CN106936731A

CN106936731A - The method and apparatus of the message forwarding in software defined network SDN

Info

Publication number: CN106936731A
Application number: CN201511032559.3A
Authority: CN
Inventors: 宋雪飞; 胡农达; 于德雷; 侯杰
Original assignee: Beijing Huawei Digital Technologies Co Ltd
Current assignee: Beijing Huawei Digital Technologies Co Ltd
Priority date: 2015-12-31
Filing date: 2015-12-31
Publication date: 2017-07-07

Abstract

The embodiment of the invention discloses a kind of method and apparatus of the message forwarding in software defined network SDN, the SDN includes at least one controller and multiple servers, and the method includes：Obtain the link-state information in the SDN；The address of the destination server of the first server in the link-state information and the plurality of server generates the load balancing table of the first server；The load balancing table is sent into the software switch to the first server, message forwarding is carried out according to the load balancing table in order to the software switch in the first server.The method of the message forwarding in the software defined network SDN of the embodiment of the present invention, by dynamic generation load balancing table, it is possible to increase the forward efficiency of load balancing data, reduces the O＆M complexity of load balancing.

Description

The method and apparatus of the message forwarding in software defined network SDN

Technical field

The present invention relates to areas of information technology, and more particularly, in software defined network SDN Message forwarding method and apparatus.

Background technology

With large-scale internet application on site, enterprise-level infrastructure service and to the demand of cloud computing application Growing, 100,000 magnitudes, the data center of even million magnitude servers are born.In this data The heart more has a large amount of mostly using fat tree network (Fat-tree) topological structure, the enormous amount of interchanger OVS SoftSwitches need management.With software defined network (Software Defined Network, Referred to as " SDN ") technology improves the configuration pipe at large-scale data center in the application of data center The flexibility of reason operating efficiency, network autgmentability and service deployment.

SDN is one kind of open network foundation (Open Network Foundation, ONF) definition The new network architecture, it is intended to the frame of personal computer (Personal Computer, PC) calculating field Structure and method redefine network calculations, regard forwarding surface as a central processing unit for execute instruction (Central Processing Unit, CPU), chain of command regards the network operating system of concentration, network meter as Calculation is programmed to by network operating system.But because south orientation agreement is out in SDN frameworks The slow of (OpenFlow) agreement development is released, for the load-balancing function of business in telecommunication transport network Support inadequate, it is impossible to realize load balancing of the Business Stream in transmitting procedure.

Mainly have in order to realize load balancing of the Business Stream in transmitting procedure, in current prior art and pass through Substantial amounts of human configuration IP address, or it is balanced by hash algorithm computational load, and these methods are main Have the drawback that：The complexity of O＆M is higher, and the forward efficiency of load balancing data is very low.

The content of the invention

The embodiment of the invention provides method and dress that the message in a kind of software defined network SDN is forwarded Put, it is possible to increase the forward efficiency of load balancing data.

First aspect, there is provided a kind of method of the message forwarding in software defined network SDN, the SDN Including at least one controller and multiple servers, the method is performed by least one controller, including：

Obtain the link-state information in the SDN；

The destination server of the first server in the link-state information and the plurality of server Address generates the load balancing list item of the load balancing table of the first server；

The load balancing table and the load balancing list item are sent to the first server, in order to this One server carries out message forwarding according to the load balancing table.

The method of the data message forwarding in the software defined network SDN of the embodiment of the present invention, Neng Goudong State generates load balancing table, carries out the forwarding of message, improves the forward efficiency of load balancing data, drops The low O＆M complexity of load balancing.

Here, link-state information includes the faulty link letter between core switch and convergence switch Breath, the faulty link information between convergence switch and access switch, access switch and server it Between faulty link information.Controller generates the list item in the load balancing table according to faulty link information.

In some possible implementations, the SDN also includes multiple core switch and multiple convergences Interchanger, the plurality of core switch is the Upper Switch of the plurality of convergence switch, the Link State Information includes fault-free link information, the first clothes in the link-state information and the plurality of server The address of the destination server of business device generates the load balancing table of the first server, including：

Address according to the fault-free link information and the destination server is generated in the load balancing table Load balancing list item, wherein, the core switch address list of the load balancing list item is the plurality of core Switch address list, the convergence switch address list of the load balancing list item is the first server institute Multiple convergence switch address lists in the node set at place.

With reference in a first aspect, in the first possible implementation, the SDN also includes multiple cores Interchanger and multiple convergence switches, the plurality of core switch are that the upper strata of the plurality of convergence switch is handed over Change planes, the link-state information includes Fisrt fault link information, the Fisrt fault link information is more for this The first convergence in the first core switch and the plurality of convergence switch in individual core switch is exchanged The information of link failure between machine；

Wherein, the purpose service of the first server in the link-state information and the plurality of server The address of device generates the load balancing table of the first server, including：

Address according to the Fisrt fault link information and the destination server is generated in the load balancing table Load balancing list item.

With reference to the first possible implementation of first aspect, in second possible implementation, In descendants's server that the first server is first convergence switch, described first converges exchange When machine is in different node sets from the destination server, the core of list item in the load balancing table Switch address list is that the multiple core switch address removes the first core switch address Address list afterwards, wherein, it is only capable of being communicated by core switch between the different node set.

With reference to the first possible implementation of first aspect, in the third possible implementation, It is not descendants's server of first convergence switch, the destination server in the first server When being descendants's server of first convergence switch,

If the Part I of the address of the destination server and the of the first core switch address It is a part of identical, and the Part II of the address of the destination server and first core switch ground The Part III of location is identical, then the core switch address list in the load balancing table is the multiple Core switch address removes the address list behind the first core switch address；

Or, the Part I of the address of the destination server and the of the first core switch address It is a part of different, or the address of the destination server Part II and first core switch ground The Part III of location is different, then the core switch address list in the load balancing table is the multiple Core switch address.

With reference to the first possible implementation of first aspect, in the 4th kind of possible implementation, In descendants's server that the first server is first convergence switch, described first converges exchange Machine is in first node set together with the destination server, and first convergence switch and first are accessed Interchanger is connected, and the destination server is connected with the second access switch, first access switch When different from second access switch, the convergence switch address list part of the load balancing list item Multiple convergence switch address lists in for the first node set.

With reference in a first aspect, in the 5th kind of possible implementation, the SDN also includes multiple cores Interchanger, multiple convergence switches and multiple access switch, the plurality of core switch is the plurality of remittance The Upper Switch of poly- interchanger, the plurality of convergence switch is that the upper strata of the plurality of access switch exchanges Machine, the link-state information includes the second faulty link information, and second faulty link information is the plurality of The 3rd access switch in the second convergence switch and the plurality of access switch in convergence switch Between link failure information；

Wherein, the purpose clothes of the first server in the link-state information and the plurality of server The address of business device generates the load balancing table of the first server, including：

Address according to second faulty link information and the destination server is generated in the load balancing table Load balancing list item.

With reference to the 5th kind of possible implementation of first aspect, in the 6th kind of possible implementation, In descendants's server that the first server is the 3rd access switch, the 3rd access switch with should Destination server is in Section Point set together, and in the destination server and the plurality of access switch Four access switch are connected, and when the 3rd access switch is different from the 4th access switch, then this is born The convergence switch address list for carrying load balancing list item in sharing table is many in the Section Point set Individual convergence switch address removes the address list behind the second convergence switch address.

With reference to the 5th kind of possible implementation of first aspect, in the 7th kind of possible implementation, In descendants's server that the first server is the 3rd access switch, the 3rd access switch with should Destination server is in different node sets, the core in load balancing list item in the load balancing table Switch address list removes the corresponding ancestral of second convergence switch for the plurality of core switch address Address list behind first core switch address, wherein, it is only capable of by core between the different node set Interchanger is communicated.

With reference to the 5th kind of possible implementation of first aspect, in the 8th kind of possible implementation, It is not descendants's server of the 3rd access switch, the first server in the first server The 3rd node set is in together with the 3rd access switch, the destination server connects for the described 3rd When entering descendants's server of interchanger,

If the Part I of the address of the destination server and the of the second convergence switch address It is a part of identical, and the Part II of the address of the destination server and second convergence switch ground The Part III of location is identical, then the convergence switch address list in the load balancing table is the described 3rd Multiple convergence switch addresses in node set remove the address behind the second convergence switch address List；

If or, the Part I of the address of the destination server and the second convergence switch address Part I is different, or the destination server address Part II and second convergence switch ground The Part III of location is different, then the convergence switch address list of the load balancing list item is the described 3rd Multiple convergence switch address lists in point set.

