CN104811387A - Equal Cost Multi-path With Bit Indexed Explicit Replication - Google Patents

Abstract

The invention discloses an equal cost multi-path with bit indexed explicit replication. Various systems and methods for performing bit indexed explicit replication (BIER). For example, one method involves receiving a packet at a node. The packet includes a bit string. The node selects forwarding information based on a flow value associated with the packet. The forwarding information includes a forwarding bit mask. The node then forwards the packet based on the bit string and the forwarding information.

Description

There is the explicit equal cost multipath copied of position index

related application

Based on united states patent law the 119th article of e money, the temporary patent application number that the application's request was submitted on January 24th, 2014 is 61/931,473, name is called the domestic priority of the patent application of " the bitmask forwarding plane for stateless multiple spot copies ", and the full content of this application is incorporated herein by reference and for all objects.

The number of patent application that the application still submits on September 17th, 2014 is 14/488, 790, name is called the part continuation application of " position index utilizing multiprotocol label to switch is explicit to be copied ", this application again then according to united states patent law the 119th article of e money request in the temporary patent application number that on September 17th, 2013 submits to be 61/878, 693, name is called " having the multicast IPv6 forwarding bitmask ", be 61/931 with the temporary patent application number submitted on January 24th, 2014, 473, the domestic right of the patent application that name is called " bitmask copied for stateless multiple spot forwards structure ".The application or number of patent application are 14/488,761, name is called the part continuation application of " position index is explicit to be copied ", this application then according to united states patent law the 119th article of e money request in the temporary patent application number that on September 17th, 2013 submits to be 61/878,693, name is called " having the multicast IPv6 forwarding bitmask ", be 61/931 with the temporary patent application number submitted on January 24th, 2014,473, the domestic right of the patent application that name is called " bitmask copied for stateless multiple spot forwards structure ".The number of patent application that the application still submits on September 17th, 2014 is 14/488, 810, name is called the part continuation application of " utilize the position index of internet network agreement the 6th edition is explicit to be copied ", this application then according to united states patent law the 119th article of e money request in the temporary patent application number that on September 17th, 2013 submits to be 61/878, 693, name is called " having the multicast IPv6 forwarding bitmask ", be 61/931473 with the temporary patent application number submitted on January 24th, 2014, the domestic right of the patent application that name is called " bitmask copied for stateless multiple spot forwards structure ".For all objects, the full content of every part application of the above-mentioned two pieces provisional application quoted and three non-provisional applications is incorporated herein by reference and for all objects, as them here by complete and all proposing.

Technical field

The application relates generally to network technology, and is specifically related to have the explicit equal cost multipath copied of position index.

Background technology

Target node data.The form of network node can be one or more router, one or more bridge, one or more switch, one or more server, or any communication processing equipment that other is suitable for.These data adopt the form of grouping commonly and utilize transmits forwarding.Be grouped into the format unit of data, it generally includes control information and payload data.Control information can comprise: identify the information (such as, address) of source and destination, the error detection code being similar to School Affairs, sequence information etc.Control information is usually located at head and the afterbody of grouping.Valid data are usually located between the head of grouping and afterbody.

These process concepts forward grouping and relate to different process, although can be simply complicated.The process forwarded involved by grouping changes along with the type of the retransmission method used.In many networks, be more prone to adopt multicast to carry out forwarding data.One of them reason is multicast is a kind of by the bandwidth conservation technology by reducing flow to multiple receiving terminal distributing data.But, in conventional multicast system, employ considerable control plane information.Set up and safeguard that this control information will have the trend becoming complicated and expend computational resource, and limiting factor main in overall performance of network can be become.Another problem of multicast is: due to the packet distribution mechanism used, grouping is sometimes forwarded to less desirable destination.These unnecessary distributions of grouping mean and cause unwelcome burden to network performance.Overcome this burden by traditional approach to relate to generation and safeguard even more control plane information.

Summary of the invention

According to embodiment of the present disclosure, propose a kind of method, comprising: receive grouping at node, wherein grouping comprises bit string; From many group forwarding informations, select first group of forwarding information, wherein this selection is based on flow valuve, and first group of forwarding information comprises forwarding bitmask; And forward grouping based on bit string and first group of forwarding information.

According to embodiment of the present disclosure, also proposed a kind of network equipment, comprising: memory, this memory stores organizes forwarding information more; Network interface, this network interface is configured to receive grouping, and wherein grouping comprises bit string; And processor, this processor is configured to: from many group forwarding informations, select first group of forwarding information, wherein select first group of forwarding information to carry out based on flow valuve, and first group of forwarding information comprises forwarding bitmask; And forward grouping based on bit string and first group of forwarding information.

According to embodiment of the present disclosure, also proposed a kind of system, comprising: storage device, this storage device organizes forwarding information for storing more; Network Interface Unit, this Network Interface Unit is for receiving grouping, and wherein grouping comprises bit string; And processing unit, this processing unit is used for: from many group forwarding informations, select first group of forwarding information, wherein select first group of forwarding information to carry out based on flow valuve, and first group of forwarding information comprises forwarding bitmask; And forward grouping based on bit string and first group of forwarding information.

Accompanying drawing explanation

By reference to accompanying drawing, the disclosure can be better understood, and its numerous objects, feature and advantage become clearer to those skilled in the art.

Fig. 1 shows the simplified block diagram of some assembly of example network.

Fig. 2 shows the simplified block diagram of some assembly of example network.

Fig. 3 A is the flow chart according to present specification shows the instantiation procedure that node adopts.

Fig. 3 B is the flow chart according to present specification shows the instantiation procedure that node adopts.

Fig. 4 shows the simplified block diagram of some assembly of example network.

Fig. 5 A is the sample table generated by node according to this specification.

Fig. 5 B is the sample table generated by node according to this specification.

Fig. 6 is the sample table generated by node according to this specification.

Fig. 7 A is the sample table generated by node according to this specification.

Fig. 7 B is the sample table generated by node according to this specification.

Fig. 7 C is the sample table generated by node according to this specification.

Fig. 7 D is the sample table generated by node according to this specification.

Fig. 8 is the sample table generated by node according to this specification.

Fig. 9 is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 10 is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 11 is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 12 is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 13 A is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 13 B is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 14 is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 15 is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 16 A is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 16 B is the flow chart according to present specification shows the instantiation procedure that node adopts.

Figure 17 is can by the block diagram of some assembly of example endpoint used according to present specification shows.

Figure 18 is the block diagram of the computer system being suitable for the embodiment realizing system described here.

Figure 19 is the block diagram of the network equipment being applicable to the embodiment realizing system as described herein.

Embodiment

general introduction

There has been described for the explicit various system and methods copying (bit indexed explicit replication, BIER) of execute bit index.Such as, a kind of method is included in reception at Node grouping.This grouping comprises bit string (bit string).This node selects forwarding information based on the flow valuve (flowvalue) be associated with this grouping.Forwarding information comprises forwarding bitmask.Then, this node forwards grouping based on bit string and forwarding information.

multicast

Multi-case data divides into groups by multicast, and (generally including the packet of the information (such as, Multicast group address) of mark multicast group) is distributed to multiple receiving terminal from source, and unduly increases the weight of the burden in source.Term used herein " receiving terminal " represents the main frame (such as, computer equipment or application) subscribing to multicast group.Not copy multi-case data by source divide into groups and the copy that this multi-case data divides into groups is sent to each receiving terminal, but sent the single copy of multi-case data grouping by source, and this grouping is copied at (one or more) some place that the router that multicast is enable (here referred to as node) is diverging towards the path of each receiving terminal.Multicast route protocol copies the next enable multicast transmission of this multi-case data grouping (that is, one-to-many connects and multi-to-multi connection) by being located in the object of dividing into groups close to multi-case data, avoids and uses multiple clean cultures of same purposes to connect.Which save the network bandwidth and improve throughput.

Fig. 1 is the simplified block diagram of the network 100 performing multicast data transmission.Multicast enable node 110,120,130 and 140 is coupled by network link 150,160 and 170.Multicast enable node 110 is also coupled to source 111 and receiving terminal (receiver) 112; Multicast enable node 120 is coupled to receiving terminal 121; Multicast enable node 130 is coupled to receiving terminal 131 and receiving terminal 132; And multicast enable node 140 is coupled to receiving terminal 141.Multicast enable node and such being coupled between source and/or receiving terminal can be also can be directly indirectly (such as, by the L2 network equipment or another node).

In order to this example, source 111 is main frames, and it is configured to send multi-case data grouping to comprising as the multicast group of the main frame 112,121,131,132 and 141 of receiving terminal.Source 111 send to multicast enable node 110 comprise the one or more multi-case datas grouping with common Multicast group address Multicast Flows (by from 111 to 110 arrow shown in).Multicast enable node 110 comprises multicast forwarding, and multicast enable node 110 uses this table to determine divides into groups to where forwarding the multi-case data be associated with this Multicast Flows.Multicast forwarding comprises and identifies the multicast enable node 110 that is connected to multicast distribution tree (MDT) and (such as, have sent the main frame adding message, the information of each interface as mentioned above) to one or more receiving terminals of multicast group.Then, multicast enable node 110 copies the multi-case data grouping in Multicast Flows, and the multi-case data copied grouping is sent to receiving terminal 112, multicast enable node 120 and multicast enable node 130 from identified interface.

Multicast enable node 120 and 130 use adds message (such as, Protocol Independent Multicast (PIM) adds message), and they are coupled to one or more receiving terminal to carry out advised nodes 110.Add message in response to receiving, multicast enable node 110 upgrades its multicast forwarding should be forwarded to identify multi-case data grouping for which interface.Multi-case data grouping can be copied as required by node 110, so that to multicast group receiving terminal (such as, receiving terminal 131 and 132) and MDT on other multicast enable node (such as, multicast enable node 140) provide this multi-case data to divide into groups.In this way, the Multicast Flows from source 111 is transferred to multiple receiving terminal by multicast network.

Can find out, be generally used in multicast for often setting up MDT and the process upgrading multicast forwarding result in state informations a large amount of in network.Especially, the multicast forwarding safeguarded by each multicast enable node can become quite large.Safeguard that such multicast forwarding shows the restriction of network scalability.

position index is explicit to be copied

As described below, some technology are used to receiving terminal information to be attached to grouping with the form of position (bit), and forward grouping based on receiving terminal information.This considerably reduce the state information amount being stored in Nodes, be therefore also referred to as " stateless multicast ".More formally, explicit copy (BIER) of term position index is used to describe these technology.That shows as the term implies is such, and position, position (bitposition) is used as entering the index transmitted, and grouping is only copied to specified node.BIER makes when not using the multicast distribution of each Nodes between source and receiving terminal set and often organize state information, grouping can be forwarded to multiple receiving terminal from source.

