CN104009903A - Flow forwarding method and device for RPR network - Google Patents
Flow forwarding method and device for RPR network Download PDFInfo
- Publication number
- CN104009903A CN104009903A CN201410247931.1A CN201410247931A CN104009903A CN 104009903 A CN104009903 A CN 104009903A CN 201410247931 A CN201410247931 A CN 201410247931A CN 104009903 A CN104009903 A CN 104009903A
- Authority
- CN
- China
- Prior art keywords
- rpr
- message
- node
- port
- arp
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Small-Scale Networks (AREA)
Abstract
The invention provides a flow forwarding method and device for an RPR network. The method comprises the steps that each node in the RPR network is composed of stacking n switches, and each switch provides an east physical port and a west physical port; for each node, the n east physical ports of the n switches of the node are aggregated, the n west physical ports of the n switches of the node are aggregated, and then two RPR physical aggregated ports, namely, the east physical aggregated port and the west physical aggregated port are obtained and are bound to an RPR logic port, wherein n is an integer larger than or equal to 2; the nodes in the RPR network share loads at the RPR physical aggregated ports. The flow forwarding method and device for the RPR network increase the bandwidth of the RPR logic port and improve the reliability of the RPR.
Description
Technical field
The present invention relates to RPR (Resilient Packet Ring, Resilient Packet Ring) technical field, relate in particular to flow forwarding method and device in RPR looped network.
Background technology
RPR (Resilient Packet Ring, Resilient Packet Ring) be a kind of novel MAC (Media Access Control, media interviews control) agreement, can run on SONET (Synchronous Optical Network, Synchronous Optical Network)/SDH (Synchronous Digital Hierarchy, synchronous digital hierarchy), DWDM (Dense Wavelength Division Multiplexing, level of confidentiality wavelength division multiplexing) and Ethernet on, for broadband IP metropolitan area network, operator provides flexible and efficient networking plan.
RPR technology designs in order to support jumbo data service in metropolitan area network, has following characteristics:
Physical layer diversity;
Bandwidth availability ratio is high;
Support broadcast and multicast;
Topological automatic discovering, the plug and play of support node;
Protection mechanism, can realize the fault self-recovery in 50ms by topology protection fast;
Provide flow grade to ensure by tenaculum bandwidth reservation business and rate limit;
Fair node bandwidth is distributed.
RPR adopts the encapsulation of RPR mac layer frame to realize the transparent transmission of Ethernet Over RPR.The ring structure of RPR and topological protection mechanism are transparent to repeating process and the access device of institute's bearer traffic.
RPR adopts reverse twin nuclei, and data forward between node along looped network, as shown in Figure 1.The basic conception relating in RPR is as follows:
0 ring: in RPR dicyclo, Frame sending direction is clockwise 0 ring that is called, and also claims Outer Ring (outer shroud);
1 ring: in RPR dicyclo, Frame sending direction is counterclockwise 1 ring that is called, and also claims Inner Ring (interior ring);
Node (Station): the equipment on RPR looped network, be responsible for receiving and forwarding data frame;
Link (Link): connect one section of transmission channel of adjacent node, between adjacent node by two link connections of opposite direction;
Section (Span): the link on RPR looped network between two adjacent nodes, is made up of two links of opposite direction;
Territory (Domain): the node on multiple continuous section and these sections has formed territory;
West is to port: on 0 ring receiving data frames, on 1 ring, send the physical port of Frame;
East orientation port: send Frame on 0 ring, the physical port of receiving data frames on 1 ring;
Limit (Edge): while breaking down when section or with the adjacent node of section, section can not just become limit by forwarding data;
Ring status: be divided into closed loop and open loop.Not having the ring on limit is closed loop, and the ring that has limit is open loop;
RPR logic interface, two physical ports (east orientation and west are to port) that on RPR ring, website is linked into RPR ring have formed RPR logic interface.
Rapidly, each equipment vendors have released the Ethernet interface of 40G, 100G to the development of current ethernet technology successively, and before the standard mesh of RPR, still can only support 10G port, in competition in unfavorable position.In addition, RPR networking is generally very high to reliability requirement, also in the urgent need to further development.
Fig. 2 has provided existing a kind of RPR polymerization looped network schematic diagram, as shown in Figure 2, in this RPR polymerization looped network, comprises four node S1, S2, S3, S4.Each node is made up of two exchange pilings respectively, configures two RPR logic interfaces on each node, and each logic interface accesses respectively the RPR logic box of an independent operating RPR agreement, and two separate RPR logic boxs make to coexist in looped network.These two RPR logic boxs are carried out to link aggregation, carry out load balancing with the flow on each node.
Taking node S1 as example, to establish two RPR logic interface and be respectively rpr1, rpr2, three stratas that this two logic interface adds are healed up for ragg1024:
1) in the time that S1 receives the Ethernet message that hangs preparation under self, ARP (the Address Resolution Protocol finding, address resolution protocol) outgoing interface that list item is corresponding is ragg1024, S1 shares flow according to stacking local first and load balancing Hash (Hash) algorithm, message can be shared one of them RPR logic interface, as rpr1, now, trigger again the ring selection logic of rpr1, according to ring selection table, flow is packaged into RPR Frame, and pressed on ring forwards.
2) suppose on looped network that in forward-path, a certain section is broken down, the topology of looped network meeting fast refresh, and again complete the routing of looped network, refresh ring selection table, now can exert an influence in the trend of this looped network to flow, but the load balancing hash algorithm on each node does not need to upgrade, and flow still can correctly arrive destination, and may there be of short duration packet loss centre because of change in topology.