With reference to the 5th kind of possible implementation of first aspect, in the 9th kind of possible implementation, Different node sets, the purpose clothes are in the first server and the 3rd access switch When business device is descendants's server of the 3rd access switch,

If ancestors' core of the Part I of the address of the destination server and second convergence switch The Part II of heart switch address is identical, and the Part II of the address of the destination server with it is described The Part III of ancestors' core switch address of the second convergence switch is identical, then the first server On software switch load balancing table in the core switch address list of load balancing list item be institute State multiple core switch addresses and remove the corresponding ancestors' core switch ground of second convergence switch Address list behind location, wherein, it is only capable of being led to by core switch between the different node set Letter；

If or, the ancestors of the Part I of the address of the destination server and second convergence switch The Part II of core switch address is different, or the destination server address Part II and institute The Part III for stating ancestors' core switch address of the second convergence switch is different, then the load balancing The core switch address list of list item be the multiple core switch address list, wherein, it is described not It is only capable of being communicated by core switch between same node set.

With reference in a first aspect, in the tenth kind of possible implementation, the SDN also includes multiple access Interchanger, the plurality of access switch is the Upper Switch of the plurality of server, the link-state information Including the 3rd faulty link information, the 3rd faulty link information is the 5 in the plurality of access switch The information of link failure between second server in access switch and the plurality of server；

Address according to the 3rd faulty link information and the destination server is generated in the load balancing table Load balancing list item.

With reference to the tenth kind of possible implementation of first aspect, in a kind of the tenth possible implementation In, when the first server is the second server, the remittance of load balancing list item in the load balancing table Poly- switch address list and core switch address list are sky.

With reference to the tenth kind of possible implementation of first aspect, in the 12nd kind of possible implementation In, it is not the second server in the first server, the destination server is second clothes Business device, the destination server is in fourth node set together with the second server,

If the address of the destination server is identical with the address of the second server, the load point The convergence switch address list and core switch address list of load balancing list item are sky in load table；

If described negative or, the address of the destination server is different from the address of the second server The core switch address for carrying load balancing list item in sharing table is the multiple core switch address, institute It is that multiple convergence the in the fourth node set exchanges to state the convergence switch address in load balancing table Machine address.

Second aspect, there is provided a kind of method of the message forwarding in software defined network SDN, the SDN Including at least one controller and multiple servers, the method is by the first server in the plurality of server On software switch perform, including：

Receive at least one controller transmission load balancing table, the load balancing table be this at least one Address life of the controller according to the destination server of the link-state information and the first server of the SDN Into；

Message forwarding is carried out according to the load balancing table.

With reference to second aspect, in the first possible implementation, this is carried out according to the load balancing table Message forwarding includes：

Next-hop switch address is selected to encapsulate the message as outer layer IP according to the load balancing table, and Forward the message.

A kind of third aspect, there is provided the controller in software defined network SDN, it is above-mentioned for performing Method in any possible implementation of first aspect or first aspect.Specifically, the device includes For the unit of the method in any possible implementation for performing above-mentioned first aspect or first aspect.

Fourth aspect, there is provided the software in the first server in a kind of software defined network SDN is handed over Change planes, for the method in any possible implementation for performing above-mentioned second aspect or second aspect. Specifically, the device includes any possible realization side for performing above-mentioned second aspect or second aspect The unit of the method in formula.

5th aspect, there is provided a kind of device of the message forwarding in software defined network SDN, the dress Put including：Receiver, transmitter, memory, processor and bus system.Wherein, the receiver, The transmitter, the memory are connected with the processor by the bus system, and the memory refers to for storage Order, the processor is used to perform the instruction of the memory storage, to control receiver to receive signal, and controls Transmitter sending signal processed, and when the instruction of the computing device memory storage, the execution makes Obtain the method in any possible implementation of the computing device first aspect or first aspect.

6th aspect, there is provided a kind of device of the message forwarding in software defined network SDN, the dress Put including：Receiver, transmitter, memory, processor and bus system.Wherein, the receiver, The transmitter, the memory are connected with the processor by the bus system, and the memory refers to for storage Order, the processor is used to perform the instruction of the memory storage, to control receiver to receive signal, and controls Transmitter sending signal processed, and when the instruction of the computing device memory storage, the execution makes Obtain the method in any possible implementation of the computing device second aspect or second aspect.

7th aspect, there is provided a kind of system, including the third aspect controller, the first of fourth aspect Server.

A kind of eighth aspect, there is provided computer-readable medium, for storing computer program, the calculating Machine program includes being used to perform the method in any possible implementation of first aspect or first aspect Instruction.

Brief description of the drawings

Technical scheme in order to illustrate more clearly the embodiments of the present invention, below will be in the embodiment of the present invention The required accompanying drawing for using is briefly described, it should be apparent that, drawings described below is only this Some embodiments of invention, for those of ordinary skill in the art, are not paying creative work Under the premise of, other accompanying drawings can also be obtained according to these accompanying drawings.

Fig. 1 is a schematic diagram of the network planning of the embodiment of the present invention.

Fig. 2 is the method for the message forwarding in software defined network SDN according to embodiments of the present invention Indicative flowchart.

Fig. 3 is the method for the message forwarding in software defined network SDN according to embodiments of the present invention Another indicative flowchart.

Fig. 4 is the schematic block diagram of controller according to embodiments of the present invention.

Fig. 5 is the software switch schematic block diagram in first server according to embodiments of the present invention.

Fig. 6 is the structural representation of controller according to embodiments of the present invention.

It according to the present invention is software switch structural representation in the first server that example is that Fig. 7 is.

Specific embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out Clearly and completely describe, it is clear that described embodiment is a part of embodiment of the invention, rather than Whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art are not making wound The every other embodiment obtained under the premise of the property made work, belongs to the scope of protection of the invention.

Technical scheme, can apply to various communication systems, for example：Global system for mobile telecommunications system System (GSM, Global System of Mobile communication), CDMA (CDMA, Code Division Multiple Access) system, WCDMA (WCDMA, Wideband Code Division Multiple Access Wireless), GPRS (GPRS, General Packet Radio Service), Long Term Evolution (LTE, Long Term Evolution) etc..

Large-scale data center generally uses fat tree network topology structure, for such network environment, often Connected as the hardware between server using the interchanger of isomorphism.For convenience, it is with reference to Fig. 1 Example is illustrated.Fig. 1 is a schematic diagram of the network planning of the embodiment of the present invention.In data center or In the network planning of enterprise, network can be divided into Access Layer, convergence-level and core layer, in this three layers In interchanger correspond to access switch, convergence switch and core switch respectively.Wherein, access Interchanger is used for access terminal, and convergence switch is used to converge the access switch of its lower floor, and core is exchanged Machine is used to converge the convergence switch of its lower floor, while also being responsible for being connected on internet.Access switch Lower end can be connected with SoftSwitch, i.e., with open virtual switch (Open Virtual Switch, abbreviation Be " OVS ") be connected, each OVS lower end can correspond to multiple virtual machines (Virtual Machine, Referred to as " VM ")., it is necessary to according to forwarding flow table or load when being communicated between different VM Sharing table carries out message forwarding, so as to realize between different VM between i.e. source VM and purpose VM Communication when being combined in same combination of nodes or cross-node.And forward flow table or load balancing table to be required to Controller is generated and issued, and could realize communication between different VM for example, first node group in Fig. 1 VM1 in conjunction under OVS1 to be combined with Section Point in OVS5 under VM1 communicate, it is necessary to root The forwarding of message is completed according to load balancing table.

In general, access switch is all box-shaped device (such as TOR), also have frame (such as EOR).Cassete exchange machine small volume, interface is few, also not expansible.Convergence switch and core are exchanged Machine is all frame, and volume is big, interface quantity is more, can insert different veneers, to meet different need Ask, favorable expandability some.For aspect of performance, from core switch, convergence switch, exchange is accessed Machine successively decreases successively, and core switch performance is best.These three interchangers are hardware switch, and support Open flows OPENFLOW agreements, software switch is open virtual switch (Open Virtual Switch, referred to as " OVS ").Wherein, an OVS can correspond to a physical server, one Individual physical server can create multiple VM.

In embodiments of the present invention, using CORE interchangers as core switch, AGG (aggregate) Interchanger is convergence switch and TOR interchangers to be illustrated as a example by access switch.For example, Core Interchanger, AGG interchangers and TOR interchangers are all isomorphism interchanger, the end of these hardware switch Mouth quantity is K.Plan in the following way address for interchanger：The volume of software switch Location mode is Area.Pod.Pos.Port, and hardware switch addressing mode is Area.Level.Pod.Pos, its In, POD represents the logic groups of switch-to-server, and Port represents port numbers；Pod mono- has The set of the node of specific position relation, is further divided to two classes：ATOPod and CorePod；Pos：Section Position Number of the point in POD, OVS/TOR/Agg/Core nodes respectively have independent Pos spaces. In Fig. 1, first node set, Section Point set and the 3rd node set correspond to different respectively POD.If hardware switch uses 96 mouth switchs, i.e. k=96, then AggPod is respectively 1 He First node set and Section Point set in 2, i.e. corresponding diagram 1；CorePod values are 3 i.e. corresponding diagram The addressing example of the 3rd node set in 1, Area=10, hardware switch and software switch is as follows：

Core switch is Core interchangers (address is Area.253.Pod.Pos, i.e. 10.253.1.1), The address of lower port i (i ∈ [1, k]) connected Agg interchangers of the interchanger is Area.252.Pod.Pos, That is 10.252.96.1..

