CN112311665A - Single-link fault routing protection method based on hybrid SDN network - Google Patents

Abstract

The invention discloses a single-link fault route protection method based on a hybrid SDN network, belongs to the technical field of Internet, and solves the problem that LFC in the traditional network cannot cope with all possible single-link fault situations in the network. Firstly, calculating all source and destination node pairs which are not protected by an LFC rule; and then, deploying SDN nodes in the network to protect all the unprotected source-destination pairs, so that the method can cope with all single fault situations which can occur in the network. The method provided by the invention can not only cope with all single link failures in the hybrid SDN network, but also has smaller running time and path stretching degree. Therefore, the invention is an efficient route protection method which can be deployed in a hybrid SDN network.

Description

Single-link fault routing protection method based on hybrid SDN network

Technical Field

The invention relates to the technical field of route protection, in particular to a single-link fault route protection method based on a hybrid SDN network.

Background

SDN is a new network architecture that has recently emerged, and the greatest feature of the architecture is to decouple the functions of the control plane and the forwarding plane. The control plane is composed of one or more SDN centralized controllers and is responsible for path selection and routing decision. The forwarding plane is formed by SDN switches and is responsible for forwarding data flows in the network. The control plane and the forwarding plane communicate through a standardized interface Openflow. The SDN centralized controller has a logic view of the whole network and controls forwarding paths of all data flows in the network. And the controller issues the routing decision information to the SDN switch through Openflow, so that the data flow is correctly forwarded. SDN networks have many advantages over traditional networks, such as flexible control of network traffic, ease of network management and enforcement of security policies, and the like. Therefore, if the SDN technology is deployed in the network, the performance of the network will be greatly improved. But it would be an impossible thing if all the network devices at present were upgraded to SDN devices. This is because, on one hand, deploying SDN faces huge economic overhead such as manpower and material resources, and on the other hand, deploying SDN may cause network disruption, which affects user experience. It is common in academia and industry to upgrade legacy networks with partially deployed SDN technology, gradually replacing legacy devices in the network. We refer to a network in which legacy devices and SDN devices coexist as a hybrid SDN network. A hybrid SDN network primarily includes SDN controllers, SDN switches, and legacy routers. The SDN controller exchanges information with the SDN switch through an Openflow protocol. The SDN switch can work in two modes, and can interact information with a traditional router through a routing protocol, and can also interact information with the SDN switch and an SDN controller through the SDN protocol. However, the conventional router only supports the conventional routing protocol and cannot exchange information with the SDN controller.

A number of studies on network failures have shown that 70% of failures in a network are single link failures and the remaining 30% are single node failures and concurrent failures. Routing protection schemes are commonly employed by the academia and industry to cope with frequent failures in the network. Equal-Cost multi-path Routing (ECMP) is the earliest and simplest route protection scheme adopted in the industry, but research proves that the scheme cannot provide a high failure protection rate. Aiming at the problems of ECMP, IETF proposes a frame for fast rerouting, and on the basis of the frame, an LFC, a Not-Via-based route protection scheme, a tunnel-based route protection scheme and the like are proposed. In all route protection schemes, LFCs are of close interest to the industry for their simplicity and are deployed and supported by vendors of watson and watson routers. Although LFCs are simple and easy to deploy, LFCs have a fatal drawback in that they cannot protect all possible single-link failure situations in the network. In order to overcome the problems of the LFC, the authors analyze the problem of LFC fault coverage by using theoretical knowledge of graph theory, and increase the fault protection rate of the LFC by adjusting the weight of the links in the network (LFAOP, LFA Cost Optimization), but this method does not necessarily guarantee to cope with all single-link fault situations. To this end, the authors theoretically analyze the relationship between the fault protection rate of the LFC and the network topology in detail, and increase the single fault protection scenario of the LFC by adding links. However, these schemes are developed based on the conventional network architecture and the SDN network architecture, and cannot be directly applied to the hybrid SDN network.

