CN1266911C - Path coordination protocol for large scale parallel path apparatus structure - Google Patents
Path coordination protocol for large scale parallel path apparatus structure Download PDFInfo
- Publication number
- CN1266911C CN1266911C CN 02157501 CN02157501A CN1266911C CN 1266911 C CN1266911 C CN 1266911C CN 02157501 CN02157501 CN 02157501 CN 02157501 A CN02157501 A CN 02157501A CN 1266911 C CN1266911 C CN 1266911C
- Authority
- CN
- China
- Prior art keywords
- routing
- data packet
- message data
- node
- link state
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
A parallel router comprising: 1) a plurality of routing nodes, each of the plurality of routing nodes capable of receiving message packets from and transmitting message packets to external devices, wherein each of the routing nodes maintains a routing table suitable for routing message packets from transmitting ones of the plurality of routing nodes to receiving ones of the plurality of routing nodes; and 2) a switch fabric capable of transmitting the messages packets between the transmitting nodes and the receiving nodes, wherein a designated one of the plurality of routing nodes is operable to receive from non-designated ones of the plurality of routing nodes link state advertisement (LSA) message packets containing link state information, wherein the designated routing node aggregates the LSA message packets to thereby form a link-state database.
Description
Related application
The present invention require application on October 5 calendar year 2001, application number be No.60/327494's and application on October 5 calendar year 2001, application number is the preference of U.S.'s patent application formerly of No.60/327230.
The cross reference of related application
The present invention is relevant with disclosed content in the following U.S. Patent application:
1) application number of October 5 calendar year 2001 application is that the name of No.60/327494 is called the sharp application of the state of formerly having an exclusive claim to fame of " matchmaking protocol of loose coupling large-scale parallel router ";
2) application number of October 5 calendar year 2001 application is that the name of No.60/327230 is called sharp application of the state of formerly having an exclusive claim to fame of " the redundancy scheme agreement of many-gigabit switch router ";
3) application number of applying for simultaneously with the application is called the sharp application of the state of having an exclusive claim to fame of " the redundancy scheme agreement of large-scale parallel router " for the name of [SAMS-00194].
Above-mentioned application is all awarded and is given of the present invention awarding and allow the people.Therefore these related application are quoted the purpose that all is used for reference here.
Technical field
The present invention relates generally to the large-scale parallel router, relate more specifically to be used for the communication protocol of loose coupling large-scale parallel router.
Background technology
Internet user's quantity increases, and from these users' various demands for services, the realization of new service such as the development of the speech business (VoIP) by IP or streaming applications, mobile Internet, has caused the rapid growth of Internet service.Play the conventional router that connects each sub-network or other router effect in process data packets, determine that time of requirement that its destination and transferring data be grouped into the destination usually less than under the transmission time situation on the network path, realized its good effect.But, the data packet transmission capacity in recent high broadband networks path and the processing capacity that combines and surpassed conventional router with the increase of Internet service.Therefore, router becomes the main bottleneck that can't raise speed in the Internet.
Early stage router realizes on main frame, so that the CPU of main frame carries out all management roles such as the transfer of the packet of calculating through shared bus and routing table.But, because the overhead of concentrating of CPU and on bus, exist congested, so this unformatted structure is inefficient.As a result, the router developer of exploitation distribution router structure provides than the more effective packet of concentrating structure and has handled.In the distribution router structure, be to distribute to each Line cards and bus that share is substituted by the high speed crossbar switch before many by the functions of concentrating CPU to carry out.
Fig. 1 represents the distribution router 100 according to the exemplary embodiment of prior art.Distribution router 100 is connected with networks of different type among Fig. 1, and networks of different type comprises optical networking (OC-192), ATM(Asynchronous Transfer Mode) net and as gigabit Ethernet and other net (not shown) of network interface.Distribution router 100 comprises Line cards module (LCMs) 111-113, switching fabric structure (switch fabric) 130, route processors 140 and Line cards module (LCMs) 151-153, contains LCM 111, LCM 112 and the LCM 113 of transfer table (FT) 121, transfer table (FT) 122 and transfer table (FT) 123 respectively.Equally, LCM151, LCM 152 and LCM 153 contain transfer table (FT) 161, transfer table (FT) 162 and transfer table (FT) 163 respectively.