3) suppose that on looped network, certain node breaks down, this node is not link point under the target of above-mentioned message, and the influenced situation of message is substantially with 2) identical.
Summary of the invention
The invention provides flow forwarding method and device in RPR looped network, to improve the bandwidth of RPR logic interface, strengthen the reliability of RPR.
Technical scheme of the present invention is achieved in that
A flow forwarding method in RPR looped network, the each node in this RPR looped network is made up of n platform exchange piling respectively, and every switch provides an east orientation physical port and a west to physical port; For each node, n east orientation physical port of the n platform switch that this node is comprised and n west are to physical port polymerization respectively, obtain two RPR physics aggregation ports: east orientation physics aggregation port and west are to physics aggregation port, these two RPR physics aggregation ports are tied to RPR logic interface, n is integer, and n >=2, the method comprises:
Node in RPR looped network carries out load balancing on RPR physics aggregation port.
Node in described RPR looped network carries out load balancing and comprises on RPR physics aggregation port:
Node in RPR looped network, by Topology Discovery, is set up RPR ring selection table, and each list item in RPR ring selection table all comprises: the MAC Address of target looped network node, the corresponding ring of optimal path from this node to this target looped network node;
In the time that described node receives the Ethernet message that hangs preparation under self, in the ARP table of self, search corresponding ARP list item according to the object IP address of message, obtain the outgoing interface of message, if the outgoing interface of this message is the MAC Address of the non-RPR looped network node in object media access control MAC address of RPR logic interface and message, in RPRMAC list item, find the MAC Address of corresponding target looped network node according to the target MAC (Media Access Control) address of message, in RPR ring selection table, find corresponding ring according to the MAC Address of this target looped network node, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded, wherein, RPR MAC address entries comprises: the MAC Address of RPR looped network node and hang the incidence relation of standby MAC Address under it.
After the described outgoing interface that obtains message, further comprise:
If the target MAC (Media Access Control) address that the outgoing interface of this message is RPR logic interface and message is the MAC Address of RPR looped network node, in RPR ring selection table, find corresponding ring according to the target MAC (Media Access Control) address of this message, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded.
Describedly go out in n the physical port that physics aggregation port comprises to select one to comprise at this:
For n the physical port that goes out physics aggregation port and comprise, be only that all physical ports of choosing adopt default load balancing algorithm to select to state.
Describedly when receiving after hanging the Ethernet message that preparation comes under self, described node further comprises:
When described node does not find corresponding ARP list item according to the object IP address of this Ethernet message in the ARP table of self, with the object IP address architecture ARP request message of this message, this ARP request message is sent; In the time receiving arp response message, generate ARP list item, or ARP list item+RPR MAC address entries;
According to the ARP list item generating, or ARP list item+RPR MAC address entries forwards described Ethernet message.
Described generation ARP list item, or ARP list item+RPR MAC address entries comprises:
When arp response message is from hanging under this node when standby, only generate ARP list item;
When arp response message is during from other node of RRP looped network, only generate ARP list item, wherein, the outgoing interface in ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding;
When arp response message hangs under from other node of RRP looped network when standby, generate ARP list item and RPR MAC address entries, wherein, the outgoing interface of ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding, in RPR MAC address entries, comprises: the source MAC of the encapsulation RPR MAC head of the source MAC of arp response message and arp response message.
Node in described RPR looped network carries out load balancing and further comprises on RPR physics aggregation port:
In the time that described node receives RPR Frame from a RPR physics aggregation port, if find, the target MAC (Media Access Control) address of the RPR MAC head of this Frame does not point to this node, according to default load balancing algorithm, in n the physical port comprising at another RPR physics aggregation port of self, select one, from the port of selecting, this RPR Frame is forwarded.
Described method further comprises:
In the time that described node is found the arbitrary RPR physics aggregation port Down of self, trigger RPR and switch fast.
A node in RPR looped network, the each node in described RPR looped network is made up of n platform exchange piling respectively, and every switch provides an east orientation physical port and a west to physical port; For each node, n east orientation physical port of the n platform switch that this node is comprised and n west are to physical port polymerization respectively, obtain two RPR physics aggregation ports: east orientation physics aggregation port and west are to physics aggregation port, these two RPR physics aggregation ports are tied to RPR logic interface, n is integer, and n >=2, this node comprises:
Traffic forwarding module: carry out load balancing on RPR physics aggregation port.
Described traffic forwarding module comprises:
Ring selection table learning module: by Topology Discovery, set up RPR ring selection table, each list item in RPR ring selection table all comprises: the MAC Address of target looped network node, the corresponding ring of optimal path from this node to this target looped network node;
Forwarding module: in the time receiving the Ethernet message that hangs preparation under this node, in the ARP table of this node, search corresponding ARP list item according to the object IP address of message, obtain the outgoing interface of message, if the outgoing interface of this message is the MAC Address of the non-RPR looped network node in object media access control MAC address of RPR logic interface and message, in RPR MAC address entries, find the MAC Address of corresponding target looped network node according to the target MAC (Media Access Control) address of message, in RPR ring selection table, find corresponding ring according to the MAC Address of this target looped network node, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded, wherein, RPR MAC address entries comprises: the MAC Address of RPR looped network node and hang the incidence relation of standby MAC Address under it.