Disclosure of Invention

The technical scheme adopted by the invention for solving the technical problems is as follows: a single link fault route protection method based on a hybrid SDN network is constructed, and the method comprises the following steps:

step 1: calculating a set L { (s, d), wherein s, d ∈ V }, of all source and destination node pairs which are not protected by the LFC rule in the network, calculating a fault protection rate, if the fault protection rate is less than 1, executing the step 2, and otherwise, ending the step;

step 2: for any source-destination node pair in the set (s, D) ∈ L, calculating all SDN nodes D (s, D) between each node pair;

and step 3: for any node i belongs to V in the network, counting the number of SDN nodes taking the node as a source destination node pair L { (s, d), s, d belongs to V }

Wherein y (i, s, d) represents whether the node i is an SDN node of the source-destination node pair s and d, if the node i is an SDN node of the source-destination node pair s and d, the value is 1, otherwise, the value is 0;

and 4, step 4: creating a queue Q having a structure of

All nodes i and corresponding

Join to queueIn Q;

and 5: setting an initial value of a deployment SDN node set as an empty set M ═ phi;

step 6: judging whether the fault protection rate is equal to 1, if not, executing the step 8, otherwise, ending;

and 7: if queue Q is empty, then end, otherwise choose one

Deploying the SDN node by the node m with the maximum value, and deleting the SDN node from the queue Q;

and 8: if the M is true, ending, otherwise, adding the node into the set M, wherein the M is { M }, U;

and step 9: for any source-destination node pair in the set (s, D) ∈ L, if m ∈ D (s, D), then the SDN node between the source-destination pair is determined, and no SDN node needs to be calculated for the SDN node, and the SDN node is calculated

Is set to 0, the value of the element in Q is updated;

step 10: updating

The value of (d);

step 11: and (4) calculating the fault protection rate, and circularly executing the steps 6-11 until the fault protection rate is 1.

The method for calculating the fault protection rate in the step 1 comprises the following steps:

where V is the set of nodes in the topology,

the method for calculating all SDN nodes between the node pairs (s, d) in step 2 is as follows: node i is the condition that the SDN node of the source-destination node pair (s, d) has to satisfy, i.e. the link (s, dn (s, d)) is not on the shortest path from node s to node i, and the shortest path from at least one neighboring node to node d does not include the link (s, dn (s, d)). Can be formally expressed as:

x ((s, dn (s, d)), s, i) ═ 0 indicates that the link (s, dn (s, d)) is not on the shortest path from node s to node i,

indicates that node i has at least one neighbor node to node d's shortest path excluding link (s, dn (s, d)), for

N (i) represents all neighbor nodes of the node, and dn (s, d) is the optimal next hop from node s to node d.

Compared with the prior art, the invention has the following advantages: the method aims at the problem that existing route protection methods are developed on the basis of a traditional network architecture or an SDN network architecture, and the methods cannot directly solve the problem of single link failure in a hybrid SDN network. The invention provides a single fault route protection method based on a hybrid SDN network, which can not only cope with all possible single fault situations in the hybrid SDN network, but also has lower cost and lower path stretching degree. Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

Drawings

The invention will be further described with reference to the accompanying drawings and examples, in which:

fig. 1 is a schematic flowchart of a single-fault route protection method based on a hybrid SDN network according to the present invention.

Fig. 2 is a schematic diagram of a network topology structure of a single-fault route protection method based on a hybrid SDN network according to the present invention.

Detailed Description

For a more clear understanding of the technical features, objects and effects of the present invention, embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

Referring to fig. 1, the present invention provides a single-fault route protection method based on a hybrid SDN network, including:

step 1: calculating a set L { (s, d), wherein s, d ∈ V }, of all source and destination node pairs which are not protected by the LFC rule in the network, calculating a fault protection rate, if the fault protection rate is less than 1, executing the step 2, and otherwise, ending the step; the method for calculating the fault protection rate is as follows

Where V is the set of nodes in the topology,

step 2: according to the set L { (s, D), s, D ∈ V } calculated in step 1, for any source-destination node pair in the set (s, D) ∈ L, all SDN nodes D (s, D) between each node pair are calculated, and the method of calculating all SDN nodes between the node pairs (s, D) is as follows: the condition that node i is the SDN node of the source-destination node pair (s, d) must satisfy, i.e. the link (s, dn (s, d)) is not on the shortest path from node s to node i, and the shortest path from node i to node d with at least one neighbor node does not include the link (s, dn (s, d)). Can be formally expressed as:

x ((s, dn (s, d)), s, i) ═ 0 denotes the link(s, dn (s, d)) is not on the shortest path from node s to node i,

indicates that node i has at least one neighbor node to node d's shortest path excluding link (s, dn (s, d)), for