Convergence switch is Agg interchangers (address is Area.252.Pod.Pos, i.e. 10.252.1.47), The address of the connected Core interchangers of upper port i of the interchanger is Area.253.Pos.i-k/2, i.e., 10.253.1.1；The address of the connected TOR interchangers of lower port i of the interchanger is Area.251.Pod.i, That is 10.251.1.1.

Access switch is TOR interchangers (address is Area.251.Pod.Pos), the interchanger it is upper The address of the connected Agg interchangers of port i is Area.252.Pod.i-k/2, i.e. 10.252.1.1；The friendship The address of the connected OVS interchangers of lower port i changed planes is Area.Pod.Pos.i, i.e. 10.1.1.1.

In Fig. 1, the corresponding IP address of controller 1 is 191.4.3.108, the corresponding IP of controller 2 Address is 191.4.3.109, and the corresponding IP address of controller 3 is 191.4.3.110, and controller is controlled respectively The different POD of system, solid line is primary control connection, and dotted line is standby control connection.Controller can be with base In above-mentioned network address planing method and network it is assumed that carrying out performing corresponding operation, realize that Business Stream exists Load balancing in transmitting procedure, wherein, load balancing is to share to be performed on multiple operating units.

Specifically, controller is pressed after network topological information (topology information, net element information) is got It is that hard interchanger and SoftSwitch configure IP address or relevant information according to address above mentioned planing method, including The IP address of the IP address, tenant's information and/or purpose interchanger of purpose virtual machine, according to what is got IP address information can generate local forwarding flow table or load balancing load balancing table, and by this local turn Hair flow table or load balancing load balancing table are sent to purpose interchanger.

For example, controller can log in OVS by telnet modes, send configuration order is OVS Configuring virtual LAN (Virtual Local Area Network, referred to as " VLAN ") and extension are empty Intend the relevant information of LAN (virtual Extensible LAN, referred to as " VXLAN ").Typically For, because 4096 VLAN are far from the demand for meeting extensive cloud computation data center, therefore Need to introduce VXLAN.

Further, the relevant configuration of VLAN can be：

Ovs-vsctl add-port brname portname vlan_mode=access tag=vlanId vm1port；

The relevant configuration of VXLan can be：

ovs-vsctl set Port<port_name>Vni_tag=<vxlan_id

Ovs-vsctl add-port brname vxlan0--set Interface vxlan0type=vxlan options:Remote_ip=flow options:Key=flow.

Local forwarding flow table on the corresponding OVS of the configuration information of the VLAN and VXLAN of above-mentioned OVS, Including VLAN forward tables and VXLAN forward tables.For example, the forward table T0 such as table 1 below of VLAN It is shown：

Table 1

For example, the forward table T1 of VXLAN is as shown in table 2 below：

Table 2

In table 2, flow table item 1 for OvS with tenant with VLAN scene；The pin of flow table item 2 To same OvS with tenant's difference VLAN scene；Flow table item 3 is directed to different with OvS difference tenants The scene of VLAN；Flow table item 4 for different OvS with tenant with VLAN scene；Flow table item 5 For different OvS with tenant's difference VLAN scene；Flow table item 6 is for different OvS difference tenants The scene of different VLAN.

Above-mentioned flow table item can not simultaneously be present, and controller, can after tenant's information and VM information is obtained To generate and issue above-mentioned flow table item 1,2,3, wherein, tenant can access same data or exchange The set of the user of machine a, user can access multiple switch, and tenant can be with the network in corresponding table Identifier (VXLAN Network Identifier, referred to as " VNI ").

Controller generates local load balancing table through the above way, and local load balancing table is handed down to OVS

It should be understood that the present invention is only illustrated, or other with above-mentioned network address planning Plan, can be manually set, do not limit the invention in address.

It should also be understood that Fig. 1 is only an example, for the number of the VM of appearance and various friendships in figure The number changed planes is not restricted.

The method of the message forwarding in the software defined network SDN of the embodiment of the present invention may apply to figure Network system in 1, network system is only by taking Fig. 1 as an example, however it is not limited to this.

Fig. 2 shows the side of the message forwarding in software defined network SDN according to embodiments of the present invention Method 200.Message forwarding in software defined network SDN below in conjunction with Fig. 2 to the embodiment of the present invention Method be described.The SDN includes at least one controller and multiple servers, and the method is by this At least one controller is performed, and the method 200 includes：

S210, obtains the link-state information in the SDN；

S220, the purpose service of the first server in the link-state information and the plurality of server The address of device generates the load balancing table of the first server；

S230, sends the software switch to the first server, in order to this by the load balancing table Software switch in first server carries out message forwarding according to the load balancing table.

Specifically, in SDN, controller can get link-state information, for example, When certain link breaks down, controller can receive the notice report of the link-failure state from interchanger Text, the destination server of the first server in these link-state informations and the plurality of server The load balancing table that address generates the first server includes load balancing list item, for example, controller can be with The IP address of network identifier information, destination server according to source server and the IP of purpose virtual machine Address generates load balancing table, and the load balancing table of first server is load balancing table, and this is loaded Sharing table is issued to the software switch on source server, for example, first server is source server, with The software switch being easy on source server carries out message forwarding according to the load balancing table.Wherein, software Interchanger can be understood as OVS.

In above-mentioned table 2, list item 1,2,3 is directed to for identical OVS, as long as controller Get the local forwarding flow table of generation by the IP address and tenant's information of purpose virtual machine.And for not The situation of same OVS, controller also needs to get the address of purpose OVS, could generate and issue List item 4,5,6.And for list item 4,5,6, controller is according in the SDN of the embodiment of the present invention The method generation load balancing table of message forwarding.

In embodiments of the present invention, the link-state information that controller gets can include fault-free link Information, the i.e. state without failed link, for the load balancing table of no-failure link in, only comprising one Bar default route forwarding-table item, the example of load balancing table can be as shown in table 3 below and table 4, and table 3 is Convergence switch address list when occurrence is default route is convergence switch list AggrLBListTable, table 4 is core switch address list i.e. core when occurrence is default route Interchanger list CoreLBListTable：

Table 3

Match Fields	Actions
		0.0.0.0	AggLocList

Table 4

Match Fields	Actions
		0.0.0.0	CoreLocList

When failure link hereafter, load balancing table is all based on above-mentioned table 3 and table 4 Modify, such as, new route table items, removal and failure chain are increased for corresponding load balancing table The connected address in road.

The method of the message forwarding in the software defined network SDN of the embodiment of the present invention, by dynamic life Into load balancing table, it is possible to increase the forward efficiency of load balancing data in data center, load is reduced O＆M complexity in a balanced way.

In SDN, the link-state information that controller gets can include three kinds of feelings of faulty link Condition, respectively：Core switch is lacked with the link of convergence switch, convergence switch is exchanged with access Link missing between machine, the link missing between access switch and server.Below will be to these three In the case of load balancing table be described in detail respectively.

In the first case, alternatively, the SDN also includes multiple core switch and multiple convergences Interchanger, the plurality of core switch is the Upper Switch of the plurality of convergence switch, the Link State During information includes that Fisrt fault link information, the Fisrt fault link information are the plurality of core switch Link failure between the first convergence switch in first core switch and the plurality of convergence switch Information；

Address according to the Fisrt fault link information and the destination server generates the load balancing table.

Here, faulty link is the link between the first core switch and the first convergence switch.

For the server being connected with the faulty link, three kinds can be segmented：

1) alternatively, in descendants's server that the first server is first convergence switch, this When one convergence switch is in different node sets from the destination server, the core of the load balancing list item After the list of heart switch address removes the first core switch address for the plurality of core switch address Address list, wherein, be only capable of being communicated by core switch between the different node set.

Specifically, destination server is in different nodes from the first convergence switch on faulty link It is only capable of being communicated by core switch between set, the different node set.Here load balancing Table is, it is necessary to core switch address list in changing default route, i.e. removal are connected with faulty link Core switch address list.Amended load balancing table can be with as shown in table 5 below：

Table 5

Destination server address field Corresponding matching (Match Fields) in table 5, can be handed over next-hop Address list of changing planes correspondence performs action (Actions), i.e. destination server can be in the address column of Actions An occurrence is selected in table, is performed.In table below occur destination server address field and can It is identical with implication here with the list of next-hop switch address, for sake of simplicity, will no longer when appearing below Repeat.