The outgoing interface that described forwarding module obtains message is further used for afterwards,
If the target MAC (Media Access Control) address that the outgoing interface of this message is RPR logic interface and message is the MAC Address of RPR looped network node, in RPR ring selection table, find corresponding ring according to the target MAC (Media Access Control) address of this message, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded.
Described forwarding module goes out in n the physical port that physics aggregation port comprises to select one to comprise at this:
For n the physical port that goes out physics aggregation port and comprise, be only that all physical ports of choosing adopt default load balancing algorithm to select to state.
Described forwarding module is further used for after receiving down and hanging the next Ethernet message of preparation,
While not finding corresponding ARP list item according to the object IP address of this Ethernet message in the ARP of this node table, with the object IP address architecture ARP request message of this message, this ARP request message is sent; In the time receiving arp response message, generate ARP list item or ARP list item+RPRMAC list item, and forward described Ethernet message according to the ARP list item or the ARP list item+RPR MAC address entries that generate.
Described forwarding module generates ARP list item or ARP list item+RPR MAC address entries comprises:
When arp response message is from hanging under this node when standby, only generate ARP list item;
When arp response message is during from other node of RRP looped network, only generate ARP list item, wherein, the outgoing interface in ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding;
When arp response message hangs under from other node of RRP looped network when standby, generate ARP list item and RPR MAC address entries, wherein, the outgoing interface of ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding, in RPR MAC address entries, comprises: the source MAC of the encapsulation RPR MAC head of the source MAC of arp response message and arp response message.
Described traffic forwarding module is further used for,
In the time receiving RPR Frame from a RPR physics aggregation port of this node, if find, the target MAC (Media Access Control) address of the RPR MAC head of this Frame does not point to this node, according to default load balancing algorithm, in n the physical port comprising at another RPR physics aggregation port of this node, select one, from the port of selecting, this RPR Frame is forwarded.
Visible, the present invention, by by all east orientation physical ports in the stacked group of looped network node and western to physical port polymerization respectively, has improved the bandwidth of RPR logic interface, has strengthened the reliability of RPR.
Brief description of the drawings
Fig. 1 is the schematic diagram of existing RPR;
Fig. 2 is existing RPR polymerization looped network schematic diagram;
Flow forwarding method flow chart in the RPR looped network that Fig. 3 provides for the embodiment of the present invention;
Fig. 4 is the RPR looped network schematic diagram of application example of the present invention;
The composition schematic diagram of the RPR looped network node that Fig. 5 provides for the embodiment of the present invention.
Embodiment
Still taking Fig. 2 as example, suppose to hang under S1 link point failure under the target of the message that preparation comes, now cannot arrive destination by ring corresponding to rpr1, need to be switched to the ring that rpr2 is corresponding.But on S1, select rpr1 or rpr2 to be determined by load balancing hash algorithm completely, cannot intervene adjustment for named place of destination location.
Visible, existing RPR standard is only supported 10G bandwidth, does not have the standard of upgrading can support higher speed.The technology of existing expansion RPR bandwidth is the polymerization technique based on RPR logic interface, this technology can promote the bandwidth of RPR, but while selecting forward-path due to aggregated links, can only rely on hash algorithm, in the time of node failure on one of them RPR ring, the flow of sharing this node by hash algorithm cannot be switched to separately another RPR ring fast, and use scenes is limited.
Flow forwarding method flow chart in the RPR looped network that Fig. 3 provides for the embodiment of the present invention, its concrete steps are as follows:
Step 301: build RPR looped network, the each node in this RPR looped network is respectively by n (n is integer, and n >=2) platform exchange piling composition, and every switch provides an east orientation physical port and a west to physical port; For each node, n east orientation physical port of the n platform switch that this node is comprised and n west are to physical port polymerization respectively, obtain two RPR physics aggregation ports: east orientation physics aggregation port and west, to physics aggregation port, are tied to RPR logic interface by these two RPR physics aggregation ports.
Node in step 302:RPR looped network carries out load balancing on RPR physics aggregation port.
In the specific implementation, step 302 can comprise:
Node in step 3021:RPR looped network, by Topology Discovery, is set up RPR ring selection table.
The process of establishing of RRP ring selection table is same as the prior art.
Each list item in RPR ring selection table all comprises: the MAC Address of target looped network node, the corresponding ring of optimal path from this node to this target looped network node (0 ring or 1 ring).
Wherein, target looped network node is arbitrary other node on RPR looped network.For the arbitrary node in RPR looped network, need to set up a ring selection list item for other the each node on looped network.
By Topology Discovery, the each node in RPR looped network can be recognized the jumping figure of the complete structure of looped network, each this node of nodal distance and ability that each node possesses etc., thereby learns the optimal path of this node to other each node.RPR Topology Discovery is a kind of cyclic activity, also can be initiated by the some nodes that need to know topological structure
Step 3022: in the time that the node in RPR looped network receives the Ethernet message that hangs preparation under self, in the ARP table of self, search corresponding ARP list item according to the object IP address of message, obtain the outgoing interface of message, if the outgoing interface of this message is the MAC Address of the non-RPR looped network node of the target MAC (Media Access Control) address of RPR logic interface and message, node finds the MAC Address of corresponding target looped network node in RPRMAC list item according to the target MAC (Media Access Control) address of message, in RPR ring selection table, find corresponding ring according to the MAC Address of this target looped network node, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded.
If the outgoing interface of Ethernet message is physical port, node is directly gone out this Ethernet message repeating from this outgoing interface; If the outgoing interface of message is RPR logic interface, and the target MAC (Media Access Control) address of message is the MAC Address of a certain node on RPR looped network, in RPR ring selection table, find corresponding ring according to the target MAC (Media Access Control) address of this message, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded.
