CN101136730A - Credible synchronization method of distributed network equipment - Google Patents
Credible synchronization method of distributed network equipment Download PDFInfo
- Publication number
- CN101136730A CN101136730A CNA2007101640946A CN200710164094A CN101136730A CN 101136730 A CN101136730 A CN 101136730A CN A2007101640946 A CNA2007101640946 A CN A2007101640946A CN 200710164094 A CN200710164094 A CN 200710164094A CN 101136730 A CN101136730 A CN 101136730A
- Authority
- CN
- China
- Prior art keywords
- control card
- standby
- card
- master
- main
- 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.)
- Pending
Links
Images
Abstract
The invention realizes following functions synthetically: implementing state negotiation among main and standby control cards of exchanger in core layer; reliable point-to-point transmission among main and standby control cards under QoS is ensured; reliable synchronization between network management information and route management information among main and standby control cards of exchanger in core layer; implementing compression algorithm in time for synchronizing batch data. Moreover, gentler synchronization between main and standby control cards is implemented so as to guarantee high reliability of core equipment in network layer.
Description
Technical field
The invention belongs to field of computer technology, relate between the active and standby control card of core layer rack switch synchronously and switching technique.
Background technology
High reliability is the requisite characteristic of telecom carrier-class device.The high reliability of system, when breaking down, can be fast, the accurately normal operation of recovery system, thus shorten the MTTR (Mean Time To Repair, average time for repair of breakdowns) of system.
Along with the growth of data business volume with to the raising of quality of service requirement, high reliability becomes one of high performance network most important characteristic day by day.The high availability of network is meant that system exchanges maximum running time for limited cost, and the service disruption loss that fault is caused drops to minimum.Network system with high availability need reduce the hardware and software fault on the one hand as far as possible, must do corresponding backup to valuable source on the other hand.Be about to occur in case detect fault, system can be rapidly with affected task transfers to the resource backup to continue to provide service.
Should support the high reliability characteristic for core layer switch in the network, one of important realization that for example active and standby master control borad technology is exactly the high reliability characteristic.Because network site of living in is extremely important, does not allow to occur Single Point of Faliure, this kind equipment generally all is equipped with two or more master control borad, is called main control board (Master) and slave control board (Slave).Wherein, main control board carries out service communication as the core and the outside of control plane, finishes the normal function of each module in the system; Slave control board does not then only have service communication as the backup and the outside of main control board.When main control board broke down, system carried out active and standby switching automatically, took over the work of main control board by slave control board, guaranteed professional normal operation.
When the master breaks down with control board, need be switched to standby control board, this just need be in the process of moving, and active and standby control board must carry out reliably synchronously, this synchronous, the synchronously real-time and synchronous three phases of data smoothing of batch that is divided into synchronously.
After slave control board started, the diversity ratio of main control board and slave control board was bigger owing to this moment, and main control board can be synchronized to slave control board with the Backup Data of current all modules, and this process is referred to as in batches synchronously.
With behind the EOS, system enters real-time synchronizing process, in this process, when the main control board data change, will be synchronized to standby plate in real time in batches.
Behind the masterslave switchover, standby plate is upgraded to new main board, can notify each module to carry out data collection and synchronous to business board, and it is synchronous that this process is referred to as data smoothing.In the smoothing process, each module initiatively communicates with business board, confirms aspect three of hardware states, link layer state, configuration data and synchronously, to guarantee that the data that whole system is safeguarded are consistent with state, thereby guarantee that after the active and standby switching, system can normally move.Level and smooth synchronous phase finishes, and new main board just is referred to as the main control board on the complete meaning.
In the synchronizing process of above three phases, main synchronous information is that active and standby state machine with control board is safeguarded, active and standby synchronous, active and standby with various transmit synchronous on the control board with the control board network management information.
At present, the core layer Ethernet switch active and standby following mechanism of using between the control board of synchronous employing commonly used:
● the Transmission Control Protocol link is realized
For the mode that on main standby card, adopts the TCP Handshake Protocol to realize, need set up more TCP on main standby card connects, wherein some TCP connects the state that is used to safeguard main standby card, this connection need remain, so that the negotiation of main standby card state then need be taked to connect to the batch data in synchronization in addition temporarily.In the clearly above implementation, need set up a large amount of TCP and connect, and each TCP connection all need be carried out hand shaking, need expend a large amount of cpu resources.And concerning core switch, main task is to handle complicated data forwarding, and the negotiation of main standby card adopts TCP to connect, and often implements too complexity, and the stability of equipment can reduce greatly.
● the udp protocol link is realized
On stand-by equipment, adopt UDP to connect and realize,, also need active and standbyly to carry out port snooping with control card though UDP need not too many shaking hands.The active and standby control card of Jian Liing in this way, compare TCP and connect, data volume can reduce, but because that UDP connects is unreliable, the negotiation of main standby control card state may cause asynchronous, and configuration information is lost or situation such as flow interrupt when causing active and standby card state to switch.
● the TCP+UDP protocol link is realized
Current, the active and standby method that adopts with the synchronization mechanism between the control board is: adopt TCP to connect to the negotiation of state, adopt UDP to be connected to batch data and real time information and realize.This implementation method is based upon on the ip protocol layer, to having relatively high expectations of equipment.Active and standbyly be not easy to expand, only be applicable to two control boards with control board.
More than all these realizations all be to be based upon on the IP agreement.
Before technical scheme of the present invention is described, introduce some Essential Terms earlier.
IP (Intemet Protocol (Internet protocol)): this agreement is the basic of the computer nowadays network interconnection, and it mainly acts on is various packet networks in the world to be carried out interconnected, please refer to RFC791 about the detailed introduction of this agreement.
TCP (Transmission Control Protocol (transmission control protocol)): run on the IP agreement, function is to guarantee that data correctly transmit between two nodes of IP network, please refer to RFC793 about the detailed introduction of this agreement.
UDP (User Datagram Protocol (User Datagram Protoco (UDP))): run on the IP agreement, function is to carry out insecure nothing to connect the packet transport service, please refer to RFC768 about the detailed introduction of this agreement.
Internet: literal translating is the Internet, refers to the general name of all in the world networks of coupling together by TCP/IP at present.
PC:personal computer, individual main frame.