Here, for the load balancing on the descendants's server on the convergence switch that is connected with faulty link Table A ggrLBListTable, either destination server are in identical node set also with faulty link It is that different nodes combinations does not all influence.

2) alternatively, it is not descendants's server of first convergence switch in the first server, should When destination server is descendants's server of first convergence switch,

If the Part I of the address of the destination server and first of the first core switch address Split-phase is same, and the Part II of the address of the destination server and the 3rd of the first core switch address Part is identical, then the core switch address list of the load balancing list item is the plurality of core switch ground Location removes the address list behind the first core switch address；

Or, first of the Part I of the address of the destination server and the first core switch address Point difference, or the destination server address Part II and the first core switch address the 3rd Part is different, then the core switch address list of the load balancing list item is the plurality of core switch ground Location.

Specifically, it is that source server is not connected with faulty link with difference 1), for example, First server is not descendants's server of first convergence switch.Meanwhile, destination server is failure Descendants's server of convergence switch under link, for example, destination server is first convergence switch Descendants's server.Here corresponding load balancing table needs to increase by one in core switch address list Route, amended load balancing table can be with as shown in table 6 below：

Table 6

Two kinds of situations are included in table 6：When on the Part I and faulty link of destination server address field The Part I of the first core switch is identical, and destination server address field Part II and failure chain When the Part III of the first core convergence switch is identical on road, corresponding available next-hop switch address List；Or when the of the first core switch on Part I and the faulty link of destination server address field It is a part of different, or the first core switch on the Part II and faulty link of destination server address field Part III it is different when, corresponding available next-hop switch address list.

Here, destination server address, core switch address are according to hereinbefore network address planning side Method determination, i.e. destination server address Area.Pod.Pos.i, core switch address is Area.252.Pod.Pos, the core switch that the Area of destination server address is connected with the faulty link The Area of address is identical, the core switch that the Pod of destination server address is connected with the faulty link The Pod of address is identical.For example, being Area, second for the Part I of the address of destination server It is divided into Pod, is Area, Part III for the Part I of the core switch address on faulty link It is Pod.For example, in matching item part subnet mask for 16, it is necessary to increase in load balancing table Can be with the list of next-hop switch address for all of core is exchanged in one list item, i.e. load balancing list item Address list after machine address removal failure core switch address, for example, remove the first core switch ground Location.

If in addition, the core switch address that the Area of destination server address is connected with the faulty link Area it is different, or destination server address the Pod core switch that is connected with the faulty link The Pod of location is different, then can use the list of next-hop switch address is all of core switch address.

It should be understood that the Area of address field is described as into Part I here, Pod be described as Part II or Part III, other do not cause the implication of ambiguity, merely to description is convenient, not to the present invention It is construed as limiting, also, the Part I, Part II or the Part III that appear below are also such.

It should also be understood that only being illustrated so that subnet mask is 16 as an example here, or other conjunctions Reason mode, the invention is not limited in this regard.

For the load balancing table on the server in other node sets outside faulty link, if Destination server is to be located at different node sets and the non-convergence being connected with faulty link from faulty link Descendants's server of interchanger, then load balancing table there is no any influence.

3) alternatively, in descendants's server that the first server is first convergence switch, this One convergence switch and the destination server are in first node set together, first convergence switch and the One access switch is connected, and the destination server is connected with the second access switch, and first access is exchanged When machine is different from second access switch, then the convergence switch address list of the load balancing list item is Multiple convergence switch address lists in the first node set.

Specifically, source server is connected with faulty link, for example, first server is the first convergence handing over The descendants's server changed planes.Meanwhile, destination server and the first convergence switch are in identical set of node Close, but be connected from different access switch respectively.In this case, load balancing table does not receive shadow Ring.

In second situation, alternatively, the SDN also includes multiple core switch, multiple convergences Interchanger and multiple access switch, the plurality of core switch are that the upper strata of the plurality of convergence switch is handed over Change planes, the plurality of convergence switch is the Upper Switch of the plurality of access switch, Link State letter Breath includes the second faulty link information, and second faulty link information is the in the plurality of convergence switch The letter of link failure between the 3rd access switch in two convergence switches and the plurality of access switch Breath；

Here, faulty link is the link between the second convergence switch and the 3rd access switch.

For the server device being connected with the faulty link, four kinds can be subdivided into：

1) alternatively, in descendants's server that the first server is the 3rd access switch, this Three access switch and the destination server are in Section Point set together, the destination server with it is the plurality of The 4th access switch in access switch is connected, and the 3rd access switch is accessed with the 4th and exchanged When machine is different, the convergence switch address list of the load balancing list item is many in the Section Point set Individual convergence switch address removes the address list behind the second convergence switch address.

Specifically, source server is connected with faulty link, for example, first server is the 3rd access handing over The descendants's server changed planes.Meanwhile, destination server and the 3rd access switch are in same node set, But destination server is connected with the 4th access switch, the 3rd access switch and the 4th access switch are not Together.That is destination server and faulty link are in identical node set but the different access of connection Interchanger.Here load balancing table needs to change the AggrLBListTable acquiescences road in load balancing By the address list of next-hop interchanger needs to remove the convergence switch address being connected with faulty link. Specific load balancing table example is as shown in table 7 below：

Table 7

2) alternatively, in descendants's server that the first server is the 3rd access switch, this Three access switch are in different node sets, the core of the load balancing list item from the destination server Switch address list removes the corresponding ancestral of second convergence switch for the plurality of core switch address Address list behind first core switch address, wherein, it is only capable of by core between the different node set Interchanger is communicated.

Specifically, it is that destination server is different from faulty link with difference 1) Node set, for example, destination server is in different node sets from the 3rd access switch.Here Load balancing table need to change the default route list item in load balancing table CoreLBListTable, under The address list of one jump interchanger needs to remove ancestors' core of the convergence switch being connected with faulty link Switch address.Specific load balancing table example is as shown in table 8 below：

Table 8

3) alternatively, it is not descendants's server of the 3rd access switch in the first server, should First server and the 3rd access switch are in the 3rd node set together, the destination server for this During descendants's server of three access switch,

If the Part I of the address of the destination server and first of the second convergence switch address Split-phase is same, and the Part II of the address of the destination server and the 3rd of the second convergence switch address Part is identical, then during the convergence switch address list of the load balancing list item is the 3rd node set Multiple convergence switch addresses remove the address list behind the second convergence switch address；

If or, the Part I of the address of the destination server and the first of the second convergence switch address Part is different, or the destination server address Part II and the second convergence switch address the Three parts are different, then the convergence switch address list of the load balancing list item is in the 3rd node set Multiple convergence switch address lists.

Specifically, it is that source server is not connected with faulty link, example with difference 1) and 2) Such as, the first server is not descendants's server of the 3rd access switch.Also, the first service Device is in same Pod with the 3rd access switch, for example, be in the 3rd node set together.Meanwhile, mesh Server failure link on descendants's server, for example, descendants's server of the 3rd access switch. Here, load balancing table needs to increase a route table items on AggrLBListTable, for failure Address field on link is, it is necessary to remove the convergence switch address being connected with faulty link.Specific load Sharing table example is as shown in table 9 below, wherein, the F in table is represented on Failure, i.e. representing fault link The second convergence switch address：

Table 9

Two kinds of situations are included in table 9：When on the Part I and faulty link of destination server address field The Part I of the second convergence switch is identical, and destination server address field Part II and failure chain When the Part III of the second convergence switch is identical on road, corresponding available next-hop switch address row Table；Or when the Part I and first of the second convergence switch on faulty link of destination server address field Part is different, or the second convergence switch on the Part II and faulty link of destination server address field When Part III is different, corresponding available next-hop switch address list.

Here, destination server address, convergence switch address are according to hereinbefore network address planning side Method determination, i.e. destination server address Area.Pod.Pos.i, convergence switch address is Area.252.Pod.Pos, the convergence switch that the Area of destination server address is connected with the faulty link The Area of address is identical, the convergence switch that the Pod of destination server address is connected with the faulty link The Pod of address is identical.For example, being Area, second for the Part I of the address of destination server It is divided into Pod, is Area, Part III for the Part I of the convergence switch address on faulty link It is Pod.In this case, when subnet mask is for 24 in load balancing table, it is necessary to increase by one Can be all of remittance in this Pod with next-hop switch address list in bar list item, i.e. load balancing list item Address list behind poly- switch address removal failure convergence switch address, for example, remove second and converge friendship Change planes address.

If in addition, the convergence switch address that the Area of destination server address is connected with the faulty link Area it is different, or destination server address the Pod convergence switches that are connected with the faulty link The Pod of location is different, then can use the list of next-hop switch address is all of convergence switch in this Pod Address.