Receive from the grouped data of neighboring router or sub-network and send to exchanger arrangement 130 by Line cards module (LCMs) 111-113 and Line cards module (LCMs) 151-153.Switching fabric 130 conversions are from Line cards module 111-113 and Line cards module 151-153 or pass to Line cards module 111-113 and the packet of Line cards module 151-153, and play requisite effect in relay data packets.
Route processors 140 is equipped with routing table 141 and keeps the current state of routing table 141 by the route of the variation that upgrades in time.Route processors 140 utilizes the operation Routing Protocols to keep routing tables 141, and these Routing Protocols are in this way all: routing information protocol (rip), Open Shortest Path First (OSPF) or border gateway protocol (BGP).Transfer table 121-123 and 161-163 are supported in effectively tabling look-up and be to download from the routing table 141 of route processors 140 in each Line cards.If can not find the path of the destination of Line cards module by the leading-in box of Line cards module from transfer table, then Dui Ying packet can be transferred to predetermined default route by switching fabric 130, perhaps can be abandoned quietly on Line cards.
It is the simplicity that focuses on the management routing table with conforming mode utilization that the main cause of distributed architecture is liked by the manufacturer of router.In addition on the one hand, though separately can handling of route and forwarding function to high-speed data packets, but Qos-ability route service is introduced and cause that by the unsteadiness of network route postpones to require even faster data packet transaction ability, thereby causes the additional overhead (overhead) of route processors or the unsteadiness of router itself.
If set up effective setting of interactive operation rule, so a large amount of little router (that is, parallel) is in the same manner operated.Industrial quarters has utilized single channel to avoid this coordination problem by the server process routing issue.Therefore, it had not only limited the router scale and but also had limited maximum performance to the scale of available microprocessors disposal ability.
Therefore on prior art, exist the demand of improving the large-scale parallel router.Particularly, exist the demand to the large-scale parallel router with distributed architecture, this structure realizes active data grouping Routing Protocol, but does not limit the maximum performance of the scale of this router and available microprocessors disposal ability thereof.
Summary of the invention
Loose coupling unified standard environment (LUE) route collaboration protocols according to the principle of the invention is designed to go up the minimizing traffic carrying capacity at the routing node (RN) of the virtual region that may produce heavily communication service.The present invention proposes the improvement of following several uniquenesses.Routing Protocol allows to collect each adjacent network and each main frame that will be grouped in Open Shortest Path First (OSPF) territory.This group together with distribution have with comprise network interface routing infrastructure be called as a district.The topology in district be can't see from the outside in district.Know nothing the outside topology in detail in this district the router node inside of given area.The isolation technology of this knowledge makes that the LUE agreement that proposes is that single link field realizes on the routing to communicate business that relatively mark reduces with being treated to whole autonomous system.The routing node that belongs to same zone has the same zone link state database.
The support of routing node agreement must comprise that the network number of the proximity network of set IP classification A, B, C is the individual shared wireless network (supernet) of larger amt.For the quantity of the summary-link state advertisement in the minimizing system (LSA) packet, each RN gathers its route entry and sends their routing nodes (DRN) to appointment.Flooding scheme (flooding scheme) is a kind of uneconomic scheme for exchange LSA packet.Each RN can visit other RN by switching fabric.In this scheme, when having N routing node, the message complexity of flooding scheme equals O (N
2).The parallel router structure realizes a kind of Star topology, utilizing 2 switching processors of distribution (SWP) to reduce messaging service for a DRN and a Backup DR N is O (N), therefore competes mutually with the complexity of concentrated route and distribution transferring router structure as shown in Figure 1.
In order to reduce in each routing node (or backup) IOP or SWP desired control message of determining appointment, when system initialization, do not select DRN and Backup DR N, the SWP that wherein has minimum IP address elects DRN as with competing, therefore causes message complexity O (1) to replace O (N
2).