GR technology introduction: after nucleus equipment is restarted in the network, all Routing Protocols need rerun, since this moment this equipment to external world topology situation know nothing, so need to carry out route again mutual with the external world, normal flow according to original Routing Protocol, can cause the neighborhood of peripheral equipment and this machine to rebuild this moment, and synchronous again all routing iinformation data, with neighbours of new discovery without any difference.Calculate and on peripheral equipment, can trigger route owing to the short interruption of neighborhood, therefore still can cause the route flapping on the network and transmit disruption.In order to address this problem, IETF has proposed a series of enhancing agreements respectively at various Routing Protocols (ISIS/OSPF/BGP/LDP etc.), strengthen in the agreement at these, by transformation to original agreement flow process, make neighborhood and the route that when this machine Control Software is restarted, can notify peripheral equipment will arrive this machine within a certain period of time keep stable, and it is synchronous that this machine of assistance carries out routing iinformation, in the short as far as possible time, make this machine that various routing iinformations are returned to and restart preceding state, network route and forwarding keep highly stable in whole process, message forwarding path is also without any change, from the aspect of whole network, be the same with not restarting incident.This enhancing agreement is collectively referred to as GR (gracefulrestart), is also sometimes referred to as NSF (non-stop forwarding).
Summary of the invention
The objective of the invention is to distributed synchronization problem at core layer switch in the existing network, credible synchronization method in a kind of distributed network equipment is provided, this method has realized really that the state machine of the active and standby control card of core layer switch is safeguarded, device configuration file synchronously, the data forwarding table synchronously, and can reduce the complexity of software greatly.
Technical scheme of the present invention is as follows: the credible synchronization method in a kind of distributed network equipment, on the core layer Ethernet switch, realize the synchronous of primary, spare control card, and comprise the steps:
(1) when device start, after primary, spare control card moved by synchronization mechanism, its state machine carried out transition according to the negotiation of synchronization mechanism, and negotiation rules is: if be to start simultaneously, the control card that slot number is little is that the master uses control card, and all the other are standby control card; If have only a main control card, then oneself be the main control card of using; After if control card starts, find the existing main control card of using, then oneself is consulted be standby control card;
(2) after primary, spare control card was consulted to finish, the master carried out batch data by synchronization mechanism and is synchronized to standby control card with the network management information and the routing iinformation of control card with oneself, after this in running a large amount of network management informations is regularly carried out synchronously.
Credible synchronization method in the aforesaid distributed network equipment, wherein, in the running of core layer switch, if have network management information or Routing Protocol to carry out partial update, the master sends to lastest imformation standby control card in real time with the control card synchronization mechanism and carries out data sync.
Credible synchronization method in the aforesaid distributed network equipment; wherein; standby control card detects main unusual or restart with control card; again consult to determine the new main control card of using between the remaining standby control card; new main data forwarding with control card management service card; and the announcement associated devices restarts, and returns to the state before switching.
Credible synchronization method in the aforesaid distributed network equipment, wherein, in the state negotiations process of synchronization mechanism, the point-to-point transmission based on link layer is adopted in the communication between the primary, spare control card, simultaneously by QoS technology implementation safety guarantee.
Credible synchronization method in the aforesaid distributed network equipment wherein, at the synchronous bulk information of batch data, compresses earlier, carries out then synchronously.
Credible synchronization method in the aforesaid distributed network equipment, wherein, the affirmation message that standby control card is synchronous to real time data is undertaken incidentally by the information in the state machine negotiation, to reduce the traffic of real-time synchrodata.
Credible synchronization method in the aforesaid distributed network equipment, wherein, when the master used the hardware anomalies of control card, standby control card detected by hardware mode, can immediately oneself be converted to the main control card of using; When adopting software mode to carry out craft switching activestandby state for the master with control card, upper layer application is notified standby main control card earlier, makes standby main control card be converted into the main control card of using immediately.
Credible synchronization method in the aforesaid distributed network equipment, wherein, standby control card is converted to main with behind the control card, service card is by the hardware interrupts mode, confirm main replacing with control card, service card will be transmitted business and transmit with control card by new master, make the data forwarding of service card can not interrupt.
Beneficial effect of the present invention is: by designing a kind of reliable synchronization mechanism method, make device just pass through this method, can safeguard the state between the active and standby control card preferably, and can realize the reliable transmission of data, avoid the huge overhead that uses Transmission Control Protocol to bring, or use the unreliable of udp protocol transmission.Make state information in the distributed network equipment, network management information and various transmit information be able to smooth realization synchronously, and then the redundancy backup of the equipment of assurance, thus guarantee the reliability of core network device.
Description of drawings
The state machine diagram that Fig. 1 safeguards for synchronization mechanism.
Fig. 2 is the data packet head schematic diagram of synchronization mechanism.
Fig. 3 is synchronization mechanism activestandby state machine transition figure.
Fig. 4 is a core Ethernet switch distributed platform schematic diagram.
Fig. 5 is a core Ethernet switch software function module schematic diagram.
Embodiment
The invention provides a kind of transmission method of setting up the reliable point-to-point on the link layer, the guarantee of QoS is arranged simultaneously.Can not only be similar to the TCP Handshake Protocol and carry out reliable transmission, can the data that need not reliable transmission directly be received and dispatched fast as udp protocol simultaneously.Its principle is: two master controls safeguard that all one sends buffer memory and a reception buffer memory, send buffer memory and be used to deposit the unacknowledged bag that sends, receive buffer memory and be used to deposit and receive but out of order bag, receive the buffer memory the inside according to the order of sequence bag directly give corresponding upper strata and handle.The sender safeguards local variable ack_expected (the minimum sequence number of the bag that the sender sends) and next_seq_to_send (maximum sequence number+1 that the sender sends); The recipient safeguards local variable seq_expected (smallest sequence number that expectation receives), if seq and seq_expected are unequal, send NAK and requires to retransmit seq_expected, gives corresponding processing function successively with bag according to the order of sequence and handles.
The re-transmission of bag is undertaken by the retransmission timer of each bag, and the retransmission timer inspection has or not NAK to arrive, and retransmits if having then, does not then continue to restart retransmission timer if having, and the closing to be wrapped by ACK of retransmission timer triggered.
In order to guarantee the reliable transmission of the active and standby card message of core switch, at first defined privately owned message format, support clean culture, multicast and the broadcasting each other of active and standby control card, this structure has considered that the staged operation of the data message of transmission chunk, the QoS that considers data simultaneously guarantee, i.e. the application of different stage can be with different service strategies.Data packet head structure such as Fig. 2 of this reliable synchronization mechanism definition, the detailed operation instruction of this structure will be described below.