4) different node sets alternatively, is in from the 3rd access switch in the first server, When the destination server is descendants's server of the 3rd access switch,

If the Part I of the address of the destination server is handed over ancestors' core of second convergence switch Change planes address Part II it is identical, and the Part II of the address of the destination server with this second converge The Part III of ancestors' core switch address of interchanger is identical, then the core of the load balancing list item is handed over Address list of changing planes removes the corresponding ancestors of second convergence switch for the plurality of core switch address Address list behind core switch address, wherein, it is only capable of being handed over by core between the different node set Change planes and communicated；

If or, ancestors' core of the Part I of the address of the destination server and second convergence switch The Part II of switch address is different, or the Part II of the address of the destination server second converges with this The Part III of ancestors' core switch address of poly- interchanger is different, then the core of the load balancing list item Switch address list is the plurality of core switch address list, wherein, between the different node set It is only capable of being communicated by core switch.

Specifically, it is that source server is not attached to different from faulty link with difference 3) Node set, wherein, it is only capable of being communicated by core switch between different node sets.Here, Load balancing table needs to increase a route table items on CoreLBListTable, on faulty link Address field, it is necessary to remove the core switch address being connected with faulty link.Specific load balancing table Example is as shown in table 10 below, wherein, the second convergence switch on F representing fault links in table Ancestors' core switch address：

Table 10

Two kinds of situations are included in table 10：When the Part I and faulty link of destination server address field The Part I of upper second convergence switch is identical, and destination server address field Part II and failure When the Part III of ancestors' core switch of the second convergence switch is identical on link, it is corresponding it is available under One jumps switch address list；Or when second on the Part I and faulty link of destination server address field The Part I of ancestors' core switch of convergence switch is different, or destination server address field second It is right when part is different from the Part III of ancestors' core switch of the second convergence switch on faulty link The available next-hop switch address list answered.

Here, destination server address, core switch address are according to hereinbefore network address planning side Method determination, i.e. destination server address Area.Pod.Pos.i, core switch address is Area.252.Pod.Pos, the convergence switch that the Area of destination server address is connected with the faulty link Ancestors' core switch address Area it is identical, the Pod of destination server address and the faulty link The Pod of ancestors' core switch address of connected convergence switch is identical.For example, for purpose service The Part I of the address of device is Area, and Part II is Pod, for convergence switch on faulty link Ancestors' core switch address Part I be Area, Part III is Pod.For example, in matching Item parts subnet mask is 24, it is necessary to increase a list item, i.e. load balancing in load balancing table Can be that all of core switch address removes failure convergence with the list of next-hop switch address in list item Address list behind ancestors' core switch address of interchanger, for example, remove the ancestral of the second convergence switch First core switch address.

If in addition, the ancestral of the convergence switch that the Area of destination server address is connected with the faulty link The Area of first core switch address is different, or destination server address Pod and the faulty link phase The Pod of ancestors' core switch address of convergence switch even is different, then can use next-hop interchanger ground Location list is all core switch addresses.

In the third situation, alternatively, the SDN also includes multiple access switch, the plurality of to connect Enter the Upper Switch that interchanger is the plurality of server, the link-state information includes the 3rd faulty link Information, the 3rd faulty link information is that the 5th access switch in the plurality of access switch is more with this The information of link failure between second server in individual server；

Here, faulty link is the link between the 5th access switch and second server.

For the server being connected with the faulty link, two kinds can be subdivided into：

1) alternatively, when the first server is the second server, the remittance of the load balancing list item Poly- switch address list and core switch address list are sky.

Specifically, source server is joined directly together with faulty link, for example, the first server is exactly Two servers.Here, the address list of the available next-hop interchanger of destination server is changed into empty, as " NULL " value.The instantiation of load balancing table as shown in table 11 below and table 12, wherein, table 11 It is AggLBListTable routing overhead sharing tables, table 12 is shared for CoreLBListTable routing overheads Table：

Table 11

Destination server address field	Next-hop switch address list can be used
		0.0.0.0/0	NULL

Table 12

2) be not alternatively, the second server in the first server, the destination server be this Two servers, the destination server is in fourth node set together with the second server,

If the address of the destination server is identical with the address of the second server, the load balancing list item Convergence switch address list and core switch address list for sky；

If or, the address of the destination server is different from the address of the second server, the load balancing Core switch address in table is the plurality of core switch address, and the convergence in the load balancing table is handed over Address of changing planes is multiple convergence switch addresses in the fourth node set.

Specifically, source server is not connected with faulty link, and destination server is connected for faulty link, I.e. destination server is second server.Here, load balancing table needs to increase a route, load point The instantiation of table is carried on a shoulder pole as shown in table 13 below and table 14, wherein, table 13 is AggLBListTable roads By load balancing table, table 14 is CoreLBListTable routing overhead sharing tables：

Table 13

Table 14

Destination server address field	Next-hop switch address list can be used
		0.0.0.0/0	{ all core switch addresses }
The address of the server being connected with faulty link	NULL

Two kinds of situations are included in table 13 and table 14 respectively：When the address field of destination server and second takes When the address field of business device is identical, corresponding available next-hop switch address list；Or, working as purpose service When the address field of device is different from the address field of second server, corresponding available next-hop switch address List.

Alternatively, there can also be two kinds of special treatment methods in such failed link, to reduce table resource Consumption, including：The load balancing table of source routing is not changed, and the server for now disconnecting still is serviced by source Device thinks reachable, and packet is dropped at purpose access switch；Or when local address is parsed, Notify that the destination server that source server end is come is unreachable.

It should be understood that in embodiments of the present invention, numbering " first ", " second " ... just to distinguishing not With object, such as in order to distinguish different interchangers, or in order to distinguish different servers, or in order to Different node sets are distinguished, the scope to the embodiment of the present invention is not construed as limiting.

Therefore, the method for the message forwarding in the software defined network SDN of the embodiment of the present invention, passes through Dynamic generation load balancing table, it is possible to increase the forward efficiency of load balancing data in data center, reduces The O＆M complexity of load balancing.

What the message in the software defined network SDN below in conjunction with Fig. 3 to the embodiment of the present invention was forwarded Method is described, and the SDN includes at least one controller and multiple servers, the method 300 by Software switch in first server in the plurality of server is performed, for example, the software switch is OVS, the method 300 includes：

S310, receive at least one controller transmission load balancing table, the load balancing table be this extremely Link-state information and the destination server of the first server of few controller according to the SDN Address generation；

S320, message forwarding is carried out according to the load balancing table.

Specifically, the software switch in first server is receiving the load balancing of controller transmission After table, message forwarding can be carried out according to the load balancing table.First server is source server.Its In, the load balancing table is by controller according to the link-state information of the SDN and the first server The address generation of destination server, had been described in above, for sake of simplicity, repeating no more.

Alternatively, the software switch in first server is carrying out message forwarding according to the load balancing table Can include：

Here, the forwarding form of the actual use of message is IPinIP encapsulation, that is, use IP tunnel Tunnel。

Current OpenFlow agreements cannot support outer layer IP encapsulation process of IP tunnel, therefore be The follow-up termination point of a tunnel treatment of support, the present invention to OpenFlow agreements has also made extended below：

PUSH_IP_TUNNEL after it have selected the address of outer envelope of above-mentioned load balancing, according to The instruction of this action, first server (can be OVS interchangers) the outer envelope address selected The encapsulation of IPinIP is carried out to message, the type of this action is 0x34, and length is 4 bytes； STRIP_IP_TUNNEL, this End point indication OVS interchangers of action in IPinIP tunnels peels outer layer off IP heads, the type of this action is 0x35.

It should be understood that here message forwarding form only by IPinIP encapsulation as a example by illustrate, or Other forwarding form, is not restricted to this.

Therefore, the method for the data message forwarding in the software defined network SDN of the embodiment of the present invention, By dynamic generation load balancing table, it is possible to increase the forward efficiency of load balancing data in data center, Reduce the O＆M complexity of load balancing.

Software defined network SDN according to embodiments of the present invention is described in detail above in association with Fig. 1 to Fig. 3 In data message forwarding method, describe according to embodiments of the present invention below in conjunction with Fig. 4 and Fig. 5 The device of the data message forwarding in software defined network SDN.

Fig. 4 shows the controller 600 in software defined network SDN according to embodiments of the present invention Schematic block diagram.As shown in figure 4, the controller 600 includes：

Acquisition module 610, for obtaining the link-state information in the SDN；

Generation module 620, link-state information and this for being obtained according to the acquisition module 610 is more The address of the destination server of the first server in individual server generates the load point of the first server Load table；

Sending module 630, the load balancing table for the generation module 620 to be generated send to this Software switch on one server, in order to the software switch in the first server according to the load Sharing table carries out message forwarding.

In acquisition module 610, link-state information can include fault-free link and three kinds of faulty links Situation.The load balancing table under these situations will be described below.