Because each node has a RPR logic interface, each RPR logic interface has been bound two RPR physics aggregation ports: east orientation physics aggregation port and west are to physics aggregation port, therefore, after having determined that RPR ring is 0 ring or 1 ring, just can determine that the RPR physics aggregation port that of message is: east orientation physics aggregation port or western to physics aggregation port.
The generative process of ARP list item and RPR MAC address entries is as follows:
For this step 3022, in the time that node does not find corresponding ARP list item according to the object IP address that hangs the next Ethernet message of preparation under self in the ARP of self shows, be handled as follows:
Step a: with the object IP address architecture ARP request message of this message, this message, to spot broadcasting, is packaged into RPR Frame by this message simultaneously and broadcasts in RPR looped network; In the time receiving arp response message, generate ARP list item, or ARP list item+RPR MAC address entries.
Wherein, IP, MAC Address that the source IP of ARP request message, MAC Address are this node, object IP address hangs the object IP address of the next Ethernet message of preparation under being, and target MAC (Media Access Control) address is broadcast address; The MAC Address that the source MAC of the RPR MAC head of ARP request message encapsulation is this node, target MAC (Media Access Control) address is broadcast address, arbitrary looped network node receives that this Frame can be carried out in looped network to ring after this RPR Frame forwards, simultaneously to after this RPR Frame decapsulation in spot broadcasting.
Wherein, the IP address in the ARP list item of generation is the source IP address of arp response message, and MAC Address is the source MAC of arp response message, and outgoing interface is the incoming interface of arp response message.
It should be noted that, arp response message has following three kinds of sources, and for different sources, the processing of node is not identical, specific as follows:
1) may be that this node hangs preparation under other and comes, now, only need to generate ARP list item, generating mode be same as the prior art, and wherein outgoing interface is the physical port of lower this equipment of extension, is non-RPR logic interface.
2) may be that on RPR looped network, other node is sent, now, arp response message has encapsulated RPR MAC head, now also only need to generate ARP list item, but the outgoing interface in ARP list item should be encapsulated arp response message RPR Frame enter RPR logic interface, that is, receiving after the RPR Frame that has encapsulated arp response message, should first the inbound port of message be converted to corresponding RPR logic interface and put into again ARP list item;
3) on may RPR looped network, under other nodes, hang that preparation comes, now, arp response message has also encapsulated RPR MAC head, now, need to generate ARP list item and RPR MAC address entries simultaneously, wherein, the outgoing interface in ARP list item should be also arp response message enter RPR logic interface; In RPR MAC address entries, comprise: the source MAC of the encapsulation RPR MAC head of the source MAC of arp response message and arp response message, what RPR MAC address entries represented is the MAC Address of RPR looped network node and the incidence relation that hangs standby MAC Address under it.
Step b: this node is according to the ARP list item generating, or ARP list item+RPR MAC address entries forwards this Ethernet message.
Wherein, if only generated ARP list item, repeating process is as follows:
If the outgoing interface of ARP list item is physical port, directly forward this Ethernet message from this physical port; If outgoing interface is RPR logic interface, in RPR ring selection table, find corresponding ring according to the target MAC (Media Access Control) address of this message, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded.
If generated ARP list item and RPR MAC address entries simultaneously, repeating process is as follows:
Now, the outgoing interface in ARP list item must be RPR logic interface.Node first finds corresponding ring according to the MAC Address of the looped network node in RPR MAC address entries in RPR ring selection table, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded.
Consider, for each looped network node, the physical port possibility Down (fault) in two RPR physics aggregation ports in the RPR logic interface of this node, the embodiment of the present invention has provided following solution:
1) node is according to default load balancing algorithm, going out while selecting one in n the physical port that physics aggregation port comprises, consideration state is the physical port of choosing, be only, that all physical ports of choosing adopt default load balancing algorithm to select to going out state in physics aggregation port; Wherein, only have and meet following condition when physical port: when the attribute configuration of the attribute configuration of port Up, port and physics aggregation port is consistent etc., the state of this physical port could be " choosing ".
2), in the time that node is found the arbitrary RPR physics aggregation port Down of self, trigger RPR and switch fast.
RPR physics aggregation port Down, i.e. n in this RPR physics aggregation port (east orientation or west to) the whole Down of physical port.
In addition, in the embodiment of the present invention, ring excessively and the lower ring of RPR Frame are processed as follows respectively:
One) crossing ring processes
In the time that node receives RPR Frame from a RPR physics aggregation port, if find, the target MAC (Media Access Control) address of the RPR MAC head of this Frame does not point to this node, according to default load balancing algorithm, in n the physical port comprising at another RPR physics aggregation port of self, select one, from the port of selecting, this RPR Frame is forwarded.
Two) lower ring is processed
In the time that node receives RPR Frame from a RPR physics aggregation port, if find, the target MAC (Media Access Control) address of the RPR head of this Frame points to this node, this RPR Frame is carried out to RPR decapsulation, and the message repeating after decapsulation is gone out.
If the target MAC (Media Access Control) address of the message after decapsulation points to this node, message is stayed to local processing.
Taking Fig. 4 as example, to establish in RPR looped network and comprise altogether 4 node S1, S2, S3, S4, each node is made up of two exchange pilings, and every switch provides an east orientation physical port and a west to physical port; For each node, two east orientation physical ports of two switches that this node is comprised and two west are to physical port polymerization respectively, obtain two RPR physics aggregation ports: east orientation physics aggregation port and west, to physics aggregation port, are tied to RPR logic interface by these two RPR physics aggregation ports.