N (i) represents all neighbor nodes of the node, and dn (s, d) is the optimal next hop from the node s to the node d;

and step 3: for any node i belongs to V in the network, counting the times of SDN nodes of a node pair L { (s, d), s, d belongs to V } which can be used as a source destination

Wherein y (i, s, d) represents whether the node i is an SDN node of the source-destination node pair s and d, and the value is 1 if the node i is an SDN node of the source-destination node pair s and d, and is 0 otherwise;

and 4, step 4: creating a queue Q having a structure of

All nodes i and corresponding

Adding the queue Q;

and 5: setting an initial value of a deployment SDN node set as an empty set M ═ phi;

step 6: judging whether the fault protection rate is equal to 1, if not, executing the step 8, otherwise, ending;

and 7: if Q is null, then end, otherwise choose one

The node m with the maximum value deploys the SDN technology and deletes the SDN technology from the queue Q;

and 8: if M ═ V is true, ending, otherwise, adding the node into the set M, where M ═ M }. U.M;

and step 9: for any source-destination node pair in the set (s, D) ∈ L, if m ∈ D (s, D), then the SDN node between the source-destination pair is determined, and no SDN node needs to be calculated for the SDN node, and the SDN node is calculated

Is set to 0, the value of the element in Q is updated;

step 10: updating

The value of (d);

step 11: and (4) calculating the fault protection rate, and circularly executing the steps 6-11 until the fault protection rate is 1.

Currently, the intra-domain routing protocols deployed by the internet are mainly link-state routing protocols, such as IS-IS and OSPF. In both routing protocols, all routers in the network have a complete topology within the autonomous domain. When the network is in a steady state, the topology stored in all routers is consistent. Each router in the network calculates a Shortest Path Tree (SPT) taking itself as a root by using a Shortest Path First algorithm (SPF) according to the network topology structure, and then constructs a routing table by using the Tree. According to the above description, the current intra-domain routing protocol uses the shortest path to forward the packet, but when the default next hop from the source node to the destination node fails, the packet transmitted to the node will be lost, which will cause network interruption, and greatly reduce the user experience.

As can be seen from the above description, the shortest path is used for forwarding packets in the internet at present, which may cause network interruption when a network fails, and may seriously affect the network performance. Therefore, in order to improve network availability and user experience, The Internet Engineering Task Force (IETF) proposes to cope with frequently occurring single link failures in a network by using LFC rules. The LFC rules will be described below:

LFC: for any destination address d, node c may send a message to its neighbor node x if and only if cost (x, d) < cost (x, c) + cost (c, d) is satisfied. The rule may be understood that when a link between node c and the optimal next hop of destination address d fails, node c may send a message to its neighbor node x because node c is not on the shortest path from node x to destination address d, where cost (x, d) represents the minimum cost from node x to node d in the network.

In the specific implementation process of the method of the invention,

step 1, according to the LFC rule, calculating a set L { (0,1), (0,2), (1,0), (1,3) } of all source-destination node pairs in the network which are not protected by the LFC rule, where a failure protection rate is 8/12 { (66.67%), and since the failure protection rate is less than 1, executing step 2;

step 2: calculating a set L { (0,1), (0,2), (1,0), (1,3) } calculated according to step 1, calculating all SDN nodes D (s, D) between each node pair for any source destination node pair in the set (s, D) ∈ L, D (0,1) { (2 }, D (0,2) { (1 }, D (1,0) = {3}, and D (1,3) { (0,2 };

and step 3: for any node i e V in the network, counting the times of SDN nodes of which the node can be a source destination node pair L { (0,1), (0,2), (1,0), (1,3) }

And 4, step 4: creating a queue Q having a structure of

All nodes i and corresponding

Adding the element into a queue Q, wherein the element in the queue Q is ((0,1), (1,1) (2,2) (3, 2));

and 5: setting an initial value of a deployment SDN node set as an empty set M ═ phi;

step 6: step 7 is executed because the fault protection rate is less than 1;

and 7: because the number of SDN nodes that can be the source-destination node pair L { (0,1), (0,2), (1,0), (1,3) } has a value of 2, and the number of 2 is smaller than the number of 3, m ═ 2, where the elements in Q are ((0,1), (1,1), (3, 2));

and 8: node 2 is added to set M, where M is {2 };

and step 9: because 2 ∈ D (0,1), the SDN node between the source-destination pair (0,1) is determined, and Σ y (i,0,1), i ∈ V is set to 0, and similarly, because 2 ∈ D (1,3), the SDN node between the source-destination pair (1,3) is determined, and Σ y (i,1,3), i ∈ V;

step 10: updating

At this time

The element in Q is ((0,0), (1,1), (3, 2));

step 11: since the SDN node between the source-destination pair (0,1) and (1,3) is determined, the fault protection rate at this time is 10/12-83.33%, execute step 6;

step 6: step 7 is executed because the fault protection rate is less than 1;

and 7: since node 3 may be the SDN node of the source-destination node pair L { (0,1), (0,2), (1,0), (1,3) } has a degree value of 2, m ═ 3, where the elements in Q are ((0,0), (1, 1));

and 8: node 3 is added to set M, where M is {2,3 };

and step 9: because 3 ∈ D (0,2), the SDN node between the source-destination pair (0,2) is determined, and Σ y (i,0,2), i ∈ V is set to 0, and similarly, because 2 ∈ D (1,0), the SDN node between the source-destination pair (1,0) is determined, and Σ y (i,1,0), i ∈ V;

step 10: updating

At this time

The element in Q is ((0,0), (0, 0));

step 11: since the SDN node between the source-destination pair (0,2) and (1,0) is determined, and the fault protection rate at this time is 12/12 ═ 1, step 6 is performed;

step 6: this is ended because the fail-safe rate is equal to 1.

While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (3)

1. A single link fault route protection method based on a hybrid SDN network comprises the following steps:

step 1: calculating a set L { (s, d), s, d ∈ V } of all source and destination node pairs which are not protected by the LFC rule in the network, calculating a fault protection rate, if the fault protection rate is less than 1, executing the step 2, otherwise, ending;

step 2: for any source-destination node pair in the set (s, D) ∈ L, calculating all SDN nodes D (s, D) between each node pair;

and step 3: for any node i belongs to V in the network, counting the number of SDN nodes taking the node as a source destination node pair L { (s, d), s, d belongs to V }

Whereiny (i, s, d) represents whether the node i is an SDN node of the source-destination node pair s and d, if the node i is an SDN node of the source-destination node pair s and d, the value is 1, otherwise, the value is 0;

and 4, step 4: creating a queue Q having a structure of

All nodes i and corresponding nodes

Adding the queue Q;

and 5: setting an initial value of a deployment SDN node set as an empty set M ═ phi;

step 6: judging whether the fault protection rate is equal to 1, if not, executing the step 8, otherwise, ending;

and 7: if queue Q is empty, then end, otherwise choose one

Deploying the SDN node by the node m with the maximum value, and deleting the SDN node from the queue Q;

and 8: if the M is true, ending, otherwise, adding the node into the set M, wherein the M is { M }, U;

and step 9: for any source-destination node pair in the set (s, D) ∈ L, if m ∈ D (s, D), then the SDN nodes between the source-destination pair are determined, no longer need to compute SDN nodes for it, and will be

Is set to 0, the value of the element in Q is updated;

step 10: updating

The value of (d);

step 11: and (4) calculating the fault protection rate, and circularly executing the steps 6-11 until the fault protection rate is 1.

2. The method of claim 1, wherein the method comprises: the method for calculating the fault protection rate in the step 1 comprises the following steps:

where V is the set of nodes in the topology,

3. the method of claim 1, wherein the method comprises: the method for calculating all SDN nodes between node pairs (s, d) in step 2 comprises the following steps: node i is a condition that the SDN node of the source-destination node pair (s, d) must satisfy, i.e. the link (s, dn (s, d)) is not on the shortest path from node s to node i, and the shortest path from at least one neighbor node to node d of node i does not include the link (s, dn (s, d)); the formalization is represented as:

x ((s, dn (s, d)), s, i) ═ 0 indicates that the link (s, dn (s, d)) is not on the shortest path from node s to node i,

indicating that node i has at least one neighbor node to node d's shortest path excluding link (s, dn (s, d)), for

N (i) represents all neighbor nodes of the node, and dn (s, d) is the optimal next hop from node s to node d.

Images