In a kind of situation, alternatively, the SDN also includes that multiple core switch and multiple convergence are handed over Change planes, the plurality of core switch is the Upper Switch of the plurality of convergence switch, Link State letter Breath includes fault-free link information,

The generation module 620 specifically for：

In a kind of situation, alternatively, the SDN also includes that multiple core switch and multiple convergence are handed over Change planes, the plurality of core switch is the Upper Switch of the plurality of convergence switch, the acquisition module is obtained The link-state information for taking includes Fisrt fault link information, and the Fisrt fault link information is the plurality of The first convergence switch in the first core switch and the plurality of convergence switch in core switch Between link failure information；

Wherein, the generation module 620 specifically for：

Alternatively, in descendants's server that the first server is first convergence switch, first remittance When poly- interchanger is in different node sets from the destination server, what the generation module 620 was generated should The core switch address list of load balancing list item for the plurality of core switch address remove this first Address list behind core switch address, wherein, it is only capable of being handed over by core between the different node set Change planes and communicated.

Alternatively, it is not descendants's server of first convergence switch, the purpose in the first server When server is descendants's server of first convergence switch,

Or, first of the Part I of the address of the destination server and the first core switch address Point difference, or the destination server address Part II and the first core switch address the 3rd Part is different, then the core switch address list of the load balancing list item of the generation module 620 generation It is the plurality of core switch address.

Alternatively, in descendants's server that the first server is first convergence switch, first remittance Poly- interchanger is in first node set together with the destination server, and first convergence switch connects with first Enter interchanger be connected, the destination server is connected with the second access switch, first access switch and When second access switch is different, then the convergence of the load balancing list item of the generation module 620 generation Switch address list is the multiple convergence switch address lists in the first node set.

In a kind of situation, alternatively, the SDN also includes that multiple core switch, multiple convergence are handed over Change planes and multiple access switch, the plurality of core switch is that the upper strata of the plurality of convergence switch exchanges Machine, the plurality of convergence switch is the Upper Switch of the plurality of access switch, and the acquisition module is obtained The link-state information include the second faulty link information, second faulty link information be the plurality of remittance The 3rd access switch in the second convergence switch and the plurality of access switch in poly- interchanger it Between link failure information；

Wherein, the generation module 620 specifically for：

Alternatively, in descendants's server that the first server is the 3rd access switch, the 3rd connects Enter interchanger and the destination server and be in Section Point set together, the destination server and the plurality of access The 4th access switch in interchanger is connected, and the 3rd access switch is with the 4th access switch not Meanwhile, the convergence switch address list of the load balancing list item of the generation module 620 generation for this Multiple convergence switch addresses in three node sets remove the address behind the second convergence switch address List.

Alternatively, in descendants's server that the first server is the 3rd access switch, the 3rd connects Enter interchanger and be in different node sets from the destination server, this of the generation module 620 generation is born Load shares the core switch address list of list item for the plurality of core switch address removes second remittance Address list behind the corresponding ancestors' core switch address of poly- interchanger, wherein, the different set of node It is only capable of being communicated by core switch between conjunction.

Alternatively, it is not descendants's server of the 3rd access switch in the first server, this first Server is in the 3rd node set together with the 3rd access switch, and the destination server connects for the 3rd When entering descendants's server of interchanger,

If or, the Part I of the address of the destination server and the first of the second convergence switch address Part is different, or the destination server address Part II and the second convergence switch address the Three parts are different, then the convergence switch address column of the load balancing list item of the generation module 620 generation Table is the multiple convergence switch address lists in the 3rd node set.

Alternatively, different node sets is in from the 3rd access switch in the first server, should When destination server is descendants's server of the 3rd access switch,

If the Part I of the address of the destination server is handed over ancestors' core of second convergence switch Change planes address Part II it is identical, and the Part II of the address of the destination server with this second converge The Part III of ancestors' core switch address of interchanger is identical, then what the generation module 620 was generated should The core switch address list of load balancing list item for the plurality of core switch address remove this second Address list behind the corresponding ancestors' core switch address of convergence switch, wherein, the different node It is only capable of being communicated by core switch between set；

If or, ancestors' core of the Part I of the address of the destination server and second convergence switch The Part II of switch address is different, or the Part II of the address of the destination server second converges with this The Part III of ancestors' core switch address of poly- interchanger is different, then the generation module 620 is generated The core switch address list of the load balancing list item is the plurality of core switch address list, its In, it is only capable of being communicated by core switch between the different node set.

In a kind of situation, alternatively, the SDN also includes multiple access switch, the plurality of access Interchanger is the Upper Switch of the plurality of server, the link-state information bag that the acquisition module is obtained Include the 3rd faulty link information, the 3rd faulty link information is the 5th connecing in the plurality of access switch Enter the information of link failure between the second server in interchanger and the plurality of server；

Wherein, the generation module 620 specifically for：

Alternatively, when the first server is the second server, what the generation module 620 was generated should The convergence switch address list and core switch address list of load balancing list item are sky.

Alternatively, it is not the second server in the first server, the destination server is second clothes Business device, the destination server is in fourth node set together with the second server,

If or, the address of the destination server is different from the address of the second server, the load balancing Core switch address in table is the plurality of core switch address, and what the generation module 620 was generated should Convergence switch address in load balancing table is the multiple convergence switches ground in the fourth node set Location.

Device 600 according to embodiments of the present invention may correspond to method 200 according to embodiments of the present invention The above and other operation and/or function of modules is respectively for reality in executive agent, and device 600 The corresponding flow of existing foregoing each method, for sake of simplicity, will not be repeated here.

Therefore, the controller of the message forwarding in the software defined network SDN of the embodiment of the present invention, leads to Cross dynamic generation load balancing table, it is possible to increase the forward efficiency of load balancing data in data center, drop The low O＆M complexity of load balancing.

Fig. 5 is shown in the first server in software defined network SDN according to embodiments of the present invention Software switch 700 schematic block diagram.The SDN includes at least one controller and multiple services Device, as shown in figure 5, the software switch 700 in first server in the plurality of server includes：

Receiver module 710, the load balancing table for receiving at least one controller transmission, the load Sharing table is link-state information and the first server of at least one controller according to the SDN The address generation of destination server；

Forwarding module 720, load balancing table for being received according to the receiver module 710 carries out message Forwarding.

Alternatively, the forwarding module 720 specifically for：

Therefore, the first server of the message forwarding in the software defined network SDN of the embodiment of the present invention On software switch, by according to controller dynamically generation load balancing table carry out message forwarding, energy The forward efficiency of load balancing data in data center is enough improved, the O＆M for reducing load balancing is complicated Degree.

Device 700 according to embodiments of the present invention may correspond to method 300 according to embodiments of the present invention The above and other operation and/or function of modules is respectively for reality in executive agent, and device 700 The corresponding flow of existing foregoing each method, for sake of simplicity, will not be repeated here.

Fig. 6 shows the controller in the software defined network SDN that another embodiment of the present invention is provided Structure, including at least one processor 702 (such as CPU), at least one network interface 705 or Person other communication interfaces, memory 706, and at least one communication bus 703, for realizing that these are filled Connection communication between putting.Processor 702 is used to perform the executable module stored in memory 706, Such as computer program.Memory 706 may include high-speed random access memory (RAM：Random Access Memory), it is also possible to also including non-labile memory (non-volatile memory), For example, at least one magnetic disk storage.(can be wired or nothing by least one network interface 705 Line) realize and the communication connection between at least one other network element.

In some embodiments, memory 706 stores program 7061, and program 7061 can be located Reason device 702 is performed, and this program includes：

Obtain the link-state information in the SDN；

The destination server of the first server in the link-state information and the plurality of server Address generates the load balancing table of the first server；

The load balancing table is sent into the software switch to the first server, in order to first clothes Software switch on business device carries out message forwarding according to the load balancing table.

Alternatively, the SDN also includes multiple core switch and multiple convergence switches, the plurality of core Heart interchanger is the Upper Switch of the plurality of convergence switch, and the link-state information includes Fisrt fault Link information, the Fisrt fault link information be the first core switch in the plurality of core switch and The information of link failure between the first convergence switch in the plurality of convergence switch；

Alternatively, in descendants's server that the first server is first convergence switch, first remittance When poly- interchanger is in different node sets from the destination server, the core of the load balancing list item is handed over Address list of changing planes is the ground after the plurality of core switch address removes the first core switch address Location list, wherein, it is only capable of being communicated by core switch between the different node set.

Alternatively, in descendants's server that the first server is first convergence switch, first remittance Poly- interchanger is in first node set together with the destination server, and first convergence switch connects with first Enter interchanger be connected, the destination server is connected with the second access switch, first access switch and When second access switch is different, then the convergence switch address list of the load balancing list item for this Multiple convergence switch address lists in one node set.

Alternatively, the SDN also includes multiple core switch, multiple convergence switches and multiple accesses Interchanger, the plurality of core switch is the Upper Switch of the plurality of convergence switch, the plurality of convergence Interchanger is the Upper Switch of the plurality of access switch, and the link-state information includes the second failure chain Road information, second faulty link information is the second convergence switch in the plurality of convergence switch and should The information of link failure between the 3rd access switch in multiple access switch；

Alternatively, in descendants's server that the first server is the 3rd access switch, the 3rd connects Enter interchanger and the destination server and be in Section Point set together, the destination server and the plurality of access The 4th access switch in interchanger is connected, and the 3rd access switch is with the 4th access switch not Meanwhile, the convergence switch address list of the load balancing list item is the multiple remittances in the Section Point set Poly- switch address removes the address list behind the second convergence switch address.

Alternatively, in descendants's server that the first server is the 3rd access switch, the 3rd connects Enter interchanger and be in different node sets from the destination server, the core of the load balancing list item is exchanged Machine address list removes the corresponding ancestors' core of second convergence switch for the plurality of core switch address Address list after heart switch address, wherein, it is only capable of being exchanged by core between the different node set Machine is communicated.

Alternatively, the SDN also includes multiple access switch, and the plurality of access switch is the plurality of The Upper Switch of server, the link-state information includes the 3rd faulty link information, the 3rd failure Link information is in the 5th access switch and the plurality of server in the plurality of access switch The information of link failure between two servers；

Alternatively, when the first server is the second server, the convergence of the load balancing list item is handed over Change planes address list and core switch address list is sky.

Therefore, the controller of the data message forwarding in the software defined network SDN of the embodiment of the present invention, By dynamic generation load balancing table, it is possible to increase the forward efficiency of load balancing data in data center, Reduce the O＆M complexity of load balancing.

Fig. 7 shows the first clothes in the software defined network SDN that another embodiment of the present invention is provided The structure of the software switch on business device, including at least one processor 802 (such as CPU), at least One network interface 805 or other communication interfaces, memory 806, and at least one communication bus 803, For realizing the connection communication between these devices.Processor 802 is used to perform storage in memory 806 Executable module, such as computer program.Memory 806 may include high-speed random access memory (RAM：Random Access Memory), it is also possible to also including non-labile memory (non-volatile memory), for example, at least one magnetic disk storage.By at least one network interface 805 (can be wired or wireless) are realized and the communication connection between at least one other network element.

In some embodiments, memory 806 stores program 8061, and program 8061 can be located Reason device 802 is performed, and this program includes：

Message forwarding is carried out according to the load balancing table.

Therefore, the software in the first server in the software defined network SDN of the embodiment of the present invention is handed over Change planes, message forwarding is carried out by the load balancing table dynamically generated according to controller, it is possible to increase data The forward efficiency of load balancing data in center, reduces the O＆M complexity of load balancing.

It should be understood that the terms "and/or", a kind of only incidence relation for describing affiliated partner, Expression there may be three kinds of relations, for example, A and/or B, can represent：Individualism A, while depositing In A and B, individualism B these three situations.In addition, character "/" herein, before and after typicallying represent Affiliated partner is a kind of relation of "or".

It should be understood that in various embodiments of the present invention, the size of the sequence number of above-mentioned each process is not intended to The priority of execution sequence, the execution sequence of each process should be determined with its function and internal logic, without answering Implementation process to the embodiment of the present invention constitutes any restriction.

Those of ordinary skill in the art are it is to be appreciated that with reference to each of the embodiments described herein description The unit and algorithm steps of example, can be with electronic hardware or the knot of computer software and electronic hardware Close to realize.These functions are performed with hardware or software mode actually, depending on the spy of technical scheme Fixed application and design constraint.Professional and technical personnel can use not Tongfang to each specific application Method realizes described function, but this realization it is not considered that beyond the scope of this invention.

It is apparent to those skilled in the art that, it is for convenience and simplicity of description, above-mentioned to retouch The specific work process of system, device and the unit stated, may be referred to the correspondence in preceding method embodiment Process, will not be repeated here.

In several embodiments provided herein, it should be understood that disclosed system, device and Method, can realize by another way.For example, device embodiment described above is only to show Meaning property, for example, the division of the unit, only a kind of division of logic function can when actually realizing To there is other dividing mode, such as multiple units or component can be combined or be desirably integrated into another System, or some features can be ignored, or not perform.It is another, it is shown or discussed each other Coupling or direct-coupling or communication connection can be the INDIRECT COUPLING of device or unit by some interfaces Or communication connection, can be electrical, mechanical or other forms.

It is described as separating component illustrate unit can be or may not be it is physically separate, make For the part that unit shows can be or may not be physical location, you can with positioned at a place, Or can also be distributed on multiple NEs.Can select according to the actual needs part therein or Person whole units realize the purpose of this embodiment scheme.

In addition, each functional unit in each embodiment of the invention can be integrated in a processing unit In, or unit be individually physically present, it is also possible to two or more units are integrated in one In individual unit.

If the function is to realize in the form of SFU software functional unit and as independent production marketing or make Used time, can store in a computer read/write memory medium.Based on such understanding, the present invention The part that is substantially contributed to prior art in other words of technical scheme or the technical scheme portion Dividing can be embodied in the form of software product, and the computer software product is stored in a storage medium In, including some instructions are used to so that computer equipment (can be personal computer, server, Or the network equipment etc.) perform all or part of step of each embodiment methods described of the invention.And it is preceding The storage medium stated includes：USB flash disk, mobile hard disk, read-only storage (ROM, Read-Only Memory), Random access memory (RAM, Random Access Memory), magnetic disc or CD etc. are various can With the medium of store program codes.

The above, specific embodiment only of the invention, but protection scope of the present invention do not limit to In this, any one skilled in the art the invention discloses technical scope in, can be easily Expect change or replacement, should all be included within the scope of the present invention.Therefore, protection of the invention Scope should be based on the protection scope of the described claims.

Claims

1. the method that the message in a kind of software defined network SDN is forwarded, it is characterised in that the SDN Including at least one controller and multiple servers, methods described is performed by least one controller, Methods described includes：

Obtain the link-state information in the SDN；

The purpose service of the first server in the link-state information and the multiple server The address of device generates the load balancing table of the first server；

The load balancing table is sent into the software switch to the first server, in order to described Software switch in first server carries out message forwarding according to the load balancing table.

2. method according to claim 1, it is characterised in that the SDN also includes multiple cores Interchanger and multiple convergence switches, the multiple core switch is the upper of the multiple convergence switch Layer switch, the link-state information includes fault-free link information, described according to the Link State The address generation described first of the destination server of the first server in information and the multiple server The load balancing table of server, including：

Address according to the fault-free link information and the destination server generates the load balancing Load balancing list item in table, wherein, the core switch address list of the load balancing list item is institute Multiple core switch address lists are stated, the convergence switch address list of the load balancing list item is institute State the multiple convergence switch address lists in the node set residing for first server.

3. method according to claim 1, it is characterised in that the SDN also includes multiple cores Interchanger and multiple convergence switches, the multiple core switch is the upper of the multiple convergence switch Layer switch, the link-state information includes Fisrt fault link information, the Fisrt fault link letter Cease in the first core switch and the multiple convergence switch in the multiple core switch The information of link failure between first convergence switch；

Wherein, the first server in the link-state information and the multiple server The address of destination server generates the load balancing table of the first server, including：

The address generation load point according to the Fisrt fault link information and the destination server Load balancing list item in load table.

4. method according to claim 3, it is characterised in that in the first server be described Descendants's server of the first convergence switch, first convergence switch is in the destination server During different node set, the core switch address list of the load balancing list item is the multiple core Heart switch address removes the address list behind the first core switch address, wherein, the difference Node set between be only capable of being communicated by core switch.

5. method according to claim 3, it is characterised in that in the first server be not institute Descendants's server of the first convergence switch is stated, the destination server is first convergence switch During descendants's server,

If the Part I of the address of the destination server and the of the first core switch address It is a part of identical, and the Part II of the address of the destination server and first core switch ground The Part III of location is identical, then the core switch address list of the load balancing list item is the multiple Core switch address removes the address list behind the first core switch address；

Or, the Part I of the address of the destination server and the of the first core switch address It is a part of different, or the address of the destination server Part II and first core switch ground The Part III of location is different, then the core switch address list of the load balancing list item is the multiple Core switch address.

6. method according to claim 3, it is characterised in that in the first server be described Descendants's server of the first convergence switch, first convergence switch exists together with the destination server In first node set, first convergence switch is connected with the first access switch, the purpose clothes Business device is connected with the second access switch, and first access switch is with second access switch not Meanwhile, the convergence switch address list of the load balancing list item is many in the first node set Individual convergence switch address list.

