The double-stack support extension method of layer mobile IPv 6 protocol
Technical field
The invention belongs to dual-stack network mobile communication technology field, relate in particular to the double-stack support extension method of a kind of layer mobile IPv 6 protocol (HMIPv6-Hierarchical Mobile IPv6).
Background technology
Mobile network protocol (MIP-Mobile Internet Protocol) is a kind of scheme that locomotive function is provided on Global Internet at present, it have extensibility better, reliability is strong and fail safe than characteristics such as height, and can guarantee that mobile node (MN-Mobile Node) still can keep current ongoing communication when changing Network Access Point.MIP has two kinds of MIPv4 and MIPv6, and wherein the MIPv6 agreement is the mobility characteristics of introducing on the basis of IPv6, so that serve for the mobile subscriber better.Because IPv6 is better than IPv4 at aspects such as address space, fail safes, so MIPv6 is more more superior than MIPv4 performance.
Stipulate in the MIPv6 agreement, when MN when the local moves to field network, will obtain a Care-of Address (CoA-Care-of Address).In order to keep its communication continuity, MN need register the position at its current place to its home agent (HA-Home Agent), the HA grouping that just can transmit any MN of mailing to or be sent by MN like this.Clearly, under the pass-through mode of this triangle route, increase the burden of HA, made HA become the bottleneck of communication possibly, thereby influenced the performance of whole link.In order to reduce the possibility that this hidden danger takes place, realize routing optimality, MIPv6 agreement regulation, MN also needs to register to corresponding communication opposite end (CN-Corresponding Node), like this MN just can be directly and CN communicate, no longer need the forwarding of HA.For the fail safe that guarantees to communicate by letter each other, agreement is defined in must at first be returned route before MN and the CN binding and can reach process.Whether the home address (HoA-Home Address) that its objective is checking MN can reach for HA with CoA, and then negotiates the communication key of finishing communication.
The MIPv6 agreement allows node when mobile, keep its accessibility, and safeguard ongoing communication continuity between MN and the CN in the Internet network topology.For this reason, MN needs to register to its HA and all CN transmission Binding Updates (BU-Binding Update) after each moving.And,, also approximately need the two-way time between 1.5 times of MN and the CN even under the best circumstances for the authentication of the Binding Update of CN.In addition, upgrade HA and CN and all need 1 times two-way time.The repeated use of local key can not reduce to be upgraded required two-way time of CN.When node switched, the time delay that these a series of renewal operations are produced will make current ongoing communication disruption, thereby influences communication quality.If between transfer period, eliminate or reduce these additional handover delays, then can significantly improve the performance of MIPv6, strengthen support to the delay sensitive business.
The HMIPv6 agreement that the Internet engineering duty group (IETF-Internet Engineering Task Force) proposes by adopting level type routing infrastructure, has reduced the Signalling exchange quantity of MN and HA and CN, switches the communication interruption time that causes thereby reduce.In HMIPv6, introduced a new functional entity, be called mobile anchor point (MAP-Mobility Anchor Point).A MAP zone comprises a plurality of subnets, and each subnet all has couple in router (AR-Access Router), and the number of subnet according to circumstances can change.MN disposes specific region Care-of Address (RCoA-Regional CoA) by the advertising of route (RA-Router Advertisement) that reception comprises MAP information, uses this address to register to HA and CN then.In addition, MN also can obtain an on-link care-of address (LCoA-On-LinkCoA) simultaneously.When MN was mobile in same MAP territory, RCoA did not change, and at this moment binds again to HA and CN with regard to not needing, and had so just reduced between the subnet of MN in the MAP territory when mobile and HA and the mutual frequency of CN.
From top narration as can be seen no matter be that MIPv6 agreement or HMIPv6 agreement all are only to be applicable to the IPv6 network.But current actual conditions are, the IPv6 network does not have widespread deployment, and the Internet overwhelming majority still is the network of IPv4, or perhaps the network of IPv4 and IPv6 coexistence.Therefore, MN also can not keep its existing communication by the IPv6 address, and this just makes the application of HMIPv6 agreement have certain restriction.So, be necessary the HMIPv6 agreement is expanded improvement, make it be adapted to IPv4 network and the network coexisted two stack environment of IPv6.
The MAP of known technology one: HMIPv6 can be arranged in the random layer of hierarchical network, and unlike the Foreign Agent in the mobile IPv 4, MAP has not also required an independently subnet.MAP can reduce the information interaction number in the MIPv6 subrange.The more following problems that are introduced as of MAP provide solution:
MN only need send local local Binding Update (LBU-Local Binding Update) to MAP when mobile in same MAP territory, and does not need to send to HA and CN.
After MN moved to new MAP territory, MN only need send a BU to HA and CN, and the number of the CN that communicating by letter of this process and MN has nothing to do.
MAP comes down to the HA of a part.Introduce the hierarchical mobility management model in MIPv6, purpose is in order to improve the performance of MIPv6, and the influence to MIPv6 or other IPv6 agreements simultaneously minimizes.
When MN moves to a level type field network, will receive the RA that comprises the MAP option.What fill in this option is the relevant information of MAP in this territory, and the address of distance, length and MAP etc. is wherein arranged, and utilizes this option MN can judge that also it is whether still in original MAP territory.If MN when mobile, will receive the information about same MAP among the RA in same MAP territory.In case receive the RA that the MAP address changes, MN must change accordingly to HA and all CN transmission BU.
If MN is a node of not supporting HMIPv6, then will no longer carry out the process of HMIPv6 agreement, this moment, MN will be still carries out mobile management like that according to what define among the MIPv6.If MN