● the process of reliably giving out a contract for a project of reliable synchronization mechanism:
1, application layer provides packet to be sent, obtains the slot number of oneself and the slot number of destination to be sent, fills in whether require reliability services, the COS classification of formulation.
2, synchronization mechanism is to the piece in the allocation of packets memory pool that distributes, if failure, then error count is returned.If the storage allocation success according to the MTU of link, is torn the data of application layer open part, from the internal layer pond, distribute internal layer.
If 3, whether the sequence number of synchronization mechanism inspection bag to be sent within send window, if can send, then fills in agreement packet header, send.The buffer memory timer of while log-on data bag.
4, the timing mechanism by starting if the buffer memory timer expiry and is received NAK, then retransmits above packet, if expired times reach n (n default=3) inferior after, then abandon this packet, add 1 with transmission buffering area number simultaneously, carry out error count.Send buffering area and increase progressively arrangement, therefore when inserting, operate by sequential chained list by the order that transmits Sequence Number.
5, each control card also needs the ACK that piggybacks a packet to confirm when giving out a contract for a project, so the ack item in the agreement packet header subtracts 1 for the series number that expectation receives.
6, at last by calling the driving function of network layer, the protocol type 0xff08 that defines in the synchronization mechanism in the interpolation sends to another piece main control card.
● the reliable packet receiving process of reliable synchronization mechanism:
1, the core switch bottom layer driving is received the packet that synchronization mechanism transmits.
2, whether synchronization mechanism protocol testing series number equals to expect the series number that receives, if equate, then directly gives application layer, check then and receive buffering area, check according to the order of sequence, give application layer together after the bag that needs are handled is recombinated, and the series number that expectation is received assignment again.If instead the series number of receiving bag is unequal with the series number that expectation receives, then the synchronization mechanism agreement is carried out buffer memory, checks whether the reception buffering area is arranged, if having, and buffer memory then, and the NAK information that sends expectation receiving serial number is to the other side.
3, to the synchronization mechanism protocol cache packet that gets up, start timer, if overtime, then send and retransmit NAK information to the opposite end, if expired times reach n (n default=3) inferior after, then abandon this packet.
If it is NAK information that 4 synchronization mechanisms are received packet, then the bag interpolation NAK that sends in the buffer memory is received mark.
If it is ack information that 5 synchronization mechanisms are received packet, if the ack series number sends between the series number in the ack series number and that expectation is received, then the timeing closing device discharges all smaller or equal to the transmission buffer memory of receiving the ack series number.
To insecure service, the sequence number that only needs reliably to receive and dispatch in the above in the packet procedures is not filled in data, directly receives and dispatches to get final product.
Method by above-mentioned reliable synchronization mechanism is achieved as follows the backup operation between two control cards of core switch:
Realize adopting two-dimentional tabledriven mode to realize to the state machine of reliable synchronization mechanism based on state, incident.The definition of concrete state, incident and action referring in concrete the enforcement about the specific descriptions of state machine.
To network management synchronously: synchronization mechanism monitors main control card when switching to major state, carries out configuration file; The synchronizing information of network management configuration adopts the data channel transmission of synchronization mechanism; And can be according to the operating system of this data channel online upgrading backup card.In order to accelerate transmission speed, adopt the compression algorithm that oneself defines simultaneously, transferring large number of data is compressed the back transmission.
To core switch equipment various transmit and move protocol status synchronously: needs are synchronized to transmitting of backup card, and the data transmission channel of synchronization mechanism provides interface, the transmitting of transmitting synchronous.Certainly, be to transmit after the compression equally to bigger data volume.
Below in conjunction with drawings and Examples the present invention is described in detail.
● the operation of synchronization mechanism:
This reliable synchronization mechanism state machine has following state:
RSP_STARTING: initial condition
RSP_BACKUP: Status of Backups
RSP_ACTIVE: active state
The batch synchronous regime;
Real-time synchronous regime;
Smooth state.
The main control board state machine of using moves from initial condition to the active state order, and standby control board state then moves from initial condition to the Status of Backups order.The active and standby control board state transition flow process of using is referring to Fig. 3.
Main with after the control board startup, at first be in initial condition, under the environment of single master control, the master can switch to active state with control board, if start simultaneously, then hold consultation with standby control board, the negotiation that slot number is little is the main control board of using, otherwise who starts earlier, just who is set to the main control board of using, and the master control borad that the back starts is then consulted to be standby control board.
And after standby control board starts, at first be in ready state, send own negotiation message to main with control board after, if existingly mainly transitted to active state with control board, or the result who consults be oneself to be standby, then transits to Status of Backups.
After the master receives the status message of standby control board with control board, just begin to send network management information backup and the various information of transmitting in batches to standby control board.
After standby control board is received main backup information with the control board transmission, carry out the local backup operation, so far start and consult and finish substantially.After entering the state of backing up in realtime, the state machine of primary and backup master control borad has just entered a metastable state, substantially no longer changes.
At the state of backing up in realtime, be in operation with control board when main, Routing Protocol upgrades the device forwards list item, can be synchronized to standby control card in real time, this moment can only main with control board on, entire equipment is carried out network management, and the master can be synchronized to network management information on the standby control card in real time with control board.On standby control card, can not carry out the webmaster operation.
When the synchronization mechanism on the slave control board detects the master control borad abnormal condition, then transit to smooth state.New master control borad can carry out data collection and information synchronization to business board in smoothing process, after process is finished, formally upgrade to the master and use control board, and state machine migrates to active state.And former master restarts with control board, becomes standby control board after the startup, and state machine is a Status of Backups.
● determining of reliable synchronization mechanism activestandby state
When system possessed control card more than 2, the passage of interacting message between the control card need be created by system.Stipulate that simultaneously any moment only allows a control card control system, we are called master with this card, and remaining control card is called the backup control card.