7. method according to claim 1, it is characterised in that the SDN also includes multiple cores Interchanger, multiple convergence switches and multiple access switch, the multiple core switch is described many The Upper Switch of individual convergence switch, the multiple convergence switch is the multiple access switch Upper Switch, the link-state information includes the second faulty link information, second faulty link During information is the second convergence switch and the multiple access switch in the multiple convergence switch The 3rd access switch between link failure information；

The address generation load point according to second faulty link information and the destination server Load balancing list item in load table.

8. method according to claim 7, it is characterised in that in the first server be described Descendants's server of the 3rd access switch, the 3rd access switch exists together with the destination server In Section Point set, the destination server is exchanged with the 4th access in the multiple access switch Machine is connected, when the 3rd access switch is different from the 4th access switch, the load balancing The convergence switch address list of list item is the multiple convergence switch addresses in the Section Point set Remove the address list behind the second convergence switch address.

9. method according to claim 7, it is characterised in that in the first server be described Descendants's server of the 3rd access switch, the 3rd access switch is in the destination server Different node sets, the core switch address list of the load balancing list item is the multiple core Switch address removes the address behind the corresponding ancestors' core switch address of second convergence switch List, wherein, it is only capable of being communicated by core switch between the different node set.

10. method according to claim 7, it is characterised in that in the first server be not institute Descendants's server of the 3rd access switch is stated, the first server is same with the 3rd access switch In the 3rd node set, the destination server is descendants's server of the 3rd access switch When,

If the Part I of the address of the destination server and the of the second convergence switch address It is a part of identical, and the Part II of the address of the destination server and second convergence switch ground The Part III of location is identical, then the convergence switch address list of the load balancing list item is the described 3rd Multiple convergence switch addresses in node set remove the address behind the second convergence switch address List；

If or, the Part I of the address of the destination server and the second convergence switch address Part I is different, or the address of the destination server Part II and second convergence switch The Part III of address is different, then the convergence switch address list of the load balancing list item is described the Multiple convergence switch address lists in three node sets.

11. methods according to claim 7, it is characterised in that the first server with it is described 3rd access switch is in different node sets, and the destination server is that the described 3rd access is exchanged During descendants's server of machine,

If ancestors' core of the Part I of the address of the destination server and second convergence switch The Part II of heart switch address is identical, and the Part II of the address of the destination server with it is described The Part III of ancestors' core switch address of the second convergence switch is identical, then the load balancing table The core switch address list of item is converged for the multiple core switch address removal described second and handed over Change planes the address list behind corresponding ancestors' core switch address, wherein, the different node set Between be only capable of being communicated by core switch；

12. methods according to claim 1, it is characterised in that the SDN also connects including multiple Enter interchanger, the multiple access switch is the Upper Switch of the multiple server, the link Status information includes the 3rd faulty link information, and the 3rd faulty link information is handed over for the multiple access Link failure between second server in the 5th access switch changed planes and the multiple server Information；

The address generation load point according to the 3rd faulty link information and the destination server Load balancing list item in load table.

13. methods according to claim 12, it is characterised in that in the first server be institute When stating second server, convergence switch address list and the core switch ground of the load balancing list item Location list is sky.

14. methods according to claim 12, it is characterised in that be not in the first server The second server, the destination server is the second server, the destination server and institute State second server and be in fourth node set together,

If the address of the destination server is identical with the address of the second server, the load point The convergence switch address list and core switch address list for carrying on a shoulder pole list item are sky；

If described negative or, the address of the destination server is different from the address of the second server It is the multiple core switch address, the load balancing table to carry the core switch address in sharing table In convergence switch address be multiple convergence switch addresses in the fourth node set.

The method of the message forwarding in a kind of 15. software defined network SDN, it is characterised in that the SDN Including at least one controller and multiple servers, methods described is by the first clothes in the multiple server Software switch on business device is performed, including：

Receive the load balancing table that at least one controller sends, the load balancing table be it is described extremely Purpose service of few controller according to the link-state information and the first server of the SDN The address generation of device；

Message forwarding is carried out according to the load balancing table.

16. methods according to claim 15, it is characterised in that described according to the load balancing Table carries out message forwarding to be included：

Controller in a kind of 17. software defined network SDN, it is characterised in that the SDN is included extremely A few controller and multiple servers, including：

Acquisition module, for obtaining the link-state information in the SDN；

Generation module, for the link-state information that is obtained according to the acquisition module and the multiple The address of the destination server of the first server in server generates the load point of the first server Load table；

Sending module, the load balancing table for the generation module to be generated is sent to described first Software switch on server, in order to the software switch in the first server according to described negative Carrying sharing table carries out message forwarding.

18. controllers according to claim 17, the SDN also include multiple core switch and Multiple convergence switches, the multiple core switch is that the upper strata of the multiple convergence switch exchanges Machine, the link-state information includes fault-free link information,

The generation module specifically for：

19. controllers according to claim 17, it is characterised in that the SDN also includes multiple Core switch and multiple convergence switches, the multiple core switch is the multiple convergence switch Upper Switch, the link-state information that the acquisition module is obtained includes Fisrt fault link letter Breath, the Fisrt fault link information is the first core switch and institute in the multiple core switch State the information of link failure between the first convergence switch in multiple convergence switches；

Wherein, the generation module specifically for：

20. controllers according to claim 19, it is characterised in that be in the first server Descendants's server of first convergence switch, first convergence switch and the destination server During in different node sets, the core of the load balancing list item of the generation module generation is exchanged Machine address list is after the multiple core switch address removes the first core switch address Address list, wherein, it is only capable of being communicated by core switch between the different node set.

21. controllers according to claim 19, it is characterised in that in the first server not It is descendants's server of first convergence switch, the destination server is that first convergence is exchanged During descendants's server of machine,

22. controllers according to claim 19, it is characterised in that be in the first server Descendants's server of first convergence switch, first convergence switch and the destination server It is in first node set together, first convergence switch is connected with the first access switch, the mesh Server be connected with the second access switch, first access switch with described second access exchange When machine is different, the convergence switch address list of the load balancing list item is in the first node set Multiple convergence switch address lists.

23. controllers according to claim 17, it is characterised in that the SDN also includes multiple Core switch, multiple convergence switches and multiple access switch, the multiple core switch is institute The Upper Switch of multiple convergence switches is stated, the multiple convergence switch is that the multiple access exchanges The Upper Switch of machine, the link-state information that the acquisition module is obtained includes the second faulty link Information, second faulty link information be the second convergence switch in the multiple convergence switch and The information of link failure between the 3rd access switch in the multiple access switch；

Wherein, the generation module specifically for：

24. controllers according to claim 23, it is characterised in that be in the first server Descendants's server of the 3rd access switch, the 3rd access switch and the destination server It is in Section Point set together, the destination server is accessed with the 4th in the multiple access switch Interchanger is connected, when the 3rd access switch is different from the 4th access switch, the generation The convergence switch address list of the load balancing list item of module generation is the 3rd node set In multiple convergence switch addresses remove the address list behind the second convergence switch address.

25. controllers according to claim 23, it is characterised in that be in the first server Descendants's server of the 3rd access switch, the 3rd access switch and the destination server In different node sets, the core switch of the load balancing list item of the generation module generation Address list is that the multiple core switch address removes the corresponding ancestors of second convergence switch Address list behind core switch address, wherein, it is only capable of by core between the different node set Interchanger is communicated.

26. controllers according to claim 23, it is characterised in that in the first server not It is descendants's server of the 3rd access switch, the first server is accessed with the described 3rd and exchanged Machine is in the 3rd node set together, and the destination server is descendants's service of the 3rd access switch During device,

27. controllers according to claim 23, it is characterised in that the first server with 3rd access switch is in different node sets, and the destination server is the described 3rd access During descendants's server of interchanger,

28. controllers according to claim 17, it is characterised in that the SDN also includes multiple Access switch, the multiple access switch is the Upper Switch of the multiple server, described to obtain The link-state information that modulus block is obtained includes the 3rd faulty link information, the 3rd faulty link Information is in the 5th access switch and the multiple server in the multiple access switch The information of link failure between two servers；

Wherein, the generation module specifically for：

29. controllers according to claim 28, it is characterised in that be in the first server During the second server, the convergence switch ground of the load balancing list item of the generation module generation Location list and core switch address list are sky.

30. controllers according to claim 28, it is characterised in that in the first server not The second server, the destination server is the second server, the destination server with The second server is in fourth node set together,

The software switch in first server in a kind of 31. software defined network SDN, its feature exists In the SDN includes at least one controller and multiple servers, the institute in the multiple server The software switch stated in first server includes：

Receiver module, for receiving the load balancing table that at least one controller sends, the load Sharing table is link-state information and first clothes of at least one controller according to the SDN The address generation of the destination server of business device；

Forwarding module, the load balancing table for being received according to the receiver module carries out message and turns Hair.

32. software switch according to claim 31, it is characterised in that the forwarding module tool Body is used for：