CN103825823B - Data forwarding method based on different priorities in software-defined network - Google Patents

Abstract

The invention discloses a data forwarding method based on different priorities in a software-defined network. Through the method, a routing flow table is divided into a static route, a dynamic route and a default route which respectively represent forwarding routes with different priorities. When a route is decided, an exchanger is used for detecting the route according to the priorities from the highest to the lowest and selecting the route with the highest priority to forward data. According to the route policy, the forwarding route can be selected by the data forwarding method according to real-time flow condition and user demand of the network, so that the data transmission efficiency is increased and the network jam is effectively relieved; therefore, the data forwarding method disclosed by the invention has wide application prospect.

Description

Data forwarding method in software-defined network based on different priorities

technical field

The invention belongs to the technical field of network routing, and in particular relates to a priority-based routing framework in a software-defined network (Software Defined Network, SDN).

Background technique

Due to various limitations of the existing network architecture, SDN was born. If all network devices in the network are regarded as managed resources, then the concept of a network operating system (NetworkOS) can be abstracted by referring to the principle of the operating system. On the one hand, this network operating system abstracts the specific details of the underlying network devices. At the same time, it also provides a unified management view and programming interface for upper-layer applications. In this way, based on the network operating system platform, users can develop various applications and define logical network topology through software to meet different requirements for network resources. , without having to care about the physical topology of the underlying network, this is SDN.

SDN has many advantages over traditional networks. The controller can grasp the topology information such as the status and data of the entire network. Based on this information, many functions can be realized, for example: (1) Put the path cost in the Open Shortest Path First (OSPF) network on the controller The calculation is much faster than the existing calculation method. (2) The speed of calculating the route will be much faster than the original calculation method. (3) It can provide solutions to some problems in advance, and can solve problems faster when they appear.

The generation of the routing table in the existing network relies on the self-learning algorithm, and each router can only learn the routing information from its neighbors, which is easy to generate loops. The data packets for learning and maintaining the routing table will occupy certain link bandwidth and router resources. With the development and popularization of SDN, more and more services will be provided to users in SDN. Openflow realizes the separation of the data layer and the control layer, in which the openflow switch forwards the data layer, and the Controller realizes the function of the control layer. The controller controls the flow table in the openflow switch through the standard interface of the openflow protocol, so as to realize centralized control of the entire network. A flow table consists of many flow entries, and each flow entry is a forwarding rule. The data packet entering the switch obtains the forwarding destination port by querying the flow table. A flow entry is composed of a header field, a counter, and an operation; the header field is a ten-tuple, which is the identifier of the flow entry; the counter is used to count the statistical data of the flow entry; the operation indicates that the data packet matching the flow entry should The action to perform. The routing protocols of existing networks are generally not suitable for software-defined networking, and the present invention follows.

Contents of the invention

The present invention provides an SDN routing framework based on different priorities. The framework provides a method for selecting forwarding next hops according to different priorities, and provides different network forwarding services according to different user requirements.

In order to realize above-mentioned service, the technical scheme that the present invention provides is:

A method for forwarding data in a software-defined network based on different priorities, characterized in that the method is performed according to the following steps:

(1) When the source needs to send data packets to the sink, the controller judges whether the static route from the source to the sink is enabled; if the static route is enabled, the switch selects the next hop forwarding node according to the static route; otherwise, execute the steps ( 2);

(2) The controller judges whether the dynamic routing from source to sink is enabled; if it is enabled, the controller calculates the forwarding path from source to sink according to the routing selection rules and the real-time status of the network, and the switch selects the forwarding path according to the calculated forwarding path. One-hop forwarding node; if not enabled, perform step (3);

(3) The switch selects the next hop forwarding node according to the default route.

Preferably, enabling or disabling static routing and enabling or disabling dynamic routing in the method are set in the controller according to requirements by the administrator in the software-defined network.

Preferably, in the method, the static route is a forwarding path manually set by the administrator; when the static route is enabled, the switch selects the next-hop forwarding node according to the set static route.

Preferably, the default route in the method is the forwarding path from the source to the sink calculated by the controller according to the shortest path method.

Preferably, the software-defined network in the method is a software-defined network using the openflow protocol, and the openflow switch selects a forwarding route according to a matching result of a packet header.

Preferably, in the method, the data header is provided with matching fields Match Fields, priority field Priority, counting field Counters, instruction field Instructions, timeout field Timeouts, Cookie; when there are multiple matching routing entries, select the highest priority Routing table entry forwarding; the Priority field is the priority of the routing table entry, and the order of priority is static route > dynamic route > default route.

Preferably, in the method, the administrator adds, deletes, and modifies routing table items in the openflow controller according to the openflow protocol.

Preferably, after the static route set by the administrator in the method is enabled, the openflow controller writes the static route into the openflow switch, and checks the flow table at the same time. If there is a static route with the same source and sink as the static route just written, is deleted.

Preferably, static routing is not enabled in the method, and after the administrator enables dynamic routing, the controller calculates the forwarding path from the source to the sink according to the routing rules and the real-time state of the network, and writes the calculated routing result into openflow switch.

Preferably, the default route in the method is that the openflow controller will calculate the forwarding path from the source to the sink according to the shortest path method, and write the default route into the openflow switch; when static routing or dynamic routing is not enabled, the openflow switch The next hop forwarding node will be selected according to the default route.

The principle of the present invention is to classify routes into three types: static routes, dynamic routes or default routes. The priority forwarding process is reflected through route selection, in which static routes have the highest priority, dynamic routes have the second priority, and default routes have the lowest priority. Static routing is a forwarding path manually set by the administrator, and its priority is the highest; when static routing is enabled, the switch selects the next-hop forwarding node according to the set static routing. Dynamic routing is a forwarding path from source to sink calculated by the controller according to the routing selection rules and the real-time state of the network. Its priority is lower than static routing and higher than default routing; when dynamic routing is enabled, the switch Select the next hop forwarding node. The default route is the shortest path route calculated by the controller according to the specified algorithm, and its priority is the lowest; when dynamic routing or static routing is not enabled, the switch will select the next-hop forwarding node according to the default route.

Most software-defined networks use the openflow protocol, and the switch will match the data header (packet header) to select the forwarding route. When there are multiple matching routing entries, the routing entry with the highest priority will be selected for forwarding. The Priority field in the datagram header is the priority of the routing table entry, and the priority order is: static route > dynamic route > default route. At the same time, the openflow protocol provides the functions of adding, deleting, and modifying routing table items, and provides an operation interface for implementing routing selection.

The datagram header looks like this:

MatchFields priority Counters Instructions Timeouts cookies

Among them, Match Fields is a matching field, which matches data packets, including entry ports and data packet headers; Priority is a priority, indicating the matching order of flow table items. ;Counters is a counter, used to update the count of matching packets; Instructions is an instruction, used to modify the action set or pipeline processing; Timeouts is a timeout field, indicating the maximum time count or flow valid time; cookie is an opaque data selected by the controller Values that the controller uses to filter flow statistics, flow changes, and flow deletions, but not when processing packets.

The following are some basic strategies of the framework of the present invention: (1) Static routing strategy: after the static routing set by the administrator is enabled, the controller writes the static routing into the switch, and checks the flow table at the same time. If the source and sink of the static route are the same, delete them. (2) Dynamic routing strategy: After the administrator enables dynamic routing, that is, static routing is not enabled, the controller calculates the routing information from the source based on the real-time status of the network according to the routing selection rules (such as the smallest delay, the smallest maximum load on the path, etc.). A forwarding path to the sink, that is, the network topology is regarded as a weighted graph, the weight of each edge is calculated according to the rules, the shortest path from the source to the sink is calculated using the Dijkstra shortest path algorithm, and the calculated routing result is written into the switch, that is, for dynamic routing. The routing selection here is not unique, and different routing selection rules can be set according to different applications. (3) Default routing strategy: The default routing is that the controller will calculate the forwarding path from the source to the sink according to the shortest path method, that is, the network topology is regarded as an unweighted graph, and the Dijkstra shortest path algorithm is used to calculate the shortest path from the source to the sink. And write the calculated routing result to the switch, which is the default route. When neither static routing nor dynamic routing is enabled, the switch will select the next-hop forwarding node according to the default route.

According to the above basic strategy, the software-defined network routing framework described in the present invention can be expressed as the following steps:

(1) The administrator enables or disables static routing and dynamic routing according to requirements.

(2) When the source needs to send data packets to the sink, if the static route from the source to the sink is enabled, select the next hop forwarding node according to the static route. Otherwise, go to step (3).

(3) Detect whether the dynamic routing from source to sink is enabled. If it is enabled, the controller calculates the route from source to sink according to the routing selection rules (such as minimum delay, minimum maximum load on the path, etc.) according to the real-time state of the network. a forwarding path. If not enabled, go to step (4).

(4) The source node selects the next hop forwarding node according to the default route, where the default route is the forwarding path from the source to the sink calculated by the controller according to the shortest path method.

SDN provides a new network, and it is an important service to provide different routes based on priority. The openflow protocol is a network protocol of SDN, which provides functions of adding, deleting and modifying flow tables, and provides technical possibility for the realization of the solution of the present invention. The user's needs and the real-time status of the network are all factors that affect the routing priority. Therefore, under different conditions, the network provides different routing services for users, which has a broad application prospect.

Compared with the scheme in the prior art, the advantages of the present invention are:

(1) This method is based on software-defined networking, and realizes flexible control of switch forwarding by separating the control plane of network equipment from the data plane.

(2) Different route acquisition methods correspond to different priorities, highlighting the authority of administrators. The static route is set by the administrator and has the highest priority; the dynamic route is calculated based on the real-time network status and has the second highest priority; the default route is calculated based on the shortest path method and has the lowest priority.

The invention discloses a priority-based routing framework in a software-defined network. The framework divides routing flow entries into static routing, dynamic routing and default routing, which represent forwarding paths with different priorities. When making routing decisions, the switch detects routes in descending order of priority, and selects the route with the highest priority for forwarding. According to the routing strategy, the present invention can select forwarding paths according to real-time traffic conditions of the network and user demands, thereby improving data transmission efficiency, effectively alleviating network congestion, and having broad application prospects.

Description of drawings

The present invention will be further described below in conjunction with accompanying drawing and embodiment:

Fig. 1 is the flow chart of switch judging forwarding path in the frame of the present invention;

FIG. 2 is a network topology diagram illustrating a routing policy within the framework of the present invention.

detailed description

The above solution will be further described below in conjunction with specific embodiments. It should be understood that these examples are used to illustrate the present invention and not to limit the scope of the present invention. The implementation conditions used in the examples can be further adjusted according to the conditions of specific manufacturers, and the implementation conditions not indicated are usually the conditions in routine experiments.

Example

In this embodiment, a switch in a software-defined network selects a forwarding path based on priority, which can realize dynamic forwarding. Routes are divided into static routes, default routes, and dynamic routes. Static routes are forwarding paths manually set by administrators with the highest priority; dynamic routes are calculated by the controller based on routing selection rules and real-time network status A forwarding path to the sink has a lower priority than the static route and higher than the default route; in this embodiment, the forwarding path with the smallest delay is set as a dynamic route. This rule is not unique and can be changed according to specific needs. The default route is the shortest route calculated by the controller according to the specified algorithm, and its priority is the lowest. The routing policy of this embodiment is as follows:

(1) Static routing policy: After the static routing set by the administrator is enabled, the controller writes the static routing into the switch and checks the flow table at the same time. If there is a static routing with the same source and sink as the static routing just written, delete it .

(2) Dynamic routing strategy: After the administrator enables dynamic routing, that is, static routing is not enabled, the controller calculates the route from source to sink based on the routing selection rules (the selection rule in this embodiment is the minimum delay) and the real-time status of the network. A forwarding path of a forwarding path, that is, the network topology is regarded as a weighted graph, and the real-time forwarding delay of each switch is calculated proportionally as the weight of the outgoing edge from the switch, and the shortest path from the source to the sink is calculated using the Dijkstra shortest path algorithm, that is, It is a dynamic route, and writes the calculated route result to the switch.

(3) Default routing strategy: Default routing is the forwarding path from source to sink calculated by the controller according to the shortest path method, that is, the network topology is regarded as an unweighted graph, and the shortest path from source to sink is calculated using Dijkstra algorithm, namely For the default route, write the calculated route result to the switch. When neither static routing nor dynamic routing is enabled, the switch will select the next-hop forwarding node according to the default route.

In order to illustrate the framework of the present invention, the present embodiment has used the network as shown in Figure 2, and the rule of dynamic routing selection forwarding path is that time delay is minimum, assuming that all link conditions are all the same, the implementation steps are as follows:

(1) During the entire process of the embodiment, the data packets are sent from h2 to h3 at a constant rate, so that the time required for forwarding from S6 is 10 times the time required for forwarding from an idle switch, all through S6->S5 The routing delay becomes larger;

(2) Send data packets from h1 to h4, observe the forwarding path, and stop sending data packets from h1 to h4 after the observation;

(3) Set the static route h1->S1->S2->S3->S4->S5->h4 from h1 to h4, send data packets from h1 to h4, observe the forwarding path, stop from h1 to h4 after observation h4 sends data packets;

(4) Enable dynamic routing from h1 to h4, send data packets from h1 to h4, observe the forwarding path, and stop sending data packets from h1 to h4 after observation;

(5) Turn off the static routing and dynamic routing from h1 to h4, re-enable the dynamic routing from h1 to h4, send data packets from h1 to h4, observe the forwarding path, and stop sending data packets from h1 to h4 after the observation;

(6) Turn off the dynamic routing from h1 to h4, observe the forwarding path, and stop sending data packets from h1 to h4 after the observation is completed.

The function of the step (1) is to overload the link S6->S5, and the forwarding load of the switch S6 becomes larger, so that the delay of all routes passing through S6->S5 becomes larger.

In step (2), the network topology graph is regarded as an unweighted graph, and the controller obtains the shortest path h1->S1->S6->S5->h4 according to the Dijkstra shortest path algorithm, which is the default route. So it is observed that the forwarding path is h1->S1->S6->S5->h4.

In step (3), it is observed that the forwarding path is h1->S1->S2->S3->S4->S5->h4, which is a static route set by the administrator.

In step (4), it is observed that the forwarding path is still h1->S1->S2->S3->S4->S5->h4, because the priority of static routing is higher than that of dynamic routing, so the forwarding path is still in step (3) Static routing set.

In step (5), it is observed that the forwarding path is h1->S1->S3->S4->S5->h4. After performing step (5), the controller calculates the dynamic route according to the principle of minimum delay, because After turning off the static routing and dynamic routing, the dynamic routing is turned on again. The dynamic routing turned on at this time has the highest priority. The controller calculates the dynamic routing, regards the network topology map as a weighted map, and forwards it from S6 according to step (1). The time required is 10 times the time required for forwarding from an idle switch. The controller assigns a value of 10 to the S6->S5 edge, and 1 to the other edges in the figure. According to Dijkstra’s shortest path algorithm, it is obtained from h1 to h4 The shortest path obtained is h1->S1->S3->S4->S5->h4, so the observed forwarding path is h1->S1->S3->S4->S5->h4;

In step (6), it is observed that the forwarding path is h1->S1->S6->S5->h4, because the administrator has not enabled static routing or dynamic routing, so the switch will forward according to the default route.

The above examples are only to illustrate the technical conception and characteristics of the present invention, and its purpose is to allow people familiar with this technology to understand the content of the present invention and implement it accordingly, and cannot limit the protection scope of the present invention. All equivalent changes or modifications made according to the spirit of the present invention shall fall within the protection scope of the present invention.

Claims (9)

1. A data forwarding method in a software-defined network based on different priorities, characterized in that the method is carried out according to the following steps: (1) When the source needs to send data packets to the sink, the controller judges whether the static route from the source to the sink is enabled; if the static route is enabled, the switch selects the next hop forwarding node according to the static route; otherwise, execute the steps ( 2); (2) The controller judges whether the dynamic routing from source to sink is enabled; if it is enabled, the controller calculates the forwarding path from source to sink according to the routing selection rules and the real-time status of the network, and the switch selects the forwarding path according to the calculated forwarding path. One-hop forwarding node; if not enabled, perform step (3); (3) The switch selects the next hop forwarding node according to the default route; The switch selects the forwarding route according to the matching result of the data header; the data header is set with a matching field, a priority field, a count field, an instruction field, and a timeout field; when there are multiple matching routing entries, the routing entry with the highest priority is selected for forwarding ; The priority field is the priority of the routing entry, and the priority order is static route > dynamic route > default route. 2. The data forwarding method according to claim 1, characterized in that enabling or disabling static routing and enabling or disabling dynamic routing in the method are set in the controller by an administrator in a software-defined network according to requirements. 3. The data forwarding method according to claim 2, wherein in the method, the static route is a forwarding path manually set by the administrator; when the static route is enabled, the switch selects the next hop forwarding node according to the static route set . 4. The data forwarding method according to claim 2, wherein the default route in the method is the forwarding path from the source to the sink calculated by the controller according to the shortest path method. 5. The data forwarding method according to claim 1, wherein the software-defined network in the method is a software-defined network using the openflow protocol, and the switch is an openflow switch. 6. The data forwarding method according to claim 5, characterized in that in the method, the administrator adds, deletes, and modifies routing table items in the openflow controller according to the openflow protocol. 7. The data forwarding method according to claim 5, characterized in that after the static routing set by the administrator in the method is enabled, the openflow controller writes the static routing into the openflow switch, and checks the flow table simultaneously, if there is a static routing If the source and sink of the static route just written are the same, delete it. 8. The data forwarding method according to claim 5, characterized in that static routing is not enabled in the method, and after the administrator enables dynamic routing, the controller calculates the route from source to sink according to the routing rules and the real-time status of the network. forwarding path, and write the calculated routing result to the openflow switch. 9. The data forwarding method according to claim 5, characterized in that in the method, the default route is that the openflow controller can calculate the forwarding path from the source to the sink according to the shortest path method, and write the default route into the openflow switch; When neither static routing nor dynamic routing is enabled, the openflow switch will select the next-hop forwarding node according to the default route.