Fig. 2 shows example network 200.Network 200 comprises BIER enable node 206-218.BIER enable node is configured to use BIER to forward grouping, and sometimes referred to as position forwarding router (BFR).BIER enable node 206-218 defines provider network or provides the quotient field.Such provider network can be used for transmitting grouping to user by ISP.This territory comprises transmission node 208 and 210, and provider's boundary node 206,214,216 and 218.Provider's boundary node is coupled to user's boundary node 211,213,215 and 217.Main frame 201,203,205 and 207 is the computing equipments being coupled to user's boundary node.

Each BIER enable node 206-218 has the interface identified as shown.Such as, BIER enable node 208 has three interfaces being appointed as 1-3 respectively.Each BIER enable node is assigned with unique identifier or routable address, is called as router identifier (RID).RID such as can be implemented as Internet protocol (IP) address, prefix or loopback address (loopbackaddress).Each BIER enable node is noticed or this routable address of inundation to the every other BIER enable node in network 200.Each BIER enable node uses the routable address be advertised to set up the unicast topologies of the BIER enable node in network 200.In one embodiment, router identifier mathematically can be converted into the set identifier (SI) and position, position (BP) that are assigned to BIER enable node.This conversion is based on the length of used bit string.Such as, in order to router identifier " N " is converted to set identifier and position, position, set identifier is the integer part of the business of (N-1)/bit string length.Position, position is ((N-1) modulo BitStringLength (mould of fetch bit string length))+1.In the above examples, if N equals 257 and bit string length is 256, then SI is 1 and BP is 1.BIER network 200 also comprises the node being configured to be operating as multi-case data controller (MDC) 230.MDC performs configuration and management role, as described below.

The position that BIER enable node 206 is configured to multi-case data grouping forwards ingress router (BFIR).BIER enable node 206 is coupled to source 201 by user's boundary node 211.Multi-case data grouping from source 201 enters BIER network by BFIR (BIER enable node 206).Each BIER enable node 214,216 and 218 is configured to position and forwards egress router (BFER).BFER (directly or by user's border router) can be connected to main frame (such as, receiving terminal) or other networks.BFER is the last BIER enable node on the path between source and receiving terminal.BFER can be that directly or indirectly (such as, by the enable CE node of non-BIER) is coupled to provider border (PE) node of receiving terminal.

distribute the position, position in bit string

Each BFER in BIER network is point co-ordination positions (BP) from the set of position or the array of position.The array (array of bit) of position can be carried in grouping or other internet messages.The array of position also can be stored in be transmitted and/or in routing table.For clarity, term used herein be " bit string " (when position array in a packet time) and " bitmask " (when being stored in table when the array of position).In addition, notice that BFIR can be used as BFER, vice versa.BFIR is also assigned with a position.

Bit string (or bitmask) can have regular length or variable-length.Length for the bit string in BIER network can be statically configured or dynamic assignment, and is distributed in BIER network.In one embodiment, the length of bit string between 256 and 1024, but also can use shorter or longer bit string.In one embodiment, the maximum length of bit string value is determined by the hardware of the BIER enable node in BIER network or software limitations.In one embodiment, the different B IER enable node in BIER network for its separately bit string use different length.Such as, BIER enable node can have the dominant bit string length of 128, and another BIER enable node can have the dominant bit string length of 256.Bit string is the multitransmission entry of a type, and each position wherein in multiple positions is the element that can be used to represent independently node or interface.The other types multitransmission entry with other types element can be used.

Position, position (BP) is statically or dynamically assigned to each BFER.Each BFER should have at least one position, unique position from bit string.In one embodiment, BP can be distributed to BFER by central office (such as, multicast domain controller).In one embodiment, multiple BP is distributed to single BFER by multicast domain controller, such as, to the unique BP of each interface assignment in the one or more interfaces be included in BFER.Other mechanism for distributing BP also can be implemented, and such as, from the router identifier distributing to BIER enable node, derive BP, and wherein this derivation uses mapping algorithm.In certain embodiments, the position, position in bit string is assigned to single BFER.In other embodiments, single BP can be assigned to more than one BFER.When multiple BFER is assigned with same BP, the ownership of this BP can be exercised in preset time for one in the plurality of BFER, and ownership can shift between the plurality of BFER.The ownership of this BP can be transferred in the plurality of BFER due to any one in some reasons another, such as, in response to the Failure Transfer of node or link failure, if or in the plurality of BFER one otherwise becomes unavailable, in response to change network condition, due to consideration that timesharing is shared etc.A BP is distributed to multiple BFER to contribute to being similar to the operation of appointing and broadcasting (anycast), wherein divide into groups the receiving terminal be forwarded in one group of receiving terminal, and each receiving terminal wherein in this group receiving terminal uses common address.

The enable boundary node of the BIER in BIER network (such as, BFER and BFIR) is only had to be assigned with BP.Every other BIER enable node (such as, transmission node) in this network does not need BP to participate in BIER.This contributes to the number reducing the position distributed in network.As shown in the example of figure 2, network 200 uses four long bit strings.Each in four BFER (comprising BFIR node 206) in network 200 is assigned with BP: node 206 is assigned with BP{1000}; Node 214 is assigned with BP{0001}; Node 216 is assigned with BP{0100}; Node 218 is assigned with BP{0010}.

set

The size of the bit string that the quantity that can be addressed the BFER of (distributing BP) is included in multi-case data grouping limited.If bit string be four long, then the quantity of the BFER that can be addressed is four.The concept of set allow for the increase of the quantity of the BFER that can be assigned with BP.Set identifier (SI) is such as the number between 0 and 255.SI allows BP to be unique in the context of set.Such as, each BP can reuse in each set.Have 256 set and bit string length is in the embodiment of 256,65536 (256x256) individual BFER can be supported.In one embodiment, the BIER enable node in BIER network generates respective forwarding information for each SI.Such as, if use two different set identifiers in BIER network, then BIER enable node generates two position indexes and transmits (BIFT), and each corresponds to each SI.In response to receiving the multi-case data grouping with SI, BIER enable node use SI carrys out which forwarding information of choice for use (such as, BIFT) is carried out retransmitting multi-casting packet.

signaling

When receiving terminal (such as, main frame, the main frame 203 of such as Fig. 2) to wish when adding multicast group, this receiving terminal sends information about firms (such as, use IGMP (IGMP)) to the BFER that its (directly or indirectly) is coupled to.The multicast group added wished by information about firms mark main frame, and identifies the source be associated with this multicast group alternatively.In the figure 2 example, main frame 203 can send IGMP message to CE node 213, and then this IGMP message can be transmitted to BIER enable node 214 by CE node 213.Wish to add the message of multicast group in response to receiving instruction receiving terminal, BFER sends its signal for the interest of multicast group identified in the message.In one embodiment, this comprises: BFER sends member's message to any BFIR in network or to controller, and instruction BFER is for the interest of this multicast group.In one embodiment, member's message comprises the BP of BFER.Or the address that BFIR uses the BFER initiating member's message and the information (such as, the information in the BIRT of BFIR) noticed by IGP are performed and search, the BP be associated with the BFER determined with initiate member's message.In one embodiment, BFER uses Border Gateway Protocol (BGP) to send member's message.

Member's message only can be sent to the borderline some or all of BIER enable node (BFIR and BFER) of BIER network by BFER, or can by signaling message inundation to all nodes in network.Such as, if network is using source specific multicast (SSM), then BFER knows the source (such as, the IGMP message according to receiving from receiving terminal) of multicast group, and can find the path of this particular B FIR and only send signaling message to this BFIR.If SSM is not the multicast type used, then signaling message inundation can be given all candidate BFIR by BFER.The enable boundary node of BIER is only had to resolve this message to determine group and BP information, every other node this message discardable.Add multicast group and the disturbance of the state information (such as, position index is transmitted) safeguarded by core BIER enable node can not be caused from the receiving terminal that multicast group is quit the subscription of or need any change in this state information, unlike conventional multicast.On the contrary, to add or unsubscribe message signals to upgrade the bit string that is associated with given multicast group to BFIR.This comprises and only has BFIR more new state information (such as, upgrade the group membership that is associated with this group and show), instead of core node.This indicates the remarkable improvement on conventional multicast, and in conventional multicast, multicast distribution tree is based on adding and unsubscribe message and be established throughout network and remove.

BFIR (such as, the BIER enable node 206 of Fig. 2) safeguards and comprises the state information that this BFIR receives the entry of each multicast group of member's message.In one embodiment, state-maintenance is shown in (GMT), shown in 224 of Fig. 2 group membership by BFIR.In one embodiment, each entry in GMT comprises: the information of mark multicast group (such as, multicast group name claims and/or the address in source of multicast group), interested (such as corresponding to the multicast group that have expressed identifying in group field, by member's message) the BP list of BFER, and mark have expressed the bit string to the interested all BFER of this multicast group (such as, by being provided with position set in corresponding to the position, position that have expressed the interested each BFER of this multicast group).In response to receiving this BFER of instruction from BFER to the interested member's message of multicast group, this BFIR arranges the position of the BP corresponding to BFER in the bit string corresponding to this multicast group.When this BFER is no longer interested in the multi-case data grouping receiving this multicast group, this BFER sends signal (such as, using unsubscribe message) to BFIR, and BFIR removes the corresponding positions in bit string.

BIER network divides into groups at BIER network repeating multi-case data based on bit string.BFIR by bit string and multi-case data transmitted in packets in BIER network.There is the different technologies that much can be used for transmitting bit string.This explanation refers to bit string be encapsulated in multi-case data grouping.This term not only covers and is merged in multi-case data grouping (such as, as head or payload information) by BM, also covers some or all of bit string is additional or is added to multi-case data grouping in advance.

Fig. 3 A shows the flow chart of aforesaid operations.In one embodiment, the method is performed by BFER (such as, the BFER 214 of Fig. 2).BIER information is asked at 302, BFER, such as, position, position and set identifier.In one embodiment, ask BIER information to comprise BFER and send message to multicast domain controller (such as, the multicast domain controller 230 of Fig. 2).In one embodiment, add network in response to detection BFER, or in response to some other conditions, BIER information is automatically provided to BFER.Keeper can use BP and set identifier manual configuration BFER.

Receive BIER information at 304, BFER, this BIER information is not the result as administration configuration, is exactly such as to be included in the message from MDC in response to the request for BIER information.306, receive BIER information in response to BFER, BFER is to other node advertisement some or all of its BIER information and the router identifier thereof in BIER network.In one embodiment, BFER notices itself BP and SI by Interior Gateway Protocol (IGP).Such as, Intermediate System-to-Intermediate System (ISIS) and/or ospf (OSPF) can be modified and distribute this information with auxiliary by BIER Web vector graphic link state update.Also other flooding mechanism can be used to a point photos and sending messages.All BIER enable node (being not only BFER) in BIER network go back its router identifier of inundation, and it is used to set up network topology and unicast forwarding table.In one embodiment, BIER enable node also notices additional information, and such as, BIER enable node is configured the bit string size used.With in conventional multicast based on compared with the state information often organizing maintenance, such BIER information being added to noticed information is additional information relatively in a small amount.

Fig. 3 B shows the exemplary method using and cover signaling (overlay signaling) and distribute information about firms in BIER network.In one embodiment, the method for Fig. 3 B is performed by BFER (such as, the BIER enable node 214 of Fig. 2).

Membership request is received from main frame (such as, the main frame 203 of Fig. 2) at 322, BFER.Membership request carrys out relaying by user's boundary node (such as, user's boundary node 213 of Fig. 2) alternatively.In one embodiment, membership request comprises IGMP message.Membership request comprises the information of mark multicast group, and identifies this main frame and wish to add the information that (such as, subscribing to) this multicast group still leaves this multicast group (such as, quitting the subscription of from this multicast group).In response to reception membership request, BFER upgrades the forwarding information of the member of the main frame in instruction multicast group.Such as, if membership request indicates main frame to wish to add multicast group G1, then BFER upgrades forwarding entry, makes to be received by this BFER and any multi-case data grouping being addressed to multicast group G1 will be transmitted to this main frame by this BFER.

Member's message is generated at 324, BFER.Member's message has showed the interest of BFER to multicast group.In one embodiment, member's message uses BGP to realize.This member's message carries the information of mark multicast group, and the information of mark BFER (such as, the set identifier of BFER and position, position).

Member's message is sent at 326, BFER.In one embodiment, send member's message to relate to multicast domain controller (such as, the MDC 230 of Fig. 2) forwarding member message.Then, MDC sends member's message to the some or all of border routers (such as, BFER and BFIR) in BIER territory.

position index route and transmitting

The BP that each BIER enable node in BIER network uses other BIER enable node to notice and router identifier are to generate one or more position index routing table (BIRT) and position index is transmitted (BIFT).BIER enable node uses these tables to carry out retransmitting multi-casting packet.As shown in the example BIRT 262 of Fig. 2, position index routing table stores the table of BP to the mapping (such as, being learnt by Interior Gateway Protocol (IGP)) of router identifier.Each BIER enable node receives the mapping of BP to router identifier, and is stored in BIRT.Use router identifier, BIER enable node performs recursive lookup to identify the direct down hop BIER enable node (referred to herein as neighbours (NBR)) be connected from this BIER enable node towards the shortest path of the BIER enable node be associated with BP in unicast routing table.In one embodiment, NBR is towards the down hop on the shortest path (SPT) of the BFER noticing BP.In one embodiment, BIRT comprises each BP entry.Following discussion further, when there is the path of multiple equivalence to given BFER, BIRT comprises multiple entries of the BP be associated with this BFER.

Its BIRT is converted to one or more position index and transmits (BIFT) by each BIER enable node.Fig. 2 shows the example BIFT 264 generated by BIER enable node B 208.In one embodiment, generate BIFT to relate to and first being sorted to BIRT by neighbours.For the entry in BIRT with common NBR, the bitmask of these entries by or (OR) logic by and together, thus create bitmask, it is the combination of the BP from these entries.Here this is called as forwarding bitmask (FBM) or is only called bitmask (BM).Each being arranged in FBM indicates the BFER by the neighbours Ke Da be associated with the FBM in BIFT.Multi-case data grouping uses BIFT to forward by BIER enable node.Such as, according to BIFT 264, if having use, { the multi-case data grouping of the bit string of the BP of 0001} set arrives node 208, then the grouping of this multi-case data should be forwarded to NBR C (the BIER enable node 210 in the example of Fig. 2).If had, { the multi-case data grouping of the BP of 1000} set arrives node 208, then the grouping of this multi-case data should be forwarded to NBR A (the BIER enable node 206 in the example of Fig. 2).If had, { the multi-case data grouping of 1001} bit string arrives node 208, then the grouping of this multi-case data should be forwarded to both NBR A and NBR C.

forwarding of packets

Bit string be encapsulated into after in multi-case data grouping, BFIR uses the BFTS of this BFIR by this multi-case data forwarding of packets to one or more BIER enable node.The BIER enable node receiving multi-case data grouping uses the bit string in multi-case data grouping and the BFT of BIER enable node oneself to determine whether one or more in its neighbours of multi-case data forwarding of packets, and if be, be transmitted to which or which neighbours.For this reason, BIER enable node by multi-case data divide into groups in bit string and BIER enable node BFT in FBM entry contrast.In one embodiment, actuating logic and operation between the FBM in the BFT of the BIER enable node bit string of dividing into groups at multi-case data and BIER enable node.As pointed out, in one embodiment, the BFT of BIER enable node comprises the entry of each neighbours of this BIER enable node, and each entry comprises FBM territory, and which BFER this territory indicates can be reached along shortest path by the neighbours identified in this entry.If for given neighbours, with result be true, then BIER enable node by multi-case data forwarding of packets give these neighbours.Result is really indicate the FBM of given neighbours in the BIFT of BIER enable node to have the position being set to 1, and the corresponding positions in the bit string of multi-case data grouping is also set to 1.In the bit string of multi-case data grouping, which BFER the position that is set indicates have expressed interest to multicast group, and the BFER that the position be set in the BIFT entry of BIER enable node instruction have expressed interest can be reached by the neighbours indicated in entry.BIER enable node by the multi-case data forwarding of packets comprising bit string to the step-by-step between the FBM in the BIFT of the bit string in multi-case data grouping and BIER enable node be operating as really all neighbours.

In the figure 2 example, BIER enable node 214 (BFER) sends signal to BIER enable node 206 (BFIR) and shows that BIER enable node 214 is interested for receiving the grouping that is associated with given multicast group or stream.Send signal to BIER enable node 206 (BFIR) shows that BIER enable node 216 and 218 is interested in same multicast group to BIER enable node 216 and 218 equally.Signaling dotted line as shown in Figure 2 represents.BIER enable node 206 upgrades the entry in group membership's table 224 (if there is no, then create) of this multicast group, and by arranging to BIER enable node 214,216 and 218 corresponding positions the bit string upgraded in this entry.Bit string corresponding to this multicast group is { 0111}.

BIER enable node 206 is configured to the multi-case data grouping that reception is addressed to this multicast group or stream (such as, from source 201 via CE node 211).BIER enable node 206 uses and is included in Multicast group address in multi-case data grouping and/or source address visits its GMT and selects the bit string that is associated with multicast group.Select to correspond to the bit string of this multicast group from GMT after, the bit string of this multicast group is encapsulated in multi-case data grouping by BIER enable node 206, and identifies neighbours' (such as, using its BFT) that this grouping will be forwarded to.In one embodiment, perform and operation between this each entry being included in the BIFT of bit string and BIER enable node 206.In this example, in BIFT, only there is an entry, and this entry corresponds to BIER enable node 208.This means from BIER enable node 206 to the shortest path of all three BFER network 200 all by BIER enable node 208.Because for NBR B (BIER enable node 208), with result be true, therefore BIER enable node 206 by multi-case data forwarding of packets to BIER enable node 208.BIER enable node 206 also revises the bit string in its multi-case data grouping forwarded, as described below.

In response to receiving multi-case data grouping, the bit string ({ 0111}) of BIER enable node 208 in multi-case data grouping with performs between each entry in its BIFT (being shown in 264) and operates.Result due to NBR C is true, therefore BIER enable node 208 by multi-case data forwarding of packets to BIER enable node 210.BIER enable node 208 also by remove correspond to through this multi-case data divide into groups the position of the inaccessible BFER of shortest path of the BIER enable node be just forwarded to revise its multi-case data forwarded divide into groups in bit string.In this example, because node E is via node C unreachable (from Node B to the shortest path of node E without C), so Node B clears Node B is transmitted to the position corresponding to node E in the bit string of node C.Therefore, Node B will have bit string { the multi-case data forwarding of packets of 0011} will be to node C.The result of NBR E is also true, and therefore BIER enable node 208 copies multi-case data grouping and is BFER by this multi-case data forwarding of packets to BIER enable node 216, BIER enable node 216.Node B renewal bit string also will have bit string, and { the multi-case data forwarding of packets of 0110} is to node E.

In response to receiving multi-case data grouping, perform between the bit string ({ 0011}) of BIER enable node 210 in multi-case data grouping and each entry in its BIFT and operate, and by multi-case data forwarding of packets to BIER enable node 214 (it is BFER) and BIER enable node 218.BIER enable node 216 also by this multi-case data forwarding of packets to BIER enable node 218.This causes BIER enable node 218 to receive the duplicate copies of multi-case data grouping.This situation be undesired, and be cause owing to there are two paths of equal value from BIER enable node 208 to BIER enable node 218.That is, as shown, node F uses and can reach from the single-hop path of Node B via node C or node E.This is called as equal cost multipath (ECMP), and the technology solving this situation is described below.

ECMP

Forwarding grouping comprises node and determines which neighbour this grouping should be sent to be advanced by the destination of shortest path to this grouping.Sometimes there are the multiple neighbours dividing into groups to be forwarded.Such as, many shortest paths may be there are from this node to destination, or many equative routes.The expense in path can the aspect such as type (and the operating parameter of link on path and/or node, such as bandwidth sum disposal ability) of link on jumping figure, path and/or node be assessed.Node calculates by using such as dijkstra's algorithm and the routing iinformation distributed by IGP to perform shortest path to generate route topological.

When there is many equative routes from node to the destination of grouping, this node should along in these paths and only one forward grouping, such as in the neighbours in respective paths and only one forward grouping.Undesired repetition is caused along forwarding grouping more than a paths (such as, to more than one nodes neighbors).Repeat to represent both to copy and the sending node of multiple copies forwarding grouping causes burden, also to receiving, cushion and the receiving node (one of them grouping just enough) processing (such as, abandoning) multiple grouping causes burden.In addition, packet copies causes burden to the transmission medium (such as, network link) usually with fixed-bandwidth capacity.Determining that grouping should be forwarded on which paths may be complex task.The grouping (such as, as the grouping of the part of same Multicast Flows) that logic is relevant should pass through same neighbourhood.Otherwise, if the grouping that logic is correlated with is distributed to mulitpath, then face the additional attention of the sequence about these groupings in destination.The problem paid close attention to when processing ECMP is load balance.What want is that every bar ECMP path is all used, instead of over-burden in some paths, and other fully do not use.

In typical multicast deployments, grouping uses multicast distribution tree to forward.Node can perform ECMP and calculate as the part setting up multicast distribution to set.Such as, which upstream (towards source) neighbours should be chosen as the receiving terminal adding message and can use ECMP technology to determine.But the forwarding that ECMP is not used to retransmitting multi-casting packet to downstream node in typical multicast deployments determines.

Usually, ECMP jumps (hop-by-hop) to make decision based on a jumper connection one.That is, each node receiving grouping is determined or assesses some the ECMP standards be associated with this grouping, and correspondingly selecting paths.Perform ECMP at each Nodes to determine to produce negative influence to forwarding performance.Do like this in the environment of BIER and can comprise node and determine to there are many ECMP paths, and given grouping should be sent to the node that can be reached by ECMP path.It is complicated for performing these calculating in the dataplane.As replacement, these operate in control plane and combine and arrange transmitting that BIER node uses and perform.Once ECMP path is identified, and the forwarding information of node correspondingly configures for every bar ECMP path, then this node can which forwarding information of choice for use (such as, which bar ECMP path forwarding this grouping) simply.In one embodiment, node uses the information comprised in a packet to select between ECMP path.Such as, grouping can comprise value or the field of the Multicast Flows be associated based on grouping.Based on this value, node can be selected ECMP path and the grouping from same Multicast Flows be forwarded along single ECMP path.

Be mainly concerned with the ECMP path of different neighbours herein.The ECMP path of another type relates to the mulitpath (such as, parallel interface) of same neighbourhood.When mulitpath is to same neighbourhood, the bit string for every bar ECMP path is identical with the bit string for every bar path candidate.On the other hand, if ECMP path is to different neighbours, then the bit string in each ECMP path is different for each neighbours.

Fig. 4 is the simplified block diagram of the specific components showing example network 400.BIER enable node 410 to 416 is the BFR being configured to transmission node.BIER enable node 420 to 428 is the BFR arranged as BFER, and wherein distributed position, position as shown in the figure.BIER enable node 406 is can be the BFR of transmission node or BFIR.In one embodiment, this BIER enable node (such as, by one or more user's boundary node (not shown)) is coupled to one or more main frame (not shown) with being directly or indirectly.Alternatively, network 400 also comprises multi-case data controller (not shown).

Network 400 comprises many ECMP paths.Such as, nodes X can arrive node L by the shortest path through node C or node D.The expense in path is equivalent.Nodes X can arrive node P by any one in node A, B, C or D.As the response receiving multi-case data grouping, which neighbor node the grouping of nodes X determination multi-case data should be forwarded to.Exemplarily, if multi-case data grouping is using node L as destination (such as, the bit string divided into groups at multi-case data in the position corresponding to BP{10000} is set), nodes X is determined to forward the packet to node C or node D.

Fig. 5 A is the example of the position index routing table (BIRT) 502 generated by BFR (node 406 such as, in Fig. 4).BIRT 502 comprises three row.First row comprises for determining the position, position of each BFER that BIER network (such as, the network 400 of Fig. 4) comprises and the information of (optionally) set identifier.In one embodiment, these row comprise BFR-ID simultaneously, or virtual bit position.Virtual bit position can be implemented as the unique integral in the border router of BIER network.Secondary series comprises each routable address in BFER, such as loopback or BFR prefix.3rd arranges the one or more entries comprised for these neighbours or multiple neighbours, and the BFER corresponding to this BFR prefix arrives along shortest path from this node by these neighbours.Exemplarily, the first row of BIRT502 shows exists two ECMP paths from the nodes X Fig. 4 to node L, and this node L has a position { 10000}.Article 1, path is by neighbours C, and Article 2 path is by neighbours D.Table 502 also show have a position BFER P of 00001}, this BFER P can from nodes X by neighbours A, B, C and D arrive.Therefore, BIRT 502 shows and there are four ECMP paths from nodes X to Node B.Discuss the process generating BIRT in Fig. 10 in more detail.

Fig. 5 B is that the example position index such as generated by the node 406 of Fig. 4 transmits (BIFT) 504.BIFT 504 utilizes the information in the BIRT 502 of Fig. 5 A to generate.As shown in the figure, BIFT 504 comprises the row for forwarding bitmask (FBM).BIFT 504 also comprises the row for neighbours.BIFT 504 show in multiple the positions arranged in FBM each arrive by corresponding neighbours.Exemplarily, considering the first entry in BIFT 504, is accessibility to position, position 1,2 and 3 corresponding BEFR by neighbours A.In the diagram, node P, M and N are accessibility by node A in reflection.Discuss the subsidiary details generating BIFT in fig. 11.

Fig. 6 is an example of the ECMP mapping table 602 generated, and exemplarily, is generated by the node 406 of Fig. 4.Consider the BIFT 504 of Fig. 5 B, position, position 1 (such as, for node A, Node B, node C and node D) in each entry is set.Consider that wherein nodes X (406 of Fig. 4) receives and has the bit string { example of the multi-case data grouping of 11111}.As the response receiving the grouping of this multi-case data, the FBM that this bit string is corresponding to neighbours A compares by nodes X.Because result is true ({ 11111}AND{00111} equals { 00111}), that this multi-case data divides into groups by nodes X, { duplicate of 00111} is forwarded to neighbours A to have bit string.Subsequently, the FBM of bit string and Node B compares by nodes X.Because result is true, another duplicate that this multi-case data divides into groups by nodes X is forwarded to Node B, and this duplicate has bit string, and { 01101}, namely { 11111} carries out the result operated with (AND) for this FBM and bit string.Also continue this process to node C and node D, nodes X forwards the duplicate of this multi-case data grouping to each of these nodes.BP is set and represents ECMP problem in multiple entries of BIFT.The duplicate of this multi-case data grouping is finally forwarded to such as node P from each of node A to D.These copies of multi-case data grouping are not required.

A kind of method of copy that do not need avoiding multi-case data to divide into groups is: shield the position in the bit string of the multi-case data grouping entered more afterwards at each FBM.In other words, before the bit string of the grouping entered and next FBM compare, those positions be set in the duplicate of the multi-case data grouping of going out can be cleared.Can be used to avoid the copy of unwanted grouping in this technology while, sacrifice load balance.In the examples described above, in the duplicate sending to the multi-case data of node A to divide into groups, set has been carried out to lower three positions.If nodes X reset those positions of the bit string entering grouping before being compared by the FBM of the bit string and Node B that enter grouping, the bit string of so going out can not indicate this grouping should be sent to node P.In other words, nodes X is by { entry of 11000} and Node B compares, and will have the bit string { duplicate of 01000} to the packet of Node B retransmitting multi-casting.{ the multi-case data grouping of 01000} will only be forwarded to node N to have bit string.Therefore, all groupings to node P will through node A.This causes load imbalance, and endangers network performance.

Avoid the other method reducing performance due to ECMP path to be generate multiple BIFT, wherein for each BIFT, each position is at one and be only set in a FBM at the most.In response to receiving multi-case data grouping at BFR place, this BFR selects one of BIFT and utilizes this multi-case data forwarding of packets by the BIFT that selects.In one embodiment, this selection is based on the flow valuve (flow value) of carrying in each multi-case data grouping.The BFR of the nodes X of such as Fig. 4 and so on generates BIFT as described in as shown in Fig. 7 A, 7B, 7C and 7D.In one embodiment, generate BIFT relate to generation and use ECMP mapping table (ECMP mapping table 602 as shown in Figure 6).

In one embodiment, generate ECMP mapping table 602 and relate to the maximum number determining ECMP path in BFR.To each position, BFR makes this position set to there being how many FBM in BIFT.In the example of the BIFT 504 of Fig. 5 B, the maximum number corresponding to position, position 1 is 4.In other words, position, position 1 is set to each that be arranged in the FBM of 4 shown in Fig. 5 B.This BFR generates the table of row or the entry comprised for each position.This is illustrated in the first row of ECMP mapping table 602.BFR also creates the row for every bar ECMP path.Because in BIFT, the maximum number in ECMP path is 4, nodes X generates the ECMP mapping table 602 with 4 additional column.BFR (nodes X in this example) then by distribution of information, the neighbours that this message identification position, corresponding position is set in row.Can such as by using circulation (round-robin) algorithm or other algorithms a certain to distribute this information.

To each position, this position is set at the most once in each given table.The neighbours that the multi-case data that this position is set divides into groups to be forwarded to are imported in the row corresponding with this position.Be set for 4 neighbours at BIFT 504 bit positions 1.Each neighbour comes across in the respective column of ECMP mapping table 602.Neighbours A comes across in first row, and neighbours B comes across in secondary series, and neighbours C comes across in the 3rd row, and neighbours D comes across in the 4th row.This is reflected in the first row of table 602.This means when to each column-generation BIFT, the BIFT corresponding to first row will have for neighbours A and only for the position, position 1 that neighbours A is set.If this BIFT is selected for retransmitting multi-casting packet, the grouping of this multi-case data will be forwarded to node P (corresponding to BP 1) by neighbours A by means of only neighbours A.The BIFT corresponding to secondary series will have for neighbours B and only for the position, position 1 that neighbours B is set.If this BIFT is selected for retransmitting multi-casting packet, the grouping of this multi-case data will be forwarded to node P (corresponding to BP 1) by Node B by means of only Node B

In placely transmit in 504, position, position 2 is respectively set once for neighbours A and neighbours D.This is reflected in the first two columns of the second row of ECMP mapping table 602.Owing to there are two the ECMP paths relevant to position, position 2, and in this mapping table, there are 4 row, each ECMP neighbours are repeated twice.Position 3, position is respectively set once for node A and Node B.Similar with 2, these are mapped in the third line of table 602 and are repeated.Position 4, position is only neighbours B and is set.Position 5, position is set for neighbours C and D.Wherein in the embodiment of not divided exactly by the maximum path number of this BIFT to the number in the ECMP path of position location, perform the asymmetric distribution of the neighbor information in ECMP mapping table.

Then mapping table 602 is transmitted by the positions that BFR is converted to 4 separation.As shown in Figure 7A, position is transmitted 702 and is corresponded to the tables 1 of mapping table 602 and arrange.As position is transmitted shown in 702, be set at the FBM meta 1,2 and 3 corresponding to neighbours A, be set at the FBM meta 4 corresponding to neighbours B, and be set at the FBM meta 5 corresponding to neighbours C.In BIFT 702, each in this position, five positions is only set once.Exemplarily, position, position 1 is only set in the FBM corresponding to A.Therefore, utilize that BIFT702 is forwarded and there is the multi-case data grouping that BP 1 is set be forwarded to node P merely through neighbours A.Fig. 7 B shows and arranges corresponding BIFT 704 to the table 2 of mapping table 602.Fig. 7 C shows and arranges corresponding BIFT 706 to the table 3 of mapping table 602.Fig. 7 D shows and arranges corresponding BIFT 708 to the table 4 of mapping table 602.

Fig. 8 is the example BIFT generated by BFR (nodes X (406) such as, in Fig. 4).Fig. 8 shows optional BIFT 802.Be not in BIFT by the bit string that enters compared with each FBM entry, optional forwarding mechanism is considered each (travel through this bit string, but not travel through this BIFT) in this bit string.An advantage of this mode the maximum number that the FBM in bit string and this BIFT compares is restricted to the length of this bit string, but not is restricted to the number of neighbours on the contrary, and this number may be very large.In the example present, only use one and transmit, and the respective entries for each ECMP path is created to each.For for each position, position 1 be set in 4 neighbours, in table, create 4 FBM entries.For the position, position 2 be set for 2 neighbours and BIFT 504, create 2 FBM entries.No matter use for the multilist mechanism described by Fig. 6,7A, 7B, 7C and 7D, or use for the many entry methods of single BIFT described in Fig. 8, all have employed the selection mechanism carrying out selecting between multiple table (or multiple entry).Exemplarily, flow valuve can be comprised in multi-case data grouping.For each packet of giving constant current, this flow valuve is consistent (such as, identical value).Based on the calculating to flow valuve, select BIFT (or entry).Which ensure that the multi-case data grouping of same stream is forwarded to identical neighbours and various flows can be distributed in ECMP path.Provide in Fig. 9 to 16B about generation forwarding information, select forwarding information, and the subsidiary details of retransmitting multi-casting packet.

Fig. 9 is the flow chart showing the instantiation procedure adopted by node according to this specification.Fig. 9 shows the method for generating forwarding information.In one embodiment, forwarding information is generated by the position forwarding router (BFR) of the BFR 406 and so on of such as Fig. 4.BFR generates BIER forwarding information, then determines whether there is ECMP path.If exist, BFR generates the forwarding information of the ECMP problem of copy of avoiding such as dividing into groups.

In 902, BFR generates BIRT.In one embodiment, this relate to such as use IGP receive the BIER information be advertised.Discuss the subsidiary details relevant to generating BIRT in Fig. 10.Utilize the BIRT generated in 902, BFR generates BIFT in 904.In one embodiment, generate BIFT and relate to and determine which BFER is by which neighbours Ke Da, and generate the forwarding bitmask corresponding to neighbours.Discuss the subsidiary details generating BIFT in fig. 11.

In 906, BFT determines the maximum number in ECMP path, as indicated in the BIFT that generates in 904.In one embodiment, this relates to and checks that BIFT is to determine the number of the FBM entry that this position is set wherein for each position, and determine to have be set to the maximum number of the FBM entry of position location.Discuss the subsidiary details about the maximum number determining ECMP path in fig. 12.In 908, BFR determines whether the maximum number in ECMP path is greater than 1.If not, about generating the forwarding information irrelevant with ECMP problem, do not need to take further action.But if BFR determines to there are many ECMP paths, BFR generates non-ECMP forwarding information in 910.In one embodiment, this relates to generating multiplely transmits and/or multiple forwarding entry, will to transmit and/or multiple forwarding entry is selected to provide load balancing and avoids the copy that divides into groups multiple.At Figure 13 A, in 13B, discuss the subsidiary details about generating non-ECMP forwarding information.

Figure 10 is the flow chart according to present specification shows the instantiation procedure adopted by node.Figure 10 shows the subsidiary details of the position index routing table (BIFT) generated described by 902 of Fig. 9.In one embodiment, Figure 10 is performed by the BFR of the BFR 406 and so on of such as Fig. 4.

In 1002, BFR receives the notice generated by BFER.In one embodiment, notice and received by IGP, and comprise the information identifying the routable address (such as, router identifier) be associated with BFER and the mapping between the position, position be associated with BFER and set identifier.In response to receiving notice, BFR determines the BP relevant to the BFER that generation is noticed in 1004.BFR determines set identifier simultaneously, if it is involved in notice.

In 1006, BFR accesses its topology information stored to determine the down hop neighbours of the shortest path along the BFER noticed towards generation.In 1008, BFR upgrades BIRT.In one embodiment, this comprises to add and comprises the entry of information, the position, position of this message identification BFER and BFR ID, and can arrive the neighbours of BFER via it.

Figure 11 is the flow chart according to present specification shows instantiation procedure that adopted by node, that generate BIFT.Figure 11 shows in Fig. 9 the subsidiary details operating 904.In one embodiment, Figure 11 is performed by the BFR of the BFR 406 of such as Fig. 4 and so on.

In 1102, BFR selects the first entry of the BIRT of BFR.In 1104, BFR determines the neighbours relevant to entry and position, position.In one embodiment, neighbours are identified by BFR ID or prefix or out of Memory.In 1106, BFR determines whether BIFT comprises the entry for neighbours.When a BIRT entry, by BIFT entry corresponding for the neighbours do not existed to identify this BIRT entry.If there is not the BIFT entry corresponding to neighbours, in 1108, BFT creates the BIFT entry corresponding to neighbours.In 1110, BFR arranges the position relevant to the position, position identified in BIRT entry in FBM.Determine whether BIRT comprises other entry at 1112, BFR.If so, next BIRT entry is selected at 1114, BFR.

Exemplarily, consider the BIRT 502 of Fig. 5 A, select the first entry in BIRT at 1102, BFR.In the example of BIRT 502, first entry corresponds to neighbours C.Then, determine that this entry corresponds to BP 5 at 1104, BFR.In other words, the node L being set to BP 5 can be arrived from BFR by node C.Then, BFT is neighbours C establishment entry in its BIFT, and carries out set to the 5th in FBM.Fig. 5 B shows the entry corresponding to neighbours C and has the position 5 be set.Then BFR repeats, and next selection is used for entry corresponding with position 5 of neighbours D etc.After performing the method shown in Figure 11, BFR will have the BIFT of an entry of the neighbours comprised for each BFR, and each entry will comprise forwarding bitmask, forwards bitmask and has the position be set for the accessibility each BFER of these neighbours.

Figure 12 is the flow chart according to present specification shows the instantiation procedure adopted by node.Figure 12 shows in Fig. 9 shown in 906, determines the subsidiary details of the method for the maximum number in the ECMP path presented in the forwarding information on BFR.In one embodiment, the method for Figure 12 is performed by the BFR of the BFR 406 and so on of such as Fig. 4.

In 1202, BFR initializing variable (such as, being called as maxECMP).In one embodiment, initialization maxECMP relates to and arranges maxECMP and equal 0.1204, this BFR selects first position in bit string.In one embodiment, as shown in Figure 5 B, bit string correspond to position transmit in forwarding bitmask.In another embodiment, position, position corresponds to first in BIER topology table, and topology table comprises each neighbours indicated by BFR, the accessibility information in which position.

In 1206, BFR initialization is such as called as the variable of COUNT.In one embodiment, initialization COUNT relates to and arranges COUNT and equal 0.The first entry in BIFT or topology table is selected at 1208, BFR.1210, exemplarily, BFR determines whether first of the FBM of entry be set to 1.If so, COUNT is increased at 1212, BFR.

Determine whether there is more entries in BIFT or topology table at 1216, BFR.If so, select next entry and return 1210 at 1214, BFR, wherein BFR determines whether first of FBM in entry be set to 1.If so, COUNT is increased at 1212, BFR.In this way, first position that BFR determines have how many entries to have in BIFT and is set, and value is stored in counting variable (such as, COUNT).

Next, determine whether the value of COUNT is greater than maxECMP at 1218, BFR.First time operation method time, in other words, to first position, if in these positions any one be set, will maxECMP be greater than because maxECMP is initially 0, COUNT.

1220, as determining that COUNT is greater than the response of maxECMP, maxECMP is set as equal with the value of COUNT by BFR.Determine in bit string, whether have position, more position at 1222, BFR.If so, the next bit position in bit string is selected at 1224, BFR.In the example present, next bit position is position 2.Then method returns 1206, wherein COUNT is initialized as 0 again.

Then, select the first entry of BIFT at 1208, BFR and determine whether the position 2 of first entry is set to 1.Method is carried out along each iteration of bit string.At the end of, the maximum number in ECMP path that the value of maxECMP will equal indicated by BIFT.Utilize the example BIFT504 of Fig. 5 B, corresponding to position, position 1, the maximum number in ECMP path is 4.

Figure 13 A is the flow chart according to present specification shows the instantiation procedure adopted by node.Figure 13 A shows the subsidiary details of the non-ECMP forwarding information of generation as shown in Fig. 9 910.In one embodiment, Figure 13 A is performed by the BFR of the BFR 406 and so on of such as Fig. 4.Figure 13 A shows the process BIFT comprising many ECMP paths being converted to the multiple BIFT not comprising many ECMP paths.

Mapping table is generated, such as, ECMP mapping table 602 in Fig. 6 at 1302, BFR.Mapping table comprises the row for every bar ECMP path.In one embodiment, the number of row equals the value of the maxECMP calculated in Figure 12.Mapping table also comprises the row of each position for bit string.

First position of the FBM be contained in the BIFT of BFR is selected at 1304, BFR.In 1306, BFR determines the occurrence number of this position in BIFT.In other words, BFR determines that how many neighbours' entry comprises wherein this position and is set to the FBM of 1.In one embodiment, it is 0 that BFR arranges count value, and when BFR detects that wherein this position is set to neighbours' entry of 1, increases count value.In one embodiment, the count value calculated in multiplexing Figure 12.

At 1308, BFR, corresponding neighbor information is pressed column distribution.If in the mapping table, the occurrence number of the position be set of BP is equal with the number arranged in mapping table, so will occur once with appearance (position in other words, be set) corresponding each neighbours.If the number of times that the position be set occurs is less than the maximum number arranged in mapping table, so corresponding to the position be set one or more neighbours will occur repeatedly.Exemplarily, if the true position location of BFR is set in two entries, and the maximum number in ECMP path (i.e. the columns of mapping table) is 4, and so corresponding to the entry that its meta is set each neighbours will occur twice in mapping table entry.

Determine in bit string at 1310, BFR or forward in bitmask whether there is more multidigit position.If so, next bit position is selected at 1312, BFR.After whole bit string is processed, mapping table is done and fills in.

1314, to each row of mapping table, BFR generates BIFT.In one embodiment, the establishment entry corresponding to each neighbour is related to each column-generation BIFT.Each entry comprises FBM.Each BP, BFR of mapping table are determined that this BP should be set in the FBM of which neighbour.Consider the example in Fig. 6, as shown in Figure 7 A, in a BIFT, position, position 1,2 and 3 is set in the FBM corresponding to neighbours A.To the BIFT shown in Fig. 7 B, position, position is set in the FBM corresponding from different neighbours.

Figure 13 B is the flow chart according to present specification shows the instantiation procedure adopted by node.Figure 13 B shows the another kind of method of the generation non-ECMP forwarding information as shown in Fig. 9 910.In one embodiment, method is performed by the BFR of the BFR 406 and so on of such as Fig. 4.

The optimization BIFT according to bit position index is generated at 1320, BFR.In one embodiment, this make use of BIFT, example BIFT 504 as shown in Figure 5 B.BFR determines the length of the FBM used in BIFT, and creates the entry for each BP in the BIFT optimized.First position is selected at 1322, BFR.The first neighbours in BIFT are selected at 1324, BFR.In the example of BIFT504, BFR selects neighbours A.Determine whether the position corresponding to this position is set at 1326, BFR.If so, the entry corresponding to this position is created at 1328, BFR.In one embodiment, entry comprises the forwarding bitmask coming from BIFT and the information identifying neighbours.In the example of BIFR 504, first is set, and therefore BFR creates the entry comprising the FBM{00111} of neighbours A and neighbours A in the BIFT optimized to BP 1.

Determine whether there are more neighbours in BIFT at 1330, BFR.If so, select next neighbour at 1332, BFR, and 1326, determine whether the position corresponding to position, position is set.If so, BFR creates another entry corresponding to position, position, and this entry comprises the FBM and neighbours that come from BIFT.In the example of BIFT 504, a BP is also set for neighbours B, and therefore BFR creates the entry of the information of the FBM{01101} comprising mark neighbours B and neighbours B in the BIFT optimized to BP 1.

After all neighbours in traversal BIFT, determine whether there is more multidigit position at 1334, BFR.If so, select next bit position at 1336, BFR, and return 1324.Behind each position in traversal bit string, the optimization BIFT according to position name placement of BFR BIFT 802 having generated such as Fig. 8 and so on, this optimization BIFT comprises the entry for each ECMP path.

Figure 14 is the flow chart according to present specification shows the instantiation procedure adopted by node.The BFR with non-ECMP forwarding information (transmitting such as, to each ECMP path) of the BFR 406 of such as Fig. 4 and so on selects one group of forwarding information (such as, BIFT or BIFT entry) to divide into groups in response to reception multi-case data.In one embodiment, select to relate to the flow valuve utilizing and be contained in multi-case data grouping.Figure 14 shows a kind of method generating flow valuve.

Information about firms is received at 1402, BFIR.In one embodiment, information about firms is bgp information.The multicast group identified by information about firms is identified at 1404, BFIR.In one embodiment, this relates to resolves member's message, identifies Multicast group address territory, and extracts Multicast group address.

Flow valuve is generated at 1406, BFIR.In one embodiment, generate flow valuve to relate to and create the cryptographic Hash of multicast source and/or the cryptographic Hash of group address or identifier.In an alternate embodiment of the invention, flow valuve can be implemented as digital value.Exemplarily, BFIR can join numerical value for each multicast component, and BFIR receives the member's message for each multicast group.BFIR notices by such as using the mode covering signaling (such as BGP) flow valuve being assigned to multicast group, and therefore each BFIR uses identical flow valuve.Optionally, to identical multicast group, each BFIR can use not homogeneous turbulence.In one embodiment, receive the response of multi-case data grouping as transmission node, transmission node calculates flow valuve.Transmission node can utilize multicast source and/or group address to calculate cryptographic Hash, or can add side information to cryptographic Hash, and this information is such as the information of the port identified on transmission node.

Group membership's table is upgraded at 1408, BFIR.In one embodiment, upgrade group membership's table to relate to BFIR in the entry corresponding with multicast members's information and store flow valuve.In one embodiment, upgrade group membership's table and also relate to BFIR in the bit string corresponding to multicast group by the position position corresponding with the BP of BFER.In one embodiment, member's message comprises the BP of BFER.Optionally, BFIR is performed searched to determine the BP relevant to the BFER initiating member's message by the address that uses the BFER initiating member message and the information (information in the BIRT of such as, BFIR) noticed by IGP.

Figure 15 is the flow chart according to present specification shows the instantiation procedure adopted by node.Figure 15 shows a kind of method performed by BFIR.Multi-case data grouping is received at 1502, BFIR (such as, from main frame).The flow valuve of the multicast group that multi-case data divides into groups is determined at 1504, BFIR.In one embodiment, multi-case data grouping comprises the information identifying multicast group.BFIR uses this information to search flow valuve in the GMT of BFIR.

At 1506, BFIR, flow valuve is packaged in multi-case data grouping, such as, is packaged in the head of BIER.In one embodiment, BFIR comprises flow valuve in the entropy field (entropy field) of packet header.Entropy field is generally the field of 8.It is advantageous that: entropy field is usually located at the head of grouping, therefore BFR can identify flow valuve rapidly, and without the need to resolving the main body of the multi-case data grouping entered.

The bit string of the multicast group of multi-case data grouping is determined at 1508, BFIR.In one embodiment, bit string is contained in the GMT of BFIR.At 1510, BFIR, bit string is encapsulated in multi-case data grouping, such as, is packaged in the head of BIER.BFIR can be configured to one in the some different method for packing of bit strings use by such as network manager.This depends on network configuration, and such as can make IP, MPLS or some other tunneling.

In the packet of 1512, BFIR retransmitting multi-casting.In one embodiment, this relates to and bit string and BFIR BFIT being compared.In one embodiment, retransmitting multi-casting packet relates to the BIFT accessing and safeguarded by IR, and determines multi-case data forwarding of packets to which neighbour based on bit string.In one embodiment, between the bit string that BFIR divides into groups at multi-case data and the entry of its BIFT, actuating logic and (AND) operates, and the result forwarding the packet to AND is genuine those neighbours.

Figure 16 A is according to this specification, illustrate node use the flow chart of the example process forwarded that multi-case data divided into groups.In one embodiment, this process is performed by BFR (such as, the BFR406 of Fig. 4).The process of Figure 16 A relates to use BIFT (such as, the BIFT504 of Fig. 5 B).

At 1602 places, BFR receives multi-case data grouping.BFR determines at 1604 places whether the grouping of this multi-case data is BIER grouping, thus whether comprises bit string.In one embodiment, BFR scans the head that multi-case data divides into groups, to obtain the value that this multi-case data of instruction is grouped into BIER grouping.BFR can detect that the transmitting terminal that multi-case data divides into groups is BFR, thus show that multi-case data is grouped into the conclusion of BIER multi-case data grouping.If multi-case data grouping is not the grouping of BIER multi-case data, then BFR performs alternate process at 1630 places.In one embodiment, alternate process 1630 relates to divides into groups inundation to the total interface on BFR by multi-case data, or abandons multi-case data grouping.Alternatively, if traditional multicast forwarding information can be used, then BFR can use this information to forward grouping.

At 1606 places, BFR determines whether multi-case data grouping comprises flow valuve.In one embodiment, this relates to BFR and resolves multi-case data packet header, and determines whether head comprises entropy field.If so, then the value in entropy field can be flow valuve, or can be one of class value that there is not flow valuve in the grouping of instruction multi-case data.

In response to detecting that multi-case data grouping comprises flow valuve, BFR determines the value of flow valuve at 1608 places.In one embodiment, flow valuve is that BFR changes to determine the encoded radio of actual flow valuve to it.Use flow valuve, BFR selects to be used for the forwarding information of retransmitting multi-casting packet at 1610 places.Such as, in one embodiment, BFR safeguards conversion table flow valuve and corresponding BIFT being carried out associating.Based on the flow valuve received in multi-case data grouping, BFR selects one of BIFT for retransmitting multi-casting packet.

BFR is in 1612 place's access bit strings.In one embodiment, access bit string relates to identification tunneling, and based on the position of tunneling type determination bit string in multi-case data grouping.At 1614 places, BFR is select items many in the BIFT of BFR.In one example, the first entry in BIFT is selected, and BFR travels through BIFT execution sequence.

BIFT determines whether multi-case data forwarding of packets at 1616 places to the neighbours be associated with selected BIFT entry.In one embodiment, this relates between bit string in multi-case data grouping and the FBM in selected BIFT entry and performs AND and operate.As determined at 1618 places, if the result of AND operation is true, then the method marches to 1620, and BFR upgrades the bit string in multi-case data grouping.In one embodiment, this relates between bit string in multi-case data grouping and the bit string in selected BIFT entry and performs AND and operate, and in the result write multi-case data that AND the is operated bit string of dividing into groups.This has following effect: can not reset in position, position via the divide into groups neighbours that are forwarded to and the position arrived of multi-case data.Prevent the problem of repetition or loop like this.

At 1622 places, multi-case data forwarding of packets is given the neighbours corresponding with BIFT entry by BFR.At 1624 places, BFR determines whether there are other entries in BIFT.If so, the method is back to 1614, and next entry in BIFT is selected.

Figure 16 B is according to this specification, show node use the flow chart of the example process forwarded that multi-case data divided into groups.In one embodiment, this process is performed by BFR (such as, the BFR 406 of Fig. 4).The process of Figure 16 B relates to the BIFT (such as, the BIFT 802 of the optimization of Fig. 8) using and optimize.

At 1650 places, BFR receives multi-case data grouping.BFR determines at 1652 places whether the grouping of this multi-case data is BIER grouping, thus whether comprises bit string.In one embodiment, BFR scans the head that multi-case data divides into groups, to obtain the value that this multi-case data of instruction is grouped into BIER grouping.BFR can detect that the transmitting terminal that multi-case data divides into groups is BFR, thus show that multi-case data is grouped into the conclusion of BIER multi-case data grouping.If multi-case data grouping is not the grouping of BIER multi-case data, then BFR performs alternate process at 1680 places.In one embodiment, alternate process 1680 relates to divides into groups inundation to the total interface on BFR by multi-case data, or abandons multi-case data grouping.Alternatively, if traditional multicast forwarding information can be used, then BFR can use this information to forward grouping.

If multi-case data is grouped into the grouping of BIER multi-case data, then BFR knows that the grouping of this multi-case data comprises bit string.BFR determines the position of this bit string in multi-case data grouping at 1654 places.Use bit string, BFR determines which neighbour multi-case data grouping should be forwarded to.In one embodiment, this relates to first (as Suo Shi 1656) selecting in bit string, and determines whether first of bit string be set (as Suo Shi 1658).If this position is not set, then determine whether there is more multidigit in bit string at 1676 place BFR.If so, then BFR selects next bit at 1678 places, and the method is back to 1658.

Be set in response to the position determined in bit string, BFR determines at 1660 places whether multi-case data grouping comprises flow valuve.In one embodiment, this relates to BFR and resolves multi-case data packet header, and determines whether head comprises entropy field.If so, then the value in entropy field can be flow valuve, or can be one of class value that there is not flow valuve in the grouping of instruction multi-case data.

In response to detecting that multi-case data grouping comprises flow valuve, BFR determines the value of flow valuve at 1662 places.In one embodiment, flow valuve is that BFR changes to determine the encoded radio of actual flow valuve to it.At 1664 places, BFR determines the number of entry corresponding with selected position, position in the BIFT optimized.Use flow valuve, BFR selects to be used for the forwarding entry of retransmitting multi-casting packet at 1664 places.In one embodiment, the number using flow valuve to select forwarding information to relate to determine available forwarding information collection (such as, BIFT entry) and/or the number in ECMP path be associated with given position, position.Based on the number in ECMP path and the maximum number in ECMP path, flow valuve can be modified (such as, divided by suitable value).Such as, if flow valuve is provided for the selector in four ECMP paths, but given position, position only has two entries, then flow valuve can be provided may selecting of the number that matches by displacement to the right (divided by two).

At 1670 places, BFR creates the duplicate of multi-case data grouping, and upgrades bit string.The bit string upgraded in the duplicate of grouping relates to position corresponding with not accessibility neighbours in bit strings, and wherein, these neighbours can not arrive via the shortest path of the interface that be forwarded to from the duplicate divided into groups.This can operate by performing AND between the bitmask that the position, position from the multi-case data grouping entered is corresponding with transmitting position in entry and selected.End value is used as the bit string of the duplicate of multi-case data grouping.At 1672 places, BFR by multicast forwarding of packets to interface.

At 1674 places, BFR upgrades during multi-case data divides into groups by those clearings that the position that is set in dividing into groups with the multi-case data that BFR forwards in the bit string of being divided into groups by multi-case data is corresponding the bit string arrived.In one embodiment, perform AND between this anti-tree relating to the bit string bitmask corresponding with in entry and selected position in the multi-case data grouping received to operate.This has following effect: reset those positions corresponding with the position, position be set in the bit string spreading out of grouping, and this prevents circulation and repeats.Then BFR determines whether there is more multidigit in bit string at 1676 places.Then BFR advances to the bit string of received multi-case data grouping by turn, until arrive the end of bitmask.

Figure 17 illustrates some add-on assemble of node in the network that can be used in such as shown in Fig. 3 and/or the block diagram of replacement assemblies.In this description, node 1700 comprises multiple line card (line card 1702 (1)-(N)), and these line cards are communicatively coupled to forwarding engine or forwarding of packets device 1710 and processor 1720 via data/address bus 1730 and result bus 1740.Line card 1702 (1)-(N) comprises multiple porthandler 1750 (1,1)-(N, N), and these porthandlers are controlled by port processor controllers 1760 (1)-(N).Shall also be noted that forwarding engine 1710 and processor 1720 not only intercouple via data/address bus 1730 and result bus 1740, but also be coupled with communicating with one another by communication link 1770.

The processor 1750 and 1760 of each line card 1702 can be installed on single printed circuit board.When receive grouping or grouping and head time, can by router one 700 in the following manner to grouping or divide into groups and head identify and analyze.When received, grouping (or its some control informations or its whole control information) or grouping and head are by from porthandler 1750 (1,1)-(N, one of N) (grouping or grouping and head are received at this porthandler place) is sent to and is coupled to one or more equipment in those equipment of data/address bus 1730 (such as, porthandler 1750 (1,1) other porthandlers in-(N, N), forwarding engine 1710 and/or processor 1720).Can be determined grouping or grouping and the process of head by such as forwarding engine 1710.Such as, forwarding engine 1710 can be determined to divide into groups or that grouping and head should be forwarded in porthandler 1750 (1,1)-(N, N) is one or more.This can pass through to indicate following content to corresponding one (or multiple) in port processor controllers 1760 (1)-(N): be kept at porthandler 1750 (1,1)-(N, N) grouping in (or multiple) given in or grouping and head should be forwarded to porthandler 1750 (1,1) one suitable in-(N, N).In addition, or alternatively, once grouping or grouping and head have been identified as process, forwarding engine 1710, processor 1720 etc. can be used to process in a certain manner this grouping or grouping and head, or add grouping security information to guarantee grouping safety.On the node of initiating such grouping or grouping and head, this process can comprise such as: be encrypted some or all in the information of grouping or the information of grouping and head, adds the digital signature or some other information or process that can ensure grouping or grouping and head safety.On the node receiving grouping or grouping and the head processed like this, perform corresponding process and recover or verify the information of protected grouping or the information of grouping and head.

Figure 18 is the block diagram of computing equipment, shows forwarding module and how to be implemented in software, as mentioned above.Computing system 1810 represents widely can any single processor of moving calculation machine instructions or multiple processor computing equipment or system.The example of computing system 1810 comprises but not is limited to various equipment and (comprises work station, personal computer, laptop computer, client-side terminal, server, distributed computing system, handheld device (such as, personal digital assistant and mobile phone), network appliance, switch, router, storage control (such as, array control unit, magnetic tape drive controller or hard disk drive controllers) etc.) in any one or more.In the configuration that it is the most basic, computing system 1810 can comprise at least one processor 1818 and system storage 1816.By running the software realizing forwarding module 1817, computing system 1810 becomes dedicated computing equipment, and this dedicated computing equipment is configured to perform forwarding of packets in the above described manner.

Processor 1818 generally represents can deal with data or the explanation also any type of operating instruction or the processing unit of form.In certain embodiments, processor 1818 can receive instruction from software application or module.These instructions can make processor 1818 perform the function of the one or more embodiments in embodiment that is described herein and/or that illustrate.Such as, processor 1818 can perform operation described herein and/or can be the device for performing operation described herein.Processor 1818 can also perform other operations any that are described herein and/or that illustrate, method or process and/or can be for the device performing other operations any that are described herein and/or that illustrate, method or process.

System storage 1816 generally represents can store any type of data and/or other computer-readable instructions or the volatibility of form or non-volatile memory device or medium.The example of system storage 1816 comprises but not is limited to random access memory (RAM), read-only memory (ROM), flash memory or other suitable memory devices arbitrarily.Although not requirement, but in certain embodiments, computing system 1810 can comprise both volatile memory-elements (such as, system storage 1816) and non-volatile memory device (main storage 1832 such as, will described in detail) below.In one example, the program command that can perform to realize forwarding module can be loaded in system storage 1816, and wherein, forwarding module is configured to retransmitting multi-casting packet.

In certain embodiments, except processor 1818 and system storage 1816, computing system 1810 can also comprise one or more assembly or element.Such as, as shown in figure 18, computing system 1810 can comprise Memory Controller 1818, I/O (I/O) controller 1820 and communication interface 1822, and each in these can interconnect via the communications infrastructure 1812.The communications infrastructure 1812 generally represents any type of communication between one or more assemblies that can promote computing equipment or the infrastructure of form.The example of the communications infrastructure 1812 comprises but not is limited to communication bus (such as, Industry Standard Architecture (ISA), peripheral component interconnect (PCI), PCIexpress (PCIe) or similar bus) and network.

Memory Controller 1818 generally represents any type of communication between one or more assemblies that can process memory or data or controlling calculation system 1810 or the equipment of form.Such as, in certain embodiments, Memory Controller 1818 can carry out between control processor 1818, system storage 1816 and I/O controller 1820 via the communications infrastructure 1812 communication.In certain embodiments, Memory Controller 1818 can separately or carry out combining that to perform feature that is described herein and/or that illustrate or the one or more and/or Memory Controller 1818 in operating can be for separately or carry out combining with other elements and perform feature that is described herein and/or that illustrate or the one or more device in operating with other elements.

I/O controller 1820 generally represents can be coordinated and/or any type of input and output function of controlling calculation equipment or the module of form.Such as, in certain embodiments, I/O controller 1820 can control or promote to carry out data transmission between one or more elements of computing system 1810 (such as, processor 1818, system storage 1816, communication interface 1822, display adapter 1826, input interface 1830 and memory interface 1834).

Communication interface 1822 represents widely can promote any type of the communication between computing system 1810 and one or more optional equipment or the communication equipment of form or adapter.Such as, in certain embodiments, communication interface 1822 can promote computing system 1810 and comprise the communication between the private network of additional computing systems or public network.The example of communication interface 1822 comprises but not is limited to wired network interface (network interface unit), radio network interface (wireless network interface card), modulator-demodulator and other suitable interfaces arbitrarily.In at least one embodiment, communication interface 1822 can provide the direct connection of going to remote server via the direct link going to network (such as, the Internet).Communication interface 1822 can also pass through such as local area network (LAN) (such as, ethernet network), individual territory net, phone or cable system, cell phone connects, satellite data connects or other suitable connections provide such connection indirectly arbitrarily.

In certain embodiments, communication interface 1822 can also represent host adapter, and this host adapter is configured to via external bus or communication port to promote the communication between computing system 1810 and one or more complementary network or memory device.The example of host adapter comprises but not is limited to small computer system interface (SCSI) host adapter, USB (USB) host adapter, IEEE (IEEE) 11054 host adapter, Serial Advanced Technology Attachment (SATA) and SATA (eSATA) host adapter expanded, Advanced Technology Attachment (ATA) and Parallel ATA (PATA) host adapter, Fibre Channel port adapters, Ethernet Adaptation Unit, etc.

Communication interface 1822 can also allow computing system 1810 to participate in Distributed Calculation or telecomputing.Such as, communication interface 1822 can receive instruction from remote equipment or send to remote equipment for operation instruction.

As shown in figure 18, computing system 1810 can also comprise at least one display device 1824, and this display device 1824 is coupled in the communications infrastructure 1812 via display adapter 1826.Display device 1824 generally represents visually can show any type of the information forwarded by display adapter 1826 or the equipment of form.Similarly, display adapter 1826 generally represents and is configured to forward from the communications infrastructure 1812 (or from frame buffer) be used for carrying out any type of figure, text and other data that shows or the equipment of form at display device 1824.

As shown in figure 18, computing system 1810 can also comprise at least one input equipment 1828, and this input equipment 1828 is coupled in the communications infrastructure 1812 via input interface 1830.Input equipment 1828 generally expression can provide (computer or the manually generation) any type of input or the input equipment of form to computing system 1810.The example of input equipment 1828 comprises but not is limited to keyboard, sensing equipment, speech recognition apparatus or other input equipments arbitrarily.

As shown in figure 18, computing system 1810 can also comprise main storage 1832 and spare storage device 1833,1832 and 1833 is coupled in the communications infrastructure 1812 via memory interface 1834.Memory device 1832 and 1833 generally represents can store any type of data and/or other computer-readable instructions or the memory device of form or medium.Such as, memory device 1832 and 1833 can be disc driver (such as, so-called hard disk driver device), floppy disk, tape drive, CD drive, flash drive etc.Memory interface 1834 generally represents and is used for transmitting any type of data or the interface of form or equipment between memory device 1832 and 1833 and other assemblies of computing system 1810.The information that the memory device of class main storage 1832 can store such as routing table and transmit and so on.

In certain embodiments, memory device 1832 and 1833 can be configured to carry out reading and/or writing to removable memory module from removable memory module, wherein, removable memory module is configured to store computer software, data or other computer-readable information.The example of suitable removable memory module comprises but not is limited to floppy disk, tape, CD, flash memory device etc.Memory device 1832 and 1833 can also comprise and allows computer software, data or other computer-readable instructions to be loaded into other similar structures in computing system 1810 or equipment.Such as, memory device 1832 and 1833 can be configured to read or write software, data or other computer-readable information.Memory device 1832 and 1833 can also be a part for computing system 1810, or can be the autonomous device conducted interviews by other interface systems.

Other equipment many or subsystem can be connected to computing system 1810.On the contrary, do not need to occur that all component shown in Figure 18 and equipment are to implement embodiment that is described herein and/or that illustrate.Above mentioned equipment and subsystem can also interconnect to be different from the mode shown in Figure 18.

Computing system 1810 can also adopt the software of arbitrary number, firmware and/or hardware configuration.Such as, the one or more computer programs (being also referred to as computer software, software application, computer-readable instruction or computer control logic) that can be encoded as on computer-readable recording medium in embodiment disclosed herein.The example of computer-readable recording medium comprises magnetic storage medium (such as, hard disk drive and floppy disk), optical storage media (such as, CD-ROM or DVD-ROM), electric storage medium (such as, solid-state drive and quick flashing medium) etc.Such computer program via network (such as, the Internet) or can also be transferred to computing system 1810 to store in memory in carrier media.

The computer-readable medium comprising computer program can be loaded in computing system 1810.Then all parts or the part that are stored in the computer program on computer-readable medium can be stored in each several part of system storage 1816 and/or memory device 1832 and 1833.When being performed by processor 1818, the computer program be loaded in computing system 1810 can make processor 1818 perform the one or more function in described herein and/or shown embodiment and/or make processor 1818 become device for performing the one or more function in described herein and/or shown embodiment.Additionally or alternatively, one or more in described herein and/or shown embodiment can be implemented in firmware and/or hardware.Such as, computing system 1810 can be configured to application-specific integrated circuit (ASIC) (ASIC), and it is one or more that this ASIC is suitable for realizing in embodiment disclosed herein.

Figure 19 is the block diagram of the exemplary network device that can be associated with the node in the network 200 of Fig. 2.For example, the network equipment 1950 of Figure 19 can be associated with the BIER enable node 206 in Fig. 2.In some cases, " node " used herein comprises the one or more network equipments be associated with this node." network equipment " used herein comprises and performs routing function and/or forwarding capability and the various equipment (such as, router, switch or network controller) supporting one or more route and/or exchange agreement.The one or more routing table of network equipment maintenance and/or transmit, these routing tables and/or transmit route and/or the forwarding information that storaging mark goes to each data source and/or data user.Such as, in multicast enable node, the network equipment realizes the multicast route protocol being used for multi-case data grouping being conveyed to multicast reception end from multicast source.

In the embodiment of Figure 19, the network equipment 1950 comprises the memory device for information about firms 1952, the memory device for forwarding information 1964, forwarding module 1960 and interface 1962.Interface 1962 is coupled as and sends and receive grouping and/or other internet messages.It should be noted that the network equipment 1950 can comprise additional interface, and each interface can be logic interfacing or physical interface.In one embodiment, interface 1962 comprises one or more port.

Forwarding module 1960 is configured to perform forwarding based on stored forwarding information 1964.Forwarding module 1960 is also configured to upgrade the information about firms 1952 that stores and forwarding information 1964.Forwarding module 1960 can realize the one or more examples in 3 layer protocols and/or 2 layer protocols.

Entry 1970 provides the example of the information about firms be stored in the memory of the network equipment.As shown in the figure, entry 1970 comprises set identifier (SI) 1954, identifies the information 1958 of the information 1956 of a position (BP) and mark multicast group.Node associated by SI and BP identification entry 1970, and the multicast group of multicast group information mark ordered by respective nodes.In one embodiment, the memory device of information about firms 1952 is implemented as group member table.

Entry 1972 provides the example of the forwarding information that can be stored in the memory of the network equipment.As shown in the figure, entry 1972 comprises the information 1969 of information 1966, bit string or the bit array 1968 and mark neighbours identifying BP.Forwarding module 1960 uses the information in entry 1972 by multi-case data forwarding of packets to the interface be associated with the neighbours identified in this entry.In one embodiment, the memory device of forwarding information 1964 is implemented as position index and transmits (BIFT).

Although in conjunction with some embodiments, invention has been described, and the present invention is not intended to be limited to proposed concrete form.On the contrary, it is intended to cover and is rationally included in substituting in the scope of the present invention of claims definition, amendment and equivalent.

Claims (20)

1. a method, comprising:

In reception at Node grouping, wherein said grouping comprises bit string;

From many group forwarding informations, select first group of forwarding information, wherein this selection is based on flow valuve, and described first group of forwarding information comprises forwarding bitmask; And

Described grouping is forwarded based on described bit string and described first group of forwarding information.

2. the method for claim 1, also comprises:

Determine the maximum number in the equal cost multipath ECMP path identified in single group forwarding information.

3. method as claimed in claim 2, also comprises:

Generate and described organize forwarding information more, wherein this generation comprises described single group forwarding information to be converted to and described organizes forwarding information more.

4. the method for claim 1, wherein said grouping comprises described flow valuve.

5. the method for claim 1, also comprises:

Described flow valuve is modified.

6. the method for claim 1, also comprises:

Generate described flow valuve, wherein this generation comprises the cryptographic Hash calculating the address be associated with multicast group.

7. the method for claim 1, each group in wherein said many group forwarding informations comprises to be transmitted.

8. the method for claim 1, each group in wherein said many group forwarding informations comprises transmits entry.

9. a network equipment, comprising:

Memory, this memory stores organizes forwarding information more;

Network interface, this network interface is configured to receive grouping, and wherein wherein said grouping comprises bit string; And

Processor, this processor is configured to:

From described many group forwarding informations, select first group of forwarding information, wherein select first group of forwarding information to carry out based on flow valuve, and described first group of forwarding information comprises forwarding bitmask; And

Described grouping is forwarded based on described bit string and described first group of forwarding information.

10. the network equipment as claimed in claim 9, wherein said processor is also configured to:

Determine the maximum number in the equal cost multipath ECMP path identified in single group forwarding information.

11. network equipments as claimed in claim 10, wherein said processor is also configured to:

Generate and described organize forwarding information more, wherein generate described many group forwarding informations and comprise described single group forwarding information to be converted to and described organize forwarding information more.

12. network equipments as claimed in claim 9, wherein said grouping comprises described flow valuve.

13. network equipments as claimed in claim 10, wherein said processor is also configured to:

Described flow valuve is modified.

14. network equipments as claimed in claim 10, wherein said processor is also configured to:

Generate described flow valuve, wherein generate described flow valuve and comprise the cryptographic Hash calculating the address be associated with multicast group.

15. network equipments as claimed in claim 10, each group in wherein said many group forwarding informations comprises to be transmitted.

16. 1 kinds of systems, comprising:

Storage device, this storage device organizes forwarding information for storing more;

Network Interface Unit, this Network Interface Unit is for receiving grouping, and wherein said grouping comprises bit string; And

Processing unit, this processing unit is used for:

From described many group forwarding informations, select first group of forwarding information, wherein select first group of forwarding information to carry out based on flow valuve, and described first group of forwarding information comprises forwarding bitmask; And

Described grouping is forwarded based on described bit string and described first group of forwarding information.

17. systems as claimed in claim 16, wherein said processing unit is also configured to:

Determine the maximum number in the equal cost multipath ECMP path identified in single group forwarding information.

18. systems as claimed in claim 17, wherein said processing unit is also configured to:

Generate and described organize forwarding information more, wherein generate described many group forwarding informations and comprise described single group forwarding information to be converted to and described organize forwarding information more.

19. systems as claimed in claim 16, wherein said grouping comprises described flow valuve.

20. systems as claimed in claim 16, wherein said processing unit is also configured to:

Generate described flow valuve, wherein generate described flow valuve and comprise the cryptographic Hash calculating the address be associated with multicast group.