Taking S1 as example, S1 is made up of two exchange pilings, and polymerization is distinguished to physical port in two east orientation physical ports and two west of two switches, obtains two rpr physics aggregation ports 1024 and 1025, is tied to RPR logic interface rpr1 by 1024 and 1025.
When initial, S1 is according to Topology Discovery, and study is to ring selection table.
1) in the time that S1 receives the Ethernet message that hangs preparation under self, according to the object IP address search ARP list item of message, if the outgoing interface in the ARP list item finding is rpr1, search RPR MAC address entries according to the target MAC (Media Access Control) address of ARP list item, find the MAC Address of target looped network node, search RPR ring selection table according to the MAC Address of target looped network node, if the ring finding is ring 0, determining interface is the ragg1024 in rpr1, according to RPR ring selection table, message is packaged into RPR Frame, and be transmitted to ragg1024, load balancing module on ragg1024 is selected one according to default hash algorithm in n physical port in ragg1024, message is sent from the port of selecting.
2) suppose that the part physical port in certain the RPR physics aggregation port of certain node on RPR looped network breaks down, because the state of this RPR physics aggregation port is still Up, therefore, each node can not perceive RPR ring-network topology to be changed; If the equal fault of all physical ports in certain RPR physics aggregation port, the node in RPR looped network is switched fast.
3) suppose that certain switch in the stacked group of certain node on RPR looped network breaks down, and this node is not link point under the target of above-mentioned message, only can affects aggregated links bandwidth, each node can not perceive RPR ring-network topology to be changed; If all switches in the stacked group of this node break down, the node in RPR looped network is switched fast.
4) suppose that certain node on RPR looped network is link point under the target of above-mentioned Ethernet message, in the time that all switches in the stacked group of this node all break down, this node is without any reachable path.
The composition schematic diagram of the node in the RPR looped network that Fig. 5 provides for the embodiment of the present invention, wherein, the each node in this RPR looped network is made up of n platform exchange piling respectively, and every switch provides an east orientation physical port and a west to physical port; For each node, n east orientation physical port of the n platform switch that this node is comprised and n west are to physical port polymerization respectively, obtain two RPR physics aggregation ports: east orientation physics aggregation port and west are to physics aggregation port, these two RPR physics aggregation ports are tied to RPR logic interface, n is integer, and n >=2, this node comprises:
Traffic forwarding module 51: carry out load balancing on RPR physics aggregation port.
Wherein, traffic forwarding module 51 can comprise ring selection table learning module 511 and forwarding module 512:
Ring selection table learning module 511: by Topology Discovery, set up RPR ring selection table, each list item in RPR ring selection table all comprises: the MAC Address of target looped network node, the corresponding ring of optimal path from this node to this target looped network node: 0 ring or 1 ring.
Forwarding module 512: in the time receiving the Ethernet message that hangs preparation under this node, in the ARP of this node table, search corresponding ARP list item according to the object IP address of message, obtain the outgoing interface of message, if the outgoing interface of this message is the MAC Address of the non-RPR looped network node of the target MAC (Media Access Control) address of RPR logic interface and message, in RPR MAC address entries, find the MAC Address of corresponding target looped network node according to the target MAC (Media Access Control) address of message, in RPR ring selection table according to the MAC Address of this target looped network node in ring selection table learning module 511, find corresponding ring, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded, wherein, RPR MAC address entries comprises: the MAC Address of RPR looped network node and hang the incidence relation of standby MAC Address under it.
Wherein, the outgoing interface that forwarding module 512 obtains message is further used for afterwards, if the target MAC (Media Access Control) address that the outgoing interface of this message is RPR logic interface and message is the MAC Address of RPR looped network node, in RPR ring selection table, find corresponding ring according to the target MAC (Media Access Control) address of this message, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded.
Forwarding module 512 goes out in n the physical port that physics aggregation port comprises and selects one to comprise at this: for n the physical port that goes out physics aggregation port and comprise, be only that all physical ports of choosing adopt default load balancing algorithm to select to state.
Forwarding module 512 is further used for after receiving down and hanging the next Ethernet message of preparation, while not finding corresponding ARP list item according to the object IP address of this Ethernet message in the ARP of this node table, with the object IP address architecture ARP request message of this message, this ARP request message is sent; In the time receiving arp response message, generate ARP list item or ARP list item+RPR MAC address entries, and forward described Ethernet message according to the ARP list item or the ARP list item+RPR MAC address entries that generate.
Forwarding module 512 generates ARP list item or ARP list item+RPR MAC address entries comprises: when arp response message is from hanging under this node when standby, only generate ARP list item; When arp response message is during from other node of RRP looped network, only generate ARP list item, wherein, the outgoing interface in ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding; When arp response message hangs under from other node of RRP looped network when standby, generate ARP list item and RPR MAC address entries, wherein, the outgoing interface of ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding, in RPR MAC address entries, comprises: the source MAC of the encapsulation RPRMAC head of the source MAC of arp response message and arp response message.
Traffic forwarding module 51 is further used for, in the time receiving RPR Frame from a RPR physics aggregation port of this node, if find, the target MAC (Media Access Control) address of the RPR MAC head of this Frame does not point to this node, according to default load balancing algorithm, in n the physical port comprising at another RPR physics aggregation port of this node, select one, from the port of selecting, this RPR Frame is forwarded.