In order to overcome the above-mentioned defective of prior art, main purpose of the present invention provides a kind of distribution router of improvement.According to useful embodiment of the present invention, this distribution router comprises: 1) a plurality of router node devices, each of described a plurality of routing node devices can receive message data packet and can send message data packet to external device (ED) from external device (ED), each of wherein said a plurality of routing node devices all keeps routing table, and described routing table is used for message data packet is routed to receiving node device described a plurality of routing node device from the sending node device of described a plurality of routing node devices; 2) switch can transmit described message data packet between described sending node device and described receiving node device; A specified circuit in wherein said a plurality of routing node device receives the linking status advertisement information packet that contains linking status information by the non-specified node device of node apparatus from described a plurality of routing node devices, the routing node device of wherein said appointment is collected described link state advertisement message data packet, thereby forms a link state database.
According to one embodiment of the present of invention, the routing node device of appointment periodically sends the LSA message data packet of collection at least one by node apparatus of each non-specified circuit.
According to an alternative embodiment of the invention, at least one non-specified circuit by node apparatus use the LSA message data packet of this collection upgrade with at least one non-specified circuit by the relevant routing table of node apparatus.
According to another embodiment more of the present invention, the LSA message data packet that the routing node device of appointment is periodically broadcasted collection arrives all non-specified circuits by node apparatus.
According to another embodiment of the present invention, the LSA message data packet that each non-specified circuit use to be collected by node apparatus upgrade with each non-specified circuit by the relevant routing table of node apparatus.
According to additional embodiments of the present invention, switch comprises a plurality of switching processors that can route messages packet between each routing node device.
According to one embodiment of the present of invention, each routing node device comprises and can send the input-output processor (IOP) that message data packet arrives a plurality of router node devices and receives message data packet from a plurality of router node devices through switch.
Summarized out characteristics of the present invention and technological merit above, the professional and technical personnel can understand following details of the present invention better so quite widely.Feature of the present invention and advantage will be described from the theme of claims of the present invention.The professional and technical personnel should be understood that in order to realize identical order of the present invention, they can use disclosed notion and specific embodiment to revise or design other structure at an easy rate.The professional and technical personnel it should also be appreciated that such some equivalent structures do not exceed spirit and the scope with the wideest form of the present invention.
Before the detailed description of the present invention below understanding, some speech of use and the definition of phrase all the time can be described valuably in this invention file; Term " comprises " and " composition " and derivative thereof that the meaning is that without stint comprises; Term " perhaps " comprise and/or the meaning; Phrase " with interrelating " and " with relevant " and derivative thereof, can refer to that the meaning is, be included in wherein, with interconnected, contain, with which, be connected to or with is connected, is coupled to or with coupling, can with communicate by letter, with cooperate, intersect, arranged side by side, be close to, be restricted to or with limit, have, have specific character or the like; Term " controller " means device, system or its part of an operation of any control, and this device can be used hardware, firmware or software, and perhaps its combination of two kinds is implemented.Should be noted that the function relevant with any specific controller can be concentrate or distribute, no matter be local or far-end.Some speech that in this invention file, provides all the time and the definition of phrase, the professional and technical personnel should be understood to, and if not most occasion, uses the speech and the phrase of this qualification, and many this definition can be used former and in the future.
Description of drawings
In order intactly to understand the present invention and advantage thereof, existing wherein same numeral is represented identical object in conjunction with the accompanying drawings with reference to following description, wherein:
Fig. 1 represents the distribution router structure according to the exemplary embodiment of prior art;
Fig. 2 represents the large-scale parallel router topology according to improved collaboration protocols of the present invention;
Fig. 3 represents in the input-output processor at routing node (IOP) according to the principle of the invention and the interactive operation of the software module in the switching processor (SWP);
Fig. 4 is the explanation according to the packet of the loose coupling unified standard environment (LUE) of exemplary embodiment of the present;
Fig. 5 is the explanation of database description (DD) packet according to exemplary embodiment of the present;
Fig. 6 is the message flow diagram in the DD packet of specifying the LSA packet that forms exchange between routing node (DRN) and the non-appointment routing node (non-DRN) according to exemplary embodiment of the present;
Fig. 7 is the DD data packet message flow graph that forms the LSA packet according to the utilization of exemplary embodiment of the present from the summary-LSA that specifies routing node (DRN) to send to non-appointment routing node (non-DRN).
The various embodiment that the Fig. 2 that discusses below in file of the present invention describes principle of the present invention to Fig. 7 and being used to only be the explanation mode and do not limit the present invention in any way.The professional and technical personnel will be interpreted as, principle of the present invention can be implemented in any mode that is fit to be arranged to parallel router.
Fig. 2 represents to use the parallel router structure 200 through improved route collaboration protocols according to principle of the present invention.Parallel router structure 200 is utilized nearly, and N independent routing node (RN) provides scale scalability and high-performance, comprise the exemplary routing node 210,220,230 and 240 that connects by switch 249, this switch is made up of a pair of high speed switching fabric 250A and 250B.Each routing node comprises input-output processor (IOP) and one or more physical media device (PMD).Exemplary RN 210 comprises PMD 212 (label PMD-A), PMD 214 (label PMD-B) and IOP 216.RN 220 comprises PMD 222 (label PMD-A), PMD 224 (label PMD-B) and IOP-216.RN 230 comprises PMD 232 (label PMD-A), PMD 234 (label PMD-B) and IOP 236.At last, exemplary RN 240 comprises PMD 242 (label PMD-A), PMD 244 (label PMD-B) and IOP246.
Each buffer memory of IOP216, IOP226, IOP236 and IOP246 is brought from each subnet such as router two 90 and network 295 or the input Internet protocol (IP packet) of neighbor router.Each of IOP216, IOP226, IOP236 and IOP246 institute's requested service of also classifying is searched destination-address and transferring data is grouped into departure IOP from the packet head end.In addition, each IOP also keeps the internal routing table determined by routing protocol data grouping and calculates short data path from routing table.Each IOP handles the input packet from one of its each PMD module.According to embodiments of the invention, the bus mapping function will be handled and carry out to each PMD card framing (structure) from an input packet (or unit) of IP network (or ATM switch) in IOP.
Routing node 210,220,230 and 240 that each is made of IOP and PMD and that linked by switching fabric 250A and 250B is equivalent to a router itself basically.The present invention proposes a kind of general and scalable router topology of being made up of a plurality of RN that are connected with 250B by high speed switching fabric 250A of scale.Therefore, parallel router structure 200 can think that one group of high-speed link is connected to the RN standarized component of each piece.Exchange such as the packet between exemplary exchange processor (SWP) 255A that is arranged in switching fabric 250A and 250B and the back-up system management of 255B difference and each IOP.Therefore the parts that parallel router structure 200 can utilize market to choose respectively constitute, and cause by adding each standarized component that the cost aspect is competitive to the switching fabric, flexibility and scalability.
Different with conventional router, parallel router structure 200 requires the effective mechanism of the activity (activity) of each routing node 210,220,230 of supervision and 240.The present invention introduces a kind of route collaboration protocols of novelty, be called loose coupling unified standard environment (LUE) agreement, it can be used for connecting the independent routing node that all play the individual router effect by keeping the database to the coordination linking status of each routing node.Loose coupling unified standard environment (LUE) agreement be according to the design concept of OSPF (Open Shortest Path First) Routing Protocol and be by at the demons (port monitoring program (daemon)) of routing node 210,220,230 and 240 in each by executed in parallel, and in SWP 255A and 255B, in RN 210,220,230 and 240, select the RN of an appointment, and synchronous whole routing table.As everyone knows, demons are a kind of Agents, continued operation and provide resource to client on processing node.Demons are the background process that are used to handle the low-level operation system task.In order to implement effectively, during system initialization, specify RN to distribute to main SWP and a backup specifies RN to distribute to backup SWP for one.
Fig. 3 represents in the input-output processor at routing node (IOP) according to the principle of the invention and the interactive operation of the software module in switching processor (SWP).Suppose RN 210 (or IOP 216), RN 220 (or IOP 226), RN 230 (or IOP 236) and RN 240 (or IOP 246), be initialised with SWP 255A and SWP 255B, and maintenance activates, and LUE router demons 320, appointment LUE router demons 330 and backup specify LUE router demons 340 to move on the routing node separately such as RN 216, appointment (first-selection) SWP 255A and backup SWP 255B.Operation by LUE router demons causes the change route entry, such as the LUE router demons 320 of appointment, is reflected to nuclear core routing table by the nuclear core routing table demons such as nuclear core routing table demons 310.
In IOP 216, IOP 226, IOP 236, IOP 246, each of route demons such as routing information protocol (rip) demons 350, Open Shortest Path First (OSPF) demons 360, border gateway protocol (BGP) demons 370 through nuclear core routing table demons 310 exchanging routing informations.LUE router demons 320 in IOP 216 have through for example network communication interface (socket communication) to nuclear core routing table demons 310.Be arranged in and specify each system processor of SWP 255A and backup SWP 255B must have the coordination routing iinformation of collecting from the LUE of each IOP demons.This in order to guarantee for true, LUE router demons 320,330 and 340 each have by the coordination LSD (LSDB) of specifying LUE router demons to keep.
Different with other route software module, because it only carries out cooperation, each LUE router demons does not remain in its routing table, and keeps synchronously in each routing table of each IOP.This makes all IOP have the global coordination routing table, because according to user's angle, all IOP carry out work as a router.
Each RN and each SWP are connected on the radio network.During system initialization, two SWP are distributed to one respectively and are specified a routing node (DRN) and a route stand-by node (non-DRN).Another mode, selection algorithm are suitable for use in the OSPF Routing Protocol demons O (N of the message complexity in the point to point network
2) in and broadcasting or NBMA (nonbroadcast multiaccess) network in the message complexity in, wherein N is the routing node number.In the present invention, the message complexity is reduced to just O (1).
Fig. 4 is the packet of expression according to the loose coupling unified standard environment (LUE) of embodiments of the invention.Because under the ospf protocol situation, IP network layer (being represented by IP head end 410) operation is directly passed through in the LUE packet, and do not require the service of UDP or Transmission Control Protocol.When the routing node reception had the IP packet of IP protocol number=99, it can be specialized in other suitable agreement and use, and this routing node specified data grouping comprises a LUE Payload.Remove its IP head end 410, this routing node identification comprises the LUE packet of LUE head end 420 and LUE Payload 430.
When being changed by the routing table of nuclear core routing table demons 310 management owing to connect the input packet from the outside of the IOP of correspondence, database description (DD) packet sends to the SWP of appointment from IOP.In addition, specify SWP periodically (or event driven manner) broadcasting link state advertisement (LSA) message to activating IOP.The content of link state database is also described in the DD packet.Many DD packet can be used to describe entire database, but only set DD packet if possible, then sends to from IOP and specifies SWP.
When each packet that sends during greater than network peak transfer rate (MTU), the LUE router protocol depends on IP fragmentation (fragmentation).The length of LUE packet can reach 65535 bytes, comprises IP head end 410.The LUE agreement is used IP agreement several 99.In order to be positioned at the routing table of each IOP synchronously, the present invention uses the database description packet, and wherein type field is set to 2 values.
Each linking status advertisement information is described one section LUE router protocol demons.All LSA message send to according to point-to-point mode normal LUE demons from each IOP and specify SWP LUE router demons.The set of LSA is called as link state database in specifying LUE router demons.Specify the LSA packet of set that LUE router demons periodically broadcast it to the normal LUE router demons that are positioned at each IOP.
LUE Payload 430 can be further divided into two parts: LSA head end 440 and LSA main body 450.The LUE agreement can be ignored the school inspection and and confirm that each field is in the hope of efficient.LSA head end 440 is 20 byte head ends of standard.LSA head end 440 comprises: linking status (LS) deadline field 441, linking status type field 442, linking status identification (ID) field 443, advertising router field 444, LS sequence-number field 445 and length field 446.Head end contains the enough information of unique identification LSA.LS time limit and LS sequence-number field are used for determining which situation is more forward.
LUE main body 450 comprises net mask field 451 and metric field 452.The IP address mask of net mask field 451 expression specified networks.For example, when the position of advertisement Class A Network, can use 0xff000000." cost " of metric field 452 these routes of expression.This value is represented as the identical unit of interface cost with router-LSA in ospf protocol.
Fig. 5 is the explanation of database description (DD) packet according to exemplary embodiment of the present.The DD packet comprises: interface mtu field 501, database description sequence-number field 502, #LSA field 503 and LSA field 504.Interface mtu field 501 contains the byte number that burst (fragmentation) not sends to the maximum IP datagram of relevant interface.DD sequence-number field 502 is used for the set ordering to the grouping of database data of description.Initial value should be unique.Then, DD sequence-number field 502 increases progressively, and describes being sent out until complete database.#LSA field 503 contains the number of each LSA that is included in route reflection (route reflection).At last, linking status advertisement (LSA) field 504 comprises remaining DD packet and comprises set (possible part) table of link state database section, wherein under corresponding to the IOP situation of being represented by summary-LSA packet, each LSA describes its oneself link state database.
LUE router demons are the summary-LSA of route-type-3 only.Be included in the summary 3-LSA form from each route of nuclear core routing table demons set at each IOP.In addition, the aggr_lsa packet of broadcasting from the appointment LUE router demons of Fig. 7 has identical LSA form with the DD packet.When describing default summary route, the linking status ID of summary-LSA is set to default designated value (0,0,0,0) all the time and net mask is set to 0,0,0,0.
Fig. 6 is the message flow diagram in the DD packet of specifying the LSA packet that forms exchange between routing node (DRN) and the non-appointment routing node (non-DRN) according to exemplary embodiment of the present.For reliable LSDB (link state database) exchange is provided in each RN of virtual area, the network linking of supposing to be connected to non--DRN 605A (label is NON-DRN1) and non--DRN 605B (label is NON-DRN1) DRN 610 is reliable.If non--DRN605A is from examining the route entry that core routing table demons 310 receive set, then non--DRN605A has DD sequence number=X by transmission database description (DD) packet is responded to the appointment LUE router demons (message 621) in DRN610.If non--DRN605B is from examining the route entry that core routing table demons 310 receive set, then non--DRN605B has DD sequence number=Y by transmission database description (DD) packet is responded to the appointment LUE router demons (message 622) in DRN610.
After receiving the DD packet that contains summary-LSA message, DRN 610 holds it among the LSDB of oneself.If non--DRN 605B receives additional set route entry from nuclear core routing table demons 310, then non--DRN 605B has the DD packet of sequence number=Y+1 to being responded at DRN 610 (message 623) by transmission.If non--DRN 605A receives additional set route entry from nuclear core routing table demons 310, then non--DRN 605A has the DD packet of sequence number=X+1 to being responded at DRN 610 (message 624) by transmission.
In order to reduce the number of the LSA between DRN and the non--DRN, from corresponding RN, be collected in LSA head end and LSA Payload in the DD packet by nuclear core routing table demons management routing table.When the DD of each LSA with collection packet arrives a DRN, in the LSDB of this DRN, upgrade LSA message.The LUE router demons of appointment in DRN periodically respectively are routed to non--DRN by what the DD packet form with summary-LSA Payload (being called " aggr_LSA ") was broadcasted its collection.
Fig. 7 is the LSA data packet message flow graph of utilization from specifying routing node (DRN) to collect the DD packet with summary-LSA broadcasting to non-appointment routing node (non-DRN) according to exemplary embodiment of the present.The time cycle of the qualification of the LSA packet (aggr-LSA) of collecting in message 705,710,715,720,725 and 730 by being called RxmtInterval sends.When DRN 610 receives the aggr-LSA packet, LUE router demons 320 in each non--DRN605A, 605B switch to nuclear core routing table demons 310 with it, upgrade routing table and reflect that all routes among each IOP change in these demons.
Under the home of ospf protocol, upgrade each LSA in the link state database by each all period interval at least (for example, per 30 minutes once).Be updated among the LSA after at this interval, this LSA is assumed to be no longer valid and removes from database.The time span of LS deadline field 441 expression from dying after LSA is updated at last.Remain in all LSA in the link state database of specifying RN, until they expirations.When since its LS of expiration in DRN by when the LS database is removed, LSA message is broadcast to all non--DRN, guarantees that all RN remove LSA simultaneously being similar to, and does not depend on synchronised clock.Then, all other non--DRN remove each LSA from the LSA that is complementary with " MaxAge " of its database replication broadcasting by DRN, take memory and computational workload with minimizing.
Network multicast ability allows an application program to send a single datagram, and this datagram will be sent to a plurality of receivers.Application program presents the communication pattern of one-to-many and multi-to-multi, moves to the approving and forwarding of a large amount of receivers such as walk back and forth true and identical data group of many people videoconference, video conference, remote teaching, distributed interactive, and discovery is very useful.Host computer using the Internet packet slave protocol (IGMP) can dynamically insert and separate the multicast group, the router of the multicast ability of the current slave mode of maintenance notice main frame.In the present invention, each RN receives grouping-subordinate relation LSA message, send it to DRN and then the corresponding LSA message of DRN broadcasting share the coordination link state database to all the other each RN.
Though the present invention is described in detail, the professional and technical personnel should be appreciated that not exceeding it can make various variations, replacement and change under situation of wide form.
Claims (18)
1. parallel router comprises:
A plurality of routing node devices, each of described a plurality of routing node devices can receive message data packet and can send message data packet to external device (ED) from external device (ED), each of wherein said a plurality of routing node devices all keeps routing table, and described routing table is used for message data packet is routed to receiving node device described a plurality of routing node device from the sending node device of described a plurality of routing node devices;
Switch can transmit described message data packet between described sending node device and described receiving node device;
A specified circuit in wherein said a plurality of routing node device receives the linking status advertisement information packet that contains linking status information by the non-specified node device of node apparatus from described a plurality of routing node devices, the routing node device of wherein said appointment is collected described link state advertisement message data packet, thereby forms a link state database.
2. parallel router as claimed in claim 1, wherein said specified circuit are sent described collection by node apparatus link state advertisement message data packet to each described non-specified circuit by node apparatus one of at least.
3. parallel router as claimed in claim 2, wherein said specified circuit are sent the link state advertisement message data packet of described collection by the node apparatus cycle.
4. parallel router as claimed in claim 3, wherein said at least one non-specified circuit by node apparatus use the link state advertisement message data packet of described collection upgrade with described at least one non-specified circuit by the relevant described routing table of node apparatus.
5. the described parallel router of claim 1, wherein said specified circuit are broadcasted described collection by node apparatus link state advertisement message data packet to each non-specified circuit by node apparatus.
6. the described parallel router of claim 5, wherein said specified circuit is by the link state advertisement message data packet of the described collection of node apparatus periodic broadcasting.
7. the described parallel router of claim 6, wherein said each non-specified circuit by node apparatus use the link state advertisement message data packet of described collection upgrade with described each non-specified circuit by the relevant described routing table of node apparatus.
8. the described parallel router of claim 7, wherein said switch comprise a plurality of switching processors that can routing iinformation packet between described each routing node device.
9. the described parallel router of claim 8, wherein said each routing node device comprises input-output processor, can send described information-data packet through described switch and receive described information-data packet to described a plurality of routing node devices with from described a plurality of routing node devices.
10. communication network, this communication network comprise a plurality of parallel routers that can routing iinformation packet between each telecommunication installation that is connected on the described communication network, and each described parallel router comprises:
A plurality of routing node devices, each of described a plurality of routing node devices can receive message data packet and can send message data packet to external device (ED) from external device (ED), each of wherein said a plurality of routing node devices all keeps routing table, and described routing table is used for message data packet is routed to receiving node device described a plurality of routing node device from the sending node device of described a plurality of routing node devices;
Switch can transmit described message data packet between described sending node device and described receiving node device;
The non-specified node device of a specified node device in wherein said a plurality of routing node device from described a plurality of routing node devices receives the linking status advertisement information packet that contains linking status information, the routing node device of wherein said appointment is collected described link state advertisement message data packet, thereby forms a link state database.
11. the described communication network of claim 10, wherein said specified circuit are sent described collection by node apparatus link state advertisement message data packet to each described non-specified circuit by node apparatus one of at least.
12. the described communication network of claim 11, wherein said specified circuit are sent the link state advertisement message data packet of described collection by the node apparatus cycle.