When a control card insertion system and after starting, can not obtain the control of system, but at first to the existence of control channel broadcast transmission hello message announcement oneself, if there has been master in the system at present, master will send Hello Reply to this card so, just oneself be set to back up control card after the new control card that inserts of this piece is received this message;
If after 3 hello message of the control card that inserts transmission, do not receive hello reply, do not receive the periodic hello message that master sends yet from master, so this piece control card just oneself role be set to master, beginning control system, and the broadcasting hello message on the control channel that control card uses in cycle;
If just be inserted with 2 control cards when system powers on, control card can receive the hello message of other control card when so initial, because this moment, all control cards were not master, need this moment one of them control card of election as master so, this moment election way be that groove position by the control card place conducts an election, that less piece control card of place slot number will become master, and send hello reply to other control card.
If breaking down, master do not have the cycle to send hello message, when the backup control card does not receive the hello message of master in the hold timer expiry time, the backup control card just will become master (time of hold timer is traditionally arranged to be 4 times of durations of master hello message).
After control card becomes the backup control card; the transmission hello message that also needs the cycle to master to announce existing of oneself; if master does not receive the hello message of backup control card in the hold timer expiry time (be traditionally arranged to be backup control card hello duration 4 times), just think that the backup control card worked undesired or do not exist.
When Master receives hello message, back up control card, do not need to send helloReply message to the backup control card if find the role who sends hello message; When if master receives hello message, the role who finds hello message also is master, if the smaller helloReply message that sends so of the slot number of oneself is to the other side, if the slot number of oneself is bigger than the other side, the role that oneself is set so is the backup control card.
● the synchronization mechanism message format
1, Hello form
| 2 bytes | 1 byte | 2 bytes | 1 byte | 1 byte | 2 bytes | 1 byte | The N byte |
| Length overall | Version number | The Hello type of message | Groove position (0 counting) | State | The Hello duration | Source address length | Source address |
Field description:
Length overall: take 2 bytes, the byte number (comprising the byte that the length overall field takies itself) that expression message takies;
Version number: the version of expression HA program is 1 at present;
The Hello type of message: command code is 1;
State: the state of sign control card current state machine, 0 this control card of expression are in the start state, 1 expression control card is in backup state, 2 and represents that control cards are in the active state;
The Hello duration: represent that this control card sends the cycle of Hello, unit is a millisecond
Source address length: because expectation HA module can be applied on the miscellaneous equipment, and different equipment may use different transmission plans when communication: may use other transmission medium such as what the direct use Ethernet transmission that has, presumable equipment use TCP transmitted, also have, for the address that how to identify this control card different method for expressing just be arranged so.So the time this part is designed to variablely in design, explains by each system oneself.
The HelloReply form
| 2 bytes | 1 byte | 2 bytes | 1 byte | 1 byte | 1 byte | The N byte |
| Length overall | Version number | The HelloReply type of message | Groove position (0 counting) | State | Source address length | Source address |
The meaning of each field of HelloReply message is consistent with Hello message, and just HelloReply message is encoded to 2.
The Swap form
| 2 bytes | 1 byte | 2 bytes | 1 byte | 1 byte | 1 byte | The N byte |
| Length overall (byte) | Version number | 3 (Swap message) | Groove position (0 counting) | State | Source address length | Source address |
The command code of Swap message is 3, and the description of all the other fields is the same with the front.
● synchronous state machine is described
This reliable synchronization mechanism state machine has following state:
RSP_STARTING: initial condition
RSP_BACKUP: Status of Backups
RSP_ACTIVE: active state
The batch synchronous regime;
Real-time synchronous regime;
Smooth state.
This reliable synchronization mechanism may cause that the incident of status change has:
Receive Hello message
Receive HelloReply message
Receive KeepAlive message
Hello message is overtime
Opposite end KeepAlive overtime (being that IDLE is overtime)
Extract the opposite end
Forced Switch message
The message that the various states of this reliable synchronization mechanism may send down is:
Hello message may be sent under RSP_STARTING and RSP_ACTIVE state.
HelloReply message may be sent under the RSP_ACTIVE state.
KeepAlive message may be sent under RSP_ACTIVE and RSP_BACKUP state.
State machine is as shown in the table:
| Incident state | ?Ha_Start | ?Ha_Backup | ?Ha_Active |
| Rcv?Hello | ?F1 | ?F5 | ?F8 |
| ?Rcv?HelloReply | ?F2 | ?- | ?F9 |
| ?Peer?Extract | ?F3 | ?F6 | ?F10 |
| ?Swap | ?- | ?F13 | ?F11 |
| ?Hello?Exp | ?F4 | ?F7 | ?F7 |
| ?Hold?Exp | ?- | ?F14 | ?F12 |
State description is described as follows:
HA_Start: this state representation control card starts, and does not administer whole system, surveys the existence of other control card by sending hello message.
HA_Backup: this this control card of state representation is in Status of Backups, can not send Hello message the cycle at this State Control card.
HA_Active: this state representation control card is in state of activation, has administered system, and the transmission hello message in cycle is to the backup control card.
Event description is as follows:
Rcv Hello: the Hello message from the another one control card has been received in expression;
Rcv HelloReply: the Hello Reply message from the another one control card has been received in expression;
Peer Extract: the another one control card in the expression system is extracted by administrator hand;
Swap: promptly the keeper issues an order and allows two control cards exchange roles;
Hello Exp: i.e. hello timer expiry incident;
Hold Exp:hold timer expiry incident.
Processing procedure is described:
When state machine starts, need to set some initial values: the role is the backup control card, and hello time number is set to 0; Send hello message then, and start the hello timer.
1、F1
If the role of the hello that receives bag is that master or the other side's slot number is littler than the slot number of oneself, time according to the time hold timer in the hello bag is set to N times (default suggestion is 4 times) in the hello time value, start the hold timer, the role of oneself is set to back up control card, and transition between states is to the HA_Backup state;
If the role of the hello that receives bag is not master, and the other side's slot number is bigger than oneself, the role that oneself is set so is master, record backup control card is on the throne, the time-out time that the hold timer is set is that N is doubly to backup control card hello periodic quantity, send helloReply to the backup control card, start the hold timer, transition between states is to HA_Active.
2、F2
If the role of the helloReplay that receives bag is not master, do not handle so;
If the role of the helloReplay that receives bag is master, be set to N that master sends the hello time doubly (default 4 times) according to time of the time hold timer in the hello bag, start the hold timer, transition between states is to the HA_Backup state;
3、F3
If system can provide an other control card to extract incident, the role of oneself is set to master so, and record backup control card is not on the throne, and transition between states is to HA_Active.
4、F4
Hello is sent number of times add 1, if send number of times more than or equal to 3, it is clear 0 so hello to be sent number of times, and the role who is provided with oneself be master, startup hello timer, and HA_Active is arrived in transition between states.Otherwise, send the hello bag, start the hello timer
5、F5
If the role of hello bag is master, restart the hold timer so; If the role of hello is not master, the role that oneself is set is master, and it is on the throne that the backup control card is set, and restarts the hold timer, and transition between states is to HA_Active.
6、F6
The role that oneself is set is master, stops the hold timer, and record backup control card is not on the throne, and transition between states is to HA_Active.
7、F7
Send hello message, start the hello timer, state remains unchanged;
8、F8
If the hello condition of information is the start state, send helloReply so, and start the hold timer, record backup control card is on the throne, and state remains unchanged;
If the hello condition of information is the backup state, restart the hold timer so, state remains unchanged;
If the hello condition of information is active, judge the other side and own slot number so, if oneself groove bit comparison is little, send so helloReply message to the other side, and write down back up control card on the throne, restart the hold timer, state remains unchanged; If little to the square groove bit comparison, the role that oneself is set so is the backup control card, and transition between states is to HA_Backup;
9、F9
If receive helloReply at the active state, represent promptly that the other side thinks and oneself be master and require we to become the backup control card, at this moment, the role that oneself is set for the backup control card, restart the hold timer, transition between states is to HA_Backup;
10、F10
Control card is on the throne if record shows backup, stops the hold timer so, that the backup control card is set is not on the throne, and state remains unchanged.
11、F11
(control card receives present event and represents that webmaster requirement system control card switches) if the backup control card is not on the throne, do not deal with so, and state remains unchanged;
If the backup control card is on the throne, the role that oneself is set so is the backup control card, sends the swap order, restarts the hold timer to the other side, and state becomes HA_Backup;
12、F12
It is not on the throne that the backup control card is set, and state remains unchanged.
13、F13
The role that oneself is set is master, and it is on the throne that the backup control card is set, and restarts the hold timer, and transition between states is to HA_Active.
14、F14
The role that oneself is set is master, and it is not on the throne that the backup control card is set, and transition between states is to HA_Active.
● the synchronization mechanism data packet head
As Fig. 2:
The Version:1 byte, version number;
Srv Type:1bit, reliably still unreliable service, 0: unreliable, 1: reliable;
Srv Cos:3bit, priority, 0-7,0 represents lowest priority, and 7 represent limit priority;
Reserve:4bit keeps;
Total Len:2 byte, head adds the total length of additional message;
The Src:2 byte shows the source of data, is designed to the bit of source slot number at present, from low to high, is followed successively by 0,1,2,3bit, if the groove position is 1, and then 1bit set;
The Dst:2 byte shows the purpose of data, is designed to the bit of purpose slot number, because may purpose be a plurality of master controls;
The identifier:2 byte, each part application data until that expression sends, when the data of application layer during greater than mtu, identifier is used for indicating and organizing the fragment packets that belongs to an application
Flag:4bit, more sheet MF sign, the flag of broadcasting or clean culture sign and IP fragmentation is similar
Offset:12bit, the burst skew
The seq:2 byte transmits Sequence Number, and 0 to 216-1
The ack:2 byte confirms number that 0 to 216-1
The reserved:2 byte
The checksum:2 byte, the verification of whole head and
● the reliable synchronization of configuration layer management information realizes
By the batch backup and the process of backing up in realtime, the configuration information of main control board backups to slave control board timely.When masterslave switchover maturation on opportunity, preserved the complete configuration information content of main control board on the slave control board, and configuration synchronization success.Therefore, when masterslave switchover,, can realize level and smooth transition fully, not need again extra the collection or synchronous other guide for configuration information.
● the reliable synchronization of protocol layer realizes
In level and smooth process, to transmit for guaranteeing professional free of discontinuities, all kinds of the transmitting that relates on the business board is not to relearn after the deletion.In data collection and synchronizing process that master control borad carries out, it is constant to keep original business board data substantially, only upgrades to change partly and refresh.Protocol layer mainly comprises following transmit synchronous:
1. two layers of clean culture are transmitted
Common two layers of clean culture are transmitted and have just been utilized mac address table, and message is transmitted related MAC table and all existed on business board.When masterslave switchover took place, new master control borad can be initiated mac address information to business board and be collected with synchronously, and original MAC table can't be deleted in this process on the business board, can guarantee that therefore two layers of unicast data are transmitted normally on the business board.
2. Layer 2 Multicast is transmitted
For Layer 2 Multicast, same, transmit required multicast MAC address entries and all be kept on the business board.During masterslave switchover, master control borad carries out data collection to interface board, and the original multicast MAC address entries of business board remains unchanged, and therefore can guarantee the normal forwarding of Layer 2 Multicast stream.
3. three layers of clean culture are transmitted
For three layers of clean culture, transmitting mainly is ARP and the three layers of routing table that rely on interface board.During the masterslave switchover smoothing process, original ARP and three layers of routing table remain unchanged on the business board; Simultaneously, because OSPF, BGP, ISIS Routing Protocol are all supported the GR function, during switching, peripheral neighbor device also keeps respective neighbor relation and route constant, therefore IP message forwarding path does not change in the network, has guaranteed that three layers of clean culture transmit business and do not interrupt.
4. three layers of multicast forward table
For three layers of multicast, same in level and smooth process, because original multicast list of setting up is constant in the interface board reservation, original group business stream can guarantee normal forwarding.
Fig. 4 has represented the distributed platform structure of a concrete core Ethernet switch, and this test platform has adopted the architecture of dual master control and 4 professional ply-yarn drills.This core Ethernet switch, it is flexible to have port arrangement, simple to operate, easy to maintenance, stable performance, characteristics such as powerful.Be applicable to network backbone, the exchcange core of IP metropolitan area network, large enterprise's net and garden net or converge the center.Be particularly suitable for being used in the following business of carrier network Bras and compile contact.Also can be used as the network backbone of university education net or compile.
This switch is supported IEEE 802.1d Spanning-Tree Protocol; IEEE 802.1w produces tree protocol fast; Based on port vlan and IEEE 802.1q VLAN; The management of IEEE 802.1P priority query; IGMP Snooping supports 1024 multicast group at most; Port speed control, the rate limit granularity is the 64K bits per second; Flow Control, broadcast storm controlled function; IEEE802.1x authentication, Radius; Switch cluster management SGM; DHCP RELAY, DHCPSNOOP; RIP, OSPFv2, BGPv4, PIM-SM; IPv6, MPLS etc. enrich function.