In n the physical port that traffic forwarding module 51 comprises at another RPR physics aggregation port of this node, select one to comprise: n the physical port comprising for another RPR physics aggregation port of this node is only that all of the port of choosing adopts default load balancing algorithm to select to state.
Described in Fig. 5, node further comprises and switches fast module, in the time finding arbitrary RPR physics aggregation port Down of this node, triggers RPR and switches fast.
The embodiment of the present invention has following beneficial effect:
1) by by all east orientation physical ports in the stacked group of looped network node and western to physical port polymerization respectively, improve the bandwidth of RPR logic interface, strengthened the reliability of RPR.
2), in the time of part physical port Down in certain RPR physics aggregation port of looped network node, still normal lower ring of flow, has further strengthened the reliability of RPR.
The foregoing is only preferred embodiment of the present invention, in order to limit the present invention, within the spirit and principles in the present invention not all, any amendment of making, be equal to replacement, improvement etc., within all should being included in the scope of protection of the invention.
Claims (15)
1. the flow forwarding method in RPR looped network, is characterized in that, the each node in this RPR looped network is made up of n platform exchange piling respectively, and every switch provides an east orientation physical port and a west to physical port; For each node, n east orientation physical port of the n platform switch that this node is comprised and n west are to physical port polymerization respectively, obtain two RPR physics aggregation ports: east orientation physics aggregation port and west are to physics aggregation port, these two RPR physics aggregation ports are tied to RPR logic interface, n is integer, and n >=2, the method comprises:
Node in RPR looped network carries out load balancing on RPR physics aggregation port.
2. method according to claim 1, is characterized in that, the node in described RPR looped network carries out load balancing and comprises on RPR physics aggregation port:
Node in RPR looped network, by Topology Discovery, is set up RPR ring selection table, and each list item in RPR ring selection table all comprises: the MAC Address of target looped network node, the corresponding ring of optimal path from this node to this target looped network node;
In the time that described node receives the Ethernet message that hangs preparation under self, in the ARP table of self, search corresponding ARP list item according to the object IP address of message, obtain the outgoing interface of message, if the outgoing interface of this message is the MAC Address of the non-RPR looped network node in object media access control MAC address of RPR logic interface and message, in RPR MAC address entries, find the MAC Address of corresponding target looped network node according to the target MAC (Media Access Control) address of message, in RPR ring selection table, find corresponding ring according to the MAC Address of this target looped network node, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded, wherein, RPR MAC address entries comprises: the MAC Address of RPR looped network node and hang the incidence relation of standby MAC Address under it.
3. method according to claim 2, is characterized in that, described in obtain further comprising after the outgoing interface of message:
If the target MAC (Media Access Control) address that the outgoing interface of this message is RPR logic interface and message is the MAC Address of RPR looped network node, in RPR ring selection table, find corresponding ring according to the target MAC (Media Access Control) address of this message, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded.
4. according to the method in claim 2 or 3, it is characterized in that, describedly go out in n the physical port that physics aggregation port comprises to select one to comprise at this:
For n the physical port that goes out physics aggregation port and comprise, be only that all physical ports of choosing adopt default load balancing algorithm to select to state.
5. method according to claim 2, is characterized in that, describedly further comprises when described node receives after hanging the Ethernet message that preparation comes under self:
When described node does not find corresponding ARP list item according to the object IP address of this Ethernet message in the ARP table of self, with the object IP address architecture ARP request message of this message, this ARP request message is sent; In the time receiving arp response message, generate ARP list item, or ARP list item+RPR MAC address entries;
According to the ARP list item generating, or ARP list item+RPR MAC address entries forwards described Ethernet message.
6. method according to claim 5, is characterized in that, described generation ARP list item, or ARP list item+RPR MAC address entries comprises:
When arp response message is from hanging under this node when standby, only generate ARP list item;
When arp response message is during from other node of RRP looped network, only generate ARP list item, wherein, the outgoing interface in ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding;
When arp response message hangs under from other node of RRP looped network when standby, generate ARP list item and RPR MAC address entries, wherein, the outgoing interface of ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding, in RPR MAC address entries, comprises: the source MAC of the encapsulation RPR MAC head of the source MAC of arp response message and arp response message.
7. method according to claim 1, is characterized in that, the node in described RPR looped network carries out load balancing and further comprises on RPR physics aggregation port:
In the time that described node receives RPR Frame from a RPR physics aggregation port, if find, the target MAC (Media Access Control) address of the RPR MAC head of this Frame does not point to this node, according to default load balancing algorithm, in n the physical port comprising at another RPR physics aggregation port of self, select one, from the port of selecting, this RPR Frame is forwarded.
8. method according to claim 1, is characterized in that, described method further comprises:
In the time that described node is found the arbitrary RPR physics aggregation port Down of self, trigger RPR and switch fast.
9. the node in RPR looped network, is characterized in that, the each node in described RPR looped network is made up of n platform exchange piling respectively, and every switch provides an east orientation physical port and a west to physical port; For each node, n east orientation physical port of the n platform switch that this node is comprised and n west are to physical port polymerization respectively, obtain two RPR physics aggregation ports: east orientation physics aggregation port and west are to physics aggregation port, these two RPR physics aggregation ports are tied to RPR logic interface, n is integer, and n >=2, this node comprises:
Traffic forwarding module: carry out load balancing on RPR physics aggregation port.
10. node according to claim 9, is characterized in that, described traffic forwarding module comprises:
Ring selection table learning module: by Topology Discovery, set up RPR ring selection table, each list item in RPR ring selection table all comprises: the MAC Address of target looped network node, the corresponding ring of optimal path from this node to this target looped network node;
Forwarding module: in the time receiving the Ethernet message that hangs preparation under this node, in the ARP table of this node, search corresponding ARP list item according to the object IP address of message, obtain the outgoing interface of message, if the outgoing interface of this message is the MAC Address of the non-RPR looped network node in object media access control MAC address of RPR logic interface and message, in RPR MAC address entries, find the MAC Address of corresponding target looped network node according to the target MAC (Media Access Control) address of message, in RPR ring selection table, find corresponding ring according to the MAC Address of this target looped network node, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded, wherein, RPR MAC address entries comprises: the MAC Address of RPR looped network node and hang the incidence relation of standby MAC Address under it.
11. nodes according to claim 10, is characterized in that, the outgoing interface that described forwarding module obtains message is further used for afterwards,
If the target MAC (Media Access Control) address that the outgoing interface of this message is RPR logic interface and message is the MAC Address of RPR looped network node, in RPR ring selection table, find corresponding ring according to the target MAC (Media Access Control) address of this message, this message is packaged into RPR Frame, according to the outgoing interface of message and the ring finding, that determines message goes out physics aggregation port, according to default load balancing algorithm, go out in n the physical port that physics aggregation port comprises to select one at this, from the port of selecting, this RPR Frame is forwarded.
12. nodes according to claim 10, is characterized in that, described forwarding module goes out in n the physical port that physics aggregation port comprises to select one to comprise at this:
For n the physical port that goes out physics aggregation port and comprise, be only that all physical ports of choosing adopt default load balancing algorithm to select to state.
13. nodes according to claim 10, is characterized in that, described forwarding module is further used for after receiving down and hanging the next Ethernet message of preparation,
While not finding corresponding ARP list item according to the object IP address of this Ethernet message in the ARP of this node table, with the object IP address architecture ARP request message of this message, this ARP request message is sent; In the time receiving arp response message, generate ARP list item or ARP list item+RPR MAC address entries, and forward described Ethernet message according to the ARP list item or the ARP list item+RPR MAC address entries that generate.
14. nodes according to claim 13, is characterized in that, described forwarding module generates ARP list item or ARP list item+RPR MAC address entries comprises:
When arp response message is from hanging under this node when standby, only generate ARP list item;
When arp response message is during from other node of RRP looped network, only generate ARP list item, wherein, the outgoing interface in ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding;
When arp response message hangs under from other node of RRP looped network when standby, generate ARP list item and RPR MAC address entries, wherein, the outgoing interface of ARP list item is the RPR logic interface that the inbound port of arp response message is corresponding, in RPR MAC address entries, comprises: the source MAC of the encapsulation RPR MAC head of the source MAC of arp response message and arp response message.
15. nodes according to claim 9, is characterized in that, described traffic forwarding module is further used for,
In the time receiving RPR Frame from a RPR physics aggregation port of this node, if find, the target MAC (Media Access Control) address of the RPR MAC head of this Frame does not point to this node, according to default load balancing algorithm, in n the physical port comprising at another RPR physics aggregation port of this node, select one, from the port of selecting, this RPR Frame is forwarded.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410247931.1A CN104009903B (en) | 2014-06-05 | 2014-06-05 | Flow forwarding method and device in elastic package circular network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410247931.1A CN104009903B (en) | 2014-06-05 | 2014-06-05 | Flow forwarding method and device in elastic package circular network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104009903A true CN104009903A (en) | 2014-08-27 |
| CN104009903B CN104009903B (en) | 2017-07-14 |
Family
ID=51370399
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410247931.1A Active CN104009903B (en) | 2014-06-05 | 2014-06-05 | Flow forwarding method and device in elastic package circular network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104009903B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104348730A (en) * | 2014-11-28 | 2015-02-11 | 杭州华三通信技术有限公司 | Message forwarding method and system in RPR (resilient packet ring) network |
| CN104394084A (en) * | 2014-11-28 | 2015-03-04 | 杭州华三通信技术有限公司 | Message forwarding method and system in RPR network |
| CN106899507A (en) * | 2017-01-16 | 2017-06-27 | 新华三技术有限公司 | A kind of file transmitting method and device |
| CN108134744A (en) * | 2017-12-06 | 2018-06-08 | 新华三技术有限公司 | Load sharing method and veneer in a kind of IRF stackings |
| CN108718267A (en) * | 2018-03-13 | 2018-10-30 | 新华三技术有限公司 | Data message transmission method and device in a kind of RPR rings |
| CN109194386A (en) * | 2018-09-20 | 2019-01-11 | 新华三技术有限公司 | A kind of data message forwarding method and device |
| CN116094990A (en) * | 2023-01-19 | 2023-05-09 | 重庆长安汽车股份有限公司 | Software optimization method and device based on MAC address forwarding table and electronic equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060209877A1 (en) * | 2005-03-16 | 2006-09-21 | Fujitsu Limited | Communication system, communication card, and communication method |
| US20060274770A1 (en) * | 2005-06-06 | 2006-12-07 | Fujitsu Network Communications, Inc. | Two tier traffic shaper approach for resilient packet ring stations |
| CN101051983A (en) * | 2007-05-11 | 2007-10-10 | 杭州华三通信技术有限公司 | Construction method for reliable RPR circular net and relative RPR station device |
| CN101194473A (en) * | 2005-06-06 | 2008-06-04 | Ut斯达康通讯有限公司 | Method and site for achieving link aggregation between the interconnected RPR |
| CN101640644A (en) * | 2009-09-01 | 2010-02-03 | 杭州华三通信技术有限公司 | Method and equipment for flow equilibrium based on flexible link group |
| CN101854283A (en) * | 2010-06-01 | 2010-10-06 | 杭州华三通信技术有限公司 | Communication method and equipment of RPR (Resilient Packet Ring) looped network |
-
2014
- 2014-06-05 CN CN201410247931.1A patent/CN104009903B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060209877A1 (en) * | 2005-03-16 | 2006-09-21 | Fujitsu Limited | Communication system, communication card, and communication method |
| US20060274770A1 (en) * | 2005-06-06 | 2006-12-07 | Fujitsu Network Communications, Inc. | Two tier traffic shaper approach for resilient packet ring stations |
| CN101194473A (en) * | 2005-06-06 | 2008-06-04 | Ut斯达康通讯有限公司 | Method and site for achieving link aggregation between the interconnected RPR |
| CN101051983A (en) * | 2007-05-11 | 2007-10-10 | 杭州华三通信技术有限公司 | Construction method for reliable RPR circular net and relative RPR station device |
| CN101640644A (en) * | 2009-09-01 | 2010-02-03 | 杭州华三通信技术有限公司 | Method and equipment for flow equilibrium based on flexible link group |
| CN101854283A (en) * | 2010-06-01 | 2010-10-06 | 杭州华三通信技术有限公司 | Communication method and equipment of RPR (Resilient Packet Ring) looped network |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104348730A (en) * | 2014-11-28 | 2015-02-11 | 杭州华三通信技术有限公司 | Message forwarding method and system in RPR (resilient packet ring) network |
| CN104394084A (en) * | 2014-11-28 | 2015-03-04 | 杭州华三通信技术有限公司 | Message forwarding method and system in RPR network |
| CN104394084B (en) * | 2014-11-28 | 2018-05-11 | 新华三技术有限公司 | Message forwarding method and system in a kind of RPR networks |
| CN106899507A (en) * | 2017-01-16 | 2017-06-27 | 新华三技术有限公司 | A kind of file transmitting method and device |
| CN106899507B (en) * | 2017-01-16 | 2019-12-06 | 新华三技术有限公司 | message sending method and device |
| CN108134744A (en) * | 2017-12-06 | 2018-06-08 | 新华三技术有限公司 | Load sharing method and veneer in a kind of IRF stackings |
| CN108718267A (en) * | 2018-03-13 | 2018-10-30 | 新华三技术有限公司 | Data message transmission method and device in a kind of RPR rings |
| CN109194386A (en) * | 2018-09-20 | 2019-01-11 | 新华三技术有限公司 | A kind of data message forwarding method and device |
| CN109194386B (en) * | 2018-09-20 | 2020-06-19 | 新华三技术有限公司 | Data message forwarding method and device |
| CN116094990A (en) * | 2023-01-19 | 2023-05-09 | 重庆长安汽车股份有限公司 | Software optimization method and device based on MAC address forwarding table and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104009903B (en) | 2017-07-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8588060B2 (en) | E-spring (G.8032) interworking to provide access protection | |
| CN104009903A (en) | Flow forwarding method and device for RPR network | |
| CN101710877B (en) | Method, device and system for processing traffic flow based on pseudowire | |
| EP2391085B1 (en) | Multiple ports load sharing method, apparatus and network system | |
| US8854975B2 (en) | Scaling OAM for point-to-point trunking | |
| CN101753453B (en) | Networking method for ring network of packet transport network | |
| CN102571426B (en) | Double-homing protection method and device | |
| JP4034782B2 (en) | Ring connection device and data transfer control method | |
| EP2533475B1 (en) | Method and system for host route reachability in packet transport network access ring | |
| US8462636B2 (en) | Systems and methods for communication of management traffic over link aggregation group interface for a network element with distributed architecture | |
| CN101252494B (en) | Equipment and method for sharing traffic load | |
| JP2010509880A (en) | Hash-based multihoming | |
| CN102970231B (en) | Multicast data flow forwards implementation method and route-bridge（RB） | |
| CN101854283B (en) | Communication method and equipment of RPR (Resilient Packet Ring) looped network | |
| CN110324225B (en) | Method and device for processing message | |
| CN103607293B (en) | A kind of flow rate protecting method and equipment | |
| CN101800774A (en) | Environmental-friendly accessing method and environmental-friendly accessing network | |
| WO2013152700A1 (en) | Clock source selection method and device for microwave network element | |
| CN102355402B (en) | Method, system and router for establishing LSP (Label Switched Path) based on VPN (Virtual Private Network) route | |
| García et al. | LSOM: A link state protocol over mac addresses for metropolitan backbones using optical ethernet switches | |
| Huynh et al. | RRR: Rapid ring recovery submillisecond decentralized recovery for ethernet ring | |
| CN109194386B (en) | Data message forwarding method and device | |
| US20230300062A1 (en) | Route optimization method, physical network device and computer-readable storage medium | |
| CN115277543A (en) | Message forwarding method and device | |
| CN101695041B (en) | Method for realizing routing notification between every two virtual routers and physical router |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 310052 Binjiang District Changhe Road, Zhejiang, China, No. 466, No. Applicant after: Xinhua three Technology Co., Ltd. Address before: 310052 Binjiang District Changhe Road, Zhejiang, China, No. 466, No. Applicant before: Huasan Communication Technology Co., Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |