out-degree edge of the current source point belong to the set of vertices to be updated. 6 . The virtual network function deployment method of a service function chain according to claim 4 , wherein the second preset constraint condition is a maximum reliability value. 7 . 7. The method for deploying a virtual network function of a service function chain according to claim 4, wherein the second preset constraint condition is a minimum load factor. 8. The virtual network function deployment method of the service function chain according to claim 7, wherein the calculation formula of the load factor is: <mrow><mi>&amp;delta;</mi><mo>=</mo><mfrac><mn>1</mn><msubsup><mi>w</mi><msub><mi>v</mi><mi>i</mi></msub><mi>r</mi></msubsup></mfrac><mo>+</mo><munder><mi>&amp;Sigma;</mi><mrow><msubsup><mi>e</mi><msub><mi>v</mi><mi>i</mi></msub><mi>o</mi></msubsup><mo>&amp;Element;</mo><msubsup><mi>e</mi><mi>i</mi><mi>o</mi></msubsup></mrow></munder><mfrac><mn>1</mn><msubsup><mi>m</mi><msubsup><mi>e</mi><msub><mi>v</mi><mi>i</mi></msub><mi>o</mi></msubsup><mi>r</mi></msubsup></mfrac><mo>+</mo><msubsup><mi>m</mi><msub><mi>v</mi><mi>i</mi></msub><msub><mi>v</mi><mrow><mi>s</mi><mi>o</mi></mrow></msub></msubsup><mo>,</mo><mo>&amp;ForAll;</mo><msub><mi>v</mi><mi>i</mi></msub><mo>&amp;Element;</mo><msub><mi>V</mi><mi>P</mi></msub></mrow> Among them, δ is the load factor; is the remaining computing resource capacity of vertex v _i ; is the outgoing edge of vertex v _i ; is the set of out-degree edges of vertex v _i ; is the remaining bandwidth resource of an outgoing edge of vertex v _i ; is all the bandwidth resource overhead on the path from vertex v _i to the source; V _P is the set of vertices in the underlying network. 9. The virtual network function deployment method of the service function chain according to claim 1, further comprising optimizing the deployment scheme: Obtain the link with the smallest bandwidth in the service function chain in the deployment plan, and calculate the total number of links from the link with the smallest bandwidth to the destination vertex; When the number of links is greater than zero, obtain all virtual network functions between the destination virtual network function of the link with the smallest bandwidth and the destination vertex of the service function chain to form a function set; Determine whether there is a forwarding node between the destination vertex of the service function chain and the last virtual network function and between two adjacent virtual network functions in the function set; If it exists, when the available computing resources of the forwarding node are greater than or equal to the computing resources required by the virtual network function adjacent to the destination vertex of the service function chain, and the virtual network function adjacent to the destination vertex of the service function chain moves, its deployed The remaining bandwidth resources of all links between the location of the service function chain and its previous virtual network function deployment location must be greater than or equal to the bandwidth of the virtual link between the virtual network function adjacent to the destination vertex of the service function chain and its previous virtual network function When required, deploy the virtual network function adjacent to the destination vertex of the service function chain on the forwarding node adjacent to the destination vertex of the service function chain, and reduce the number of links by one; When the number of links is equal to zero, the optimized output includes the vertex set for deploying all virtual network functions, the path set for virtual link deployment, and the deployment scheme of forwarding nodes on the path.