The software function module of this switch main control card as shown in Figure 5, the major function that the present invention describes is positioned at master control high reliability communication control module, it and system protocol module are mutual, realize between the distribution switch control card and high reliable communication and synchronization mechanism.
Claims (8)
1. the credible synchronization method in the distributed network equipment, on the core layer Ethernet switch, realize primary, spare control card synchronously, comprise the steps:
(1) when device start, after primary, spare control card moved by synchronization mechanism, its state machine carried out transition according to the negotiation of synchronization mechanism, and negotiation rules is: if be to start simultaneously, the control card that slot number is little is that the master uses control card, and all the other are standby control card; If have only a main control card, then oneself be the main control card of using; After if control card starts, find the existing main control card of using, then oneself is consulted be standby control card;
(2) after primary, spare control card was consulted to finish, the master carried out batch data by synchronization mechanism and is synchronized to standby control card with the network management information and the routing iinformation of control card with oneself, after this in running a large amount of network management informations is regularly carried out synchronously.
2. the credible synchronization method in the distributed network equipment as claimed in claim 1, it is characterized in that: in the running of core layer switch, if have network management information or Routing Protocol to carry out partial update, the master sends to lastest imformation standby control card in real time with the control card synchronization mechanism and carries out data sync.
3. the credible synchronization method in the distributed network equipment as claimed in claim 1 or 2; it is characterized in that: standby control card detects main unusual or restart with control card; again consult to determine the new main control card of using between the remaining standby control card; new main data forwarding with control card management service card; and the announcement associated devices restarts, and returns to the state before switching.
4. the credible synchronization method in the distributed network equipment as claimed in claim 1 or 2, it is characterized in that: in the state negotiations process of synchronization mechanism, the point-to-point transmission based on link layer is adopted in communication between the primary, spare control card, simultaneously by QoS technology implementation safety guarantee.
5. the credible synchronization method in the distributed network equipment as claimed in claim 4 is characterized in that: at the synchronous bulk information of batch data, compress earlier, carry out then synchronously.
6. the credible synchronization method in the distributed network equipment as claimed in claim 1 or 2 is characterized in that: the affirmation message that standby control card is synchronous to real time data, undertaken incidentally by the information in the state machine negotiation, to reduce the traffic of real-time synchrodata.
7. the credible synchronization method in the distributed network equipment as claimed in claim 3 is characterized in that: when the master used the hardware anomalies of control card, standby control card detected by hardware mode, can immediately oneself be converted to the main control card of using; When adopting software mode to carry out craft switching activestandby state for the master with control card, upper layer application is notified standby main control card earlier, makes standby main control card be converted into the main control card of using immediately.
8. the credible synchronization method in the distributed network equipment as claimed in claim 7, it is characterized in that: standby control card is converted to main with behind the control card, service card is by the hardware interrupts mode, confirm main replacing with control card, service card will be transmitted business and transmit with control card by new master, make the data forwarding of service card can not interrupt.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2007101640946A CN101136730A (en) | 2007-10-19 | 2007-10-19 | Credible synchronization method of distributed network equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2007101640946A CN101136730A (en) | 2007-10-19 | 2007-10-19 | Credible synchronization method of distributed network equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101136730A true CN101136730A (en) | 2008-03-05 |
Family
ID=39160589
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2007101640946A Pending CN101136730A (en) | 2007-10-19 | 2007-10-19 | Credible synchronization method of distributed network equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101136730A (en) |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101984573A (en) * | 2010-11-15 | 2011-03-09 | 中兴通讯股份有限公司 | Method and system for distributed realization of LACP standard state machine |
| CN102006223A (en) * | 2010-11-17 | 2011-04-06 | 福建星网锐捷网络有限公司 | Data transmission method, device and system between cards, board card and distributed system |
| CN102104596A (en) * | 2010-12-17 | 2011-06-22 | 曙光信息产业股份有限公司 | Method for reordering TCP (transmission control protocol) connections |
| CN102420746A (en) * | 2011-11-14 | 2012-04-18 | 中兴通讯股份有限公司 | Forwarding method for multicast flow and network equipment therefor |
| CN101510890B (en) * | 2009-03-31 | 2012-07-04 | 华为技术有限公司 | Method and communication equipment for holding protocol state |
| CN102694735A (en) * | 2012-06-11 | 2012-09-26 | 杭州华三通信技术有限公司 | Method and device used for synchronization after restart of FC (fiber channel) exchanger |
| CN102820989A (en) * | 2012-08-07 | 2012-12-12 | 杭州华三通信技术有限公司 | Network equipment and asset management method thereof |
| CN103117876A (en) * | 2013-01-24 | 2013-05-22 | 中兴通讯股份有限公司 | User state information synchronizing method and device |
| CN103188690A (en) * | 2011-12-29 | 2013-07-03 | 北京信威通信技术股份有限公司 | Core network smooth expansion system and core network smooth expansion method |
| CN104052586A (en) * | 2013-03-14 | 2014-09-17 | 中兴通讯股份有限公司 | Method for realizing quick list synchronization between line cards and device thereof |
| CN104065499A (en) * | 2013-03-19 | 2014-09-24 | 鼎点视讯科技有限公司 | Main control board, main/standby system and information backup method and device |
| CN104135488A (en) * | 2014-08-13 | 2014-11-05 | 上海申腾信息技术有限公司 | Remote medical system related data transmission system and method and applications thereof |
| CN105391629A (en) * | 2014-09-05 | 2016-03-09 | 杭州迪普科技有限公司 | Resource backup method and device |
| CN106936641A (en) * | 2017-03-27 | 2017-07-07 | 迈普通信技术股份有限公司 | Entry management method and system |
| CN106941418A (en) * | 2017-03-10 | 2017-07-11 | 杭州迪普科技股份有限公司 | The synchronous method and device of SSL VPN configuration informations |
| CN107395424A (en) * | 2017-07-31 | 2017-11-24 | 江西山水光电科技股份有限公司 | A kind of method for realizing TCP connections backup |
| CN107888434A (en) * | 2017-12-29 | 2018-04-06 | 杭州迪普科技股份有限公司 | Network equipments configuration synchronous method and device |
| CN108234306A (en) * | 2016-12-21 | 2018-06-29 | 丛林网络公司 | Network equipment, network method and computer readable storage medium |
| CN108712461A (en) * | 2018-04-04 | 2018-10-26 | 烽火通信科技股份有限公司 | The realization system and method for data aggregation service |
| CN109309616A (en) * | 2017-07-27 | 2019-02-05 | 中兴通讯股份有限公司 | Notifying method and device based on ISIS agreement |
| CN110086581A (en) * | 2019-05-30 | 2019-08-02 | 深圳市中航比特通讯技术有限公司 | The synchronous method and apparatus of a main prepare more data under a kind of network integration scene |
| CN110224849A (en) * | 2019-04-18 | 2019-09-10 | 广州市高科通信技术股份有限公司 | A kind of method preventing PTN device loss of data, system and storage medium |
| CN111083074A (en) * | 2019-12-20 | 2020-04-28 | 盛科网络（苏州）有限公司 | High availability method and system for main and standby dual OSPF state machines |
| CN111131500A (en) * | 2019-12-31 | 2020-05-08 | 盛科网络（苏州）有限公司 | Method and system for switching primary multicast and standby multicast |
| CN111464346A (en) * | 2020-03-23 | 2020-07-28 | 珠海高凌信息科技股份有限公司 | Main and standby control board synchronization method and system based on ATCA (advanced telecom computing architecture) |
| CN111641568A (en) * | 2020-05-09 | 2020-09-08 | 中国电子科技集团公司第五十四研究所 | Modular switch and network architecture |
| CN112073844A (en) * | 2020-08-28 | 2020-12-11 | 烽火通信科技股份有限公司 | Transmission equipment configuration verification synchronization method and system |
| US10887173B2 (en) | 2016-12-21 | 2021-01-05 | Juniper Networks, Inc. | Communicating state information in distributed operating systems |
| US11095742B2 (en) | 2019-03-27 | 2021-08-17 | Juniper Networks, Inc. | Query proxy for delivery of dynamic system state |
| US11316775B2 (en) | 2016-12-21 | 2022-04-26 | Juniper Networks, Inc. | Maintaining coherency in distributed operating systems for network devices |
-
2007
- 2007-10-19 CN CNA2007101640946A patent/CN101136730A/en active Pending
Cited By (50)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101510890B (en) * | 2009-03-31 | 2012-07-04 | 华为技术有限公司 | Method and communication equipment for holding protocol state |
| CN101984573A (en) * | 2010-11-15 | 2011-03-09 | 中兴通讯股份有限公司 | Method and system for distributed realization of LACP standard state machine |
| CN102006223A (en) * | 2010-11-17 | 2011-04-06 | 福建星网锐捷网络有限公司 | Data transmission method, device and system between cards, board card and distributed system |
| CN102006223B (en) * | 2010-11-17 | 2013-06-19 | 福建星网锐捷网络有限公司 | Data transmission method, device and system between cards, board card and distributed system |
| CN102104596A (en) * | 2010-12-17 | 2011-06-22 | 曙光信息产业股份有限公司 | Method for reordering TCP (transmission control protocol) connections |
| CN102104596B (en) * | 2010-12-17 | 2015-08-19 | 曙光信息产业股份有限公司 | A kind of TCP connects method for reordering |
| CN102420746A (en) * | 2011-11-14 | 2012-04-18 | 中兴通讯股份有限公司 | Forwarding method for multicast flow and network equipment therefor |
| CN103188690A (en) * | 2011-12-29 | 2013-07-03 | 北京信威通信技术股份有限公司 | Core network smooth expansion system and core network smooth expansion method |
| CN102694735B (en) * | 2012-06-11 | 2014-12-10 | 杭州华三通信技术有限公司 | Method and device used for synchronization after restart of FC (fiber channel) exchanger |
| CN102694735A (en) * | 2012-06-11 | 2012-09-26 | 杭州华三通信技术有限公司 | Method and device used for synchronization after restart of FC (fiber channel) exchanger |
| US9350621B2 (en) | 2012-06-11 | 2016-05-24 | Hangzhou H3C Technologies Co., Ltd. | Synchronization after restart of a FC switch |
| CN102820989B (en) * | 2012-08-07 | 2015-09-16 | 杭州华三通信技术有限公司 | The network equipment and assets management method thereof |
| CN102820989A (en) * | 2012-08-07 | 2012-12-12 | 杭州华三通信技术有限公司 | Network equipment and asset management method thereof |
| CN103117876A (en) * | 2013-01-24 | 2013-05-22 | 中兴通讯股份有限公司 | User state information synchronizing method and device |
| CN103117876B (en) * | 2013-01-24 | 2018-10-26 | 中兴通讯股份有限公司 | The synchronous method and device of user state information |
| CN104052586A (en) * | 2013-03-14 | 2014-09-17 | 中兴通讯股份有限公司 | Method for realizing quick list synchronization between line cards and device thereof |
| CN104052586B (en) * | 2013-03-14 | 2018-01-23 | 中兴通讯股份有限公司 | Realize the method and device that quick table is synchronous between line card |
| CN104065499A (en) * | 2013-03-19 | 2014-09-24 | 鼎点视讯科技有限公司 | Main control board, main/standby system and information backup method and device |
| CN104065499B (en) * | 2013-03-19 | 2017-06-27 | 鼎点视讯科技有限公司 | Master control borad, main preparation system, information backup method and device |
| CN104135488A (en) * | 2014-08-13 | 2014-11-05 | 上海申腾信息技术有限公司 | Remote medical system related data transmission system and method and applications thereof |
| CN104135488B (en) * | 2014-08-13 | 2018-04-27 | 上海申腾信息技术有限公司 | A kind of data transmission system and transmission method and application in relation to Telemedicine System |
| CN105391629A (en) * | 2014-09-05 | 2016-03-09 | 杭州迪普科技有限公司 | Resource backup method and device |
| CN105391629B (en) * | 2014-09-05 | 2019-01-22 | 杭州迪普科技股份有限公司 | A kind of method and device of resource backup |
| CN108234306A (en) * | 2016-12-21 | 2018-06-29 | 丛林网络公司 | Network equipment, network method and computer readable storage medium |
| CN108234306B (en) * | 2016-12-21 | 2021-03-19 | 瞻博网络公司 | Network device, network method, and computer-readable storage medium |
| US10887173B2 (en) | 2016-12-21 | 2021-01-05 | Juniper Networks, Inc. | Communicating state information in distributed operating systems |
| US11265216B2 (en) | 2016-12-21 | 2022-03-01 | Juniper Networks, Inc. | Communicating state information in distributed operating systems |
| US11316775B2 (en) | 2016-12-21 | 2022-04-26 | Juniper Networks, Inc. | Maintaining coherency in distributed operating systems for network devices |
| US11316744B2 (en) | 2016-12-21 | 2022-04-26 | Juniper Networks, Inc. | Organizing execution of distributed operating systems for network devices |
| US11924044B2 (en) | 2016-12-21 | 2024-03-05 | Juniper Networks, Inc. | Organizing execution of distributed operating systems for network devices |
| CN106941418B (en) * | 2017-03-10 | 2019-12-06 | 杭州迪普科技股份有限公司 | SSL VPN configuration information synchronization method and device |
| CN106941418A (en) * | 2017-03-10 | 2017-07-11 | 杭州迪普科技股份有限公司 | The synchronous method and device of SSL VPN configuration informations |
| CN106936641A (en) * | 2017-03-27 | 2017-07-07 | 迈普通信技术股份有限公司 | Entry management method and system |
| CN106936641B (en) * | 2017-03-27 | 2019-10-29 | 迈普通信技术股份有限公司 | Entry management method and system |
| CN109309616A (en) * | 2017-07-27 | 2019-02-05 | 中兴通讯股份有限公司 | Notifying method and device based on ISIS agreement |
| CN109309616B (en) * | 2017-07-27 | 2022-03-01 | 中兴通讯股份有限公司 | Notification method and device based on ISIS protocol |
| US11196658B2 (en) | 2017-07-27 | 2021-12-07 | Xi'an Zhongxing New Software Co., Ltd. | Intermediate system to intermediate system routing protocol based notification method and apparatus |
| CN107395424A (en) * | 2017-07-31 | 2017-11-24 | 江西山水光电科技股份有限公司 | A kind of method for realizing TCP connections backup |
| CN107888434A (en) * | 2017-12-29 | 2018-04-06 | 杭州迪普科技股份有限公司 | Network equipments configuration synchronous method and device |
| CN108712461A (en) * | 2018-04-04 | 2018-10-26 | 烽火通信科技股份有限公司 | The realization system and method for data aggregation service |
| US11095742B2 (en) | 2019-03-27 | 2021-08-17 | Juniper Networks, Inc. | Query proxy for delivery of dynamic system state |
| CN110224849A (en) * | 2019-04-18 | 2019-09-10 | 广州市高科通信技术股份有限公司 | A kind of method preventing PTN device loss of data, system and storage medium |
| CN110086581A (en) * | 2019-05-30 | 2019-08-02 | 深圳市中航比特通讯技术有限公司 | The synchronous method and apparatus of a main prepare more data under a kind of network integration scene |
| CN111083074A (en) * | 2019-12-20 | 2020-04-28 | 盛科网络（苏州）有限公司 | High availability method and system for main and standby dual OSPF state machines |
| CN111131500A (en) * | 2019-12-31 | 2020-05-08 | 盛科网络（苏州）有限公司 | Method and system for switching primary multicast and standby multicast |
| CN111131500B (en) * | 2019-12-31 | 2022-11-04 | 苏州盛科通信股份有限公司 | Method and system for switching main multicast and standby multicast in two layers |
| CN111464346A (en) * | 2020-03-23 | 2020-07-28 | 珠海高凌信息科技股份有限公司 | Main and standby control board synchronization method and system based on ATCA (advanced telecom computing architecture) |
| CN111641568A (en) * | 2020-05-09 | 2020-09-08 | 中国电子科技集团公司第五十四研究所 | Modular switch and network architecture |
| CN112073844A (en) * | 2020-08-28 | 2020-12-11 | 烽火通信科技股份有限公司 | Transmission equipment configuration verification synchronization method and system |
| CN112073844B (en) * | 2020-08-28 | 2022-08-12 | 烽火通信科技股份有限公司 | Transmission equipment configuration verification synchronization method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101136730A (en) | Credible synchronization method of distributed network equipment | |
| US9036466B2 (en) | Methods and apparatus for improving network communication using ethernet switching protection | |
| US8850068B2 (en) | United router farm setup | |
| US9059902B2 (en) | Procedures, apparatuses, systems, and computer-readable media for operating primary and backup network elements | |
| EP2843906B1 (en) | Method, apparatus, and system for data transmission | |
| EP3641243A1 (en) | Lag configuration learning in an extended bridge | |
| CN100417141C (en) | Group broadcasting business realizing method | |
| WO2010069199A1 (en) | Management apparatus, control apparatus, management and controlling apparatus and router system | |
| CN102084638A (en) | Deterministic session load-balancing and redundancy of access servers in a computer network | |
| CN101573940A (en) | System and method for TCP high availability | |
| WO2020168854A1 (en) | Evpn multicast method, apparatus and system | |
| CN101060533B (en) | A method, system and device for improving the reliability of VGMP protocol | |
| JP5861772B2 (en) | Network appliance redundancy system, control device, network appliance redundancy method and program | |
| WO2014121468A1 (en) | Method and device for establishing multicast data channel in network virtualization system | |
| WO2014127629A1 (en) | Message forwarding system, method and device | |
| US11546267B2 (en) | Method for determining designated forwarder (DF) of multicast flow, device, and system | |
| CN103975563A (en) | Lossless connection failover for single devices | |
| WO2021082812A1 (en) | Message sending method and first network device | |
| US20130279378A1 (en) | Cascaded Streaming of Data Through Virtual Chain of Nodes in Hub Topology | |
| US20180359181A1 (en) | Ethernet frame transmission method in software defined networks (sdn) | |
| WO2011103759A1 (en) | Method for establishing associated bidirectional label switching path and system thereof | |
| WO2021093463A1 (en) | Packet forwarding method, first network device, and first device group | |
| EP3588859A1 (en) | Network device configuration versioning | |
| WO2012129921A1 (en) | Multicast method and multicast device | |
| WO2012159461A1 (en) | Layer-2 path maximum transmission unit discovery method and node |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Open date: 20080305 |