13. the described communication network of claim 12, wherein said at least one non-specified circuit by node apparatus use the link state advertisement message data packet of described collection upgrade with described at least one non-specified circuit by the relevant described routing table of node apparatus.
14. the described communication network of claim 10, wherein said specified circuit are broadcasted described collection by node apparatus link state advertisement message data packet to each non-specified circuit by node apparatus.
15. the described communication network of claim 14, wherein said specified circuit is by the link state advertisement message data packet of the described collection of node apparatus periodic broadcasting.
16. the described communication network of claim 15, wherein said each non-specified circuit by node apparatus use the link state advertisement message data packet of described collection upgrade with described each non-specified circuit by the relevant described routing table of node apparatus.
17. the described communication network of claim 16, wherein said switch comprise a plurality of switching processors that can routing iinformation packet between described each routing node device.
18. the described communication network of claim 17, wherein said each routing node device comprises input-output processor, can send described information-data packet through described switch and receive described information-data packet to described a plurality of routing node devices with from described a plurality of routing node devices.
Applications Claiming Priority (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US32749401P | 2001-10-05 | 2001-10-05 | |
US60/327,494 | 2001-10-05 | ||
US10/193,426 | 2002-07-11 |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1430393A CN1430393A (en) | 2003-07-16 |
CN1266911C true CN1266911C (en) | 2006-07-26 |
Family
ID=23276762
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 02157501 Expired - Fee Related CN1266911C (en) | 2001-10-05 | 2002-10-08 | Path coordination protocol for large scale parallel path apparatus structure |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN1266911C (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7463654B2 (en) * | 2003-12-22 | 2008-12-09 | 3Com Corporation | Stackable routers employing a routing protocol |
CN1943184B (en) * | 2004-02-02 | 2011-08-03 | 艾利森电话股份有限公司 | Adaptive router architecture enabling efficient internal communication |
CN1921432B (en) * | 2006-09-15 | 2011-05-18 | 华为技术有限公司 | Message transmitting method and device |
-
2002
- 2002-10-08 CN CN 02157501 patent/CN1266911C/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
CN1430393A (en) | 2003-07-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7277383B2 (en) | Redundancy mechanization protocol for a massively parallel router | |
US7254111B2 (en) | Routing coordination protocol for a massively parallel router architecture | |
CA2594267C (en) | End-to-end publish/subscribe middleware architecture | |
US9253243B2 (en) | Systems and methods for network virtualization | |
CN102377666B (en) | Flooding-based routing protocol having average-rate and burst-rate control | |
US6934292B1 (en) | Method and system for emulating a single router in a switch stack | |
CN1216480C (en) | Redundance mechanism protocol used on large scale parallel router | |
US7623474B2 (en) | Techniques for distributing information using multicast subsets | |
JP3925188B2 (en) | Application layer multicast method and relay node system | |
US8170044B2 (en) | Pipeline method and system for switching packets | |
US8238324B2 (en) | Method and system for network aware virtual machines | |
Li et al. | RDCM: Reliable data center multicast | |
WO2020078043A1 (en) | Method for determining df of multicast stream, device, and system | |
US20240267324A1 (en) | Packet forwarding method and apparatus | |
US7471676B2 (en) | Apparatus and method for distributing control plane functions in a multiprocessor router | |
CN100359875C (en) | Method for realizing backup and load shared equally based on proxy of address resolution protocol | |
CN1266911C (en) | Path coordination protocol for large scale parallel path apparatus structure | |
KR100450930B1 (en) | Routing coordination protocol for a massively parallel router architecture | |
WO2022194193A1 (en) | Method and apparatus for acquiring path | |
Lowu et al. | On efficient distribution of data in multicast networks: QoS in scalable networks | |
KR100487127B1 (en) | Management method of multicast routing table entries in distributed architecture router | |
Sheng | A New Solution of Multicast Packets Management for Managed Ethernet Switch | |
Kim et al. | Scalable redundant IPC network | |
Li et al. | Grid Communication Environment Based on Multi-RP Multicast Technology |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20060726 Termination date: 20161008 |
|
CF01 | Termination of patent right due to non-payment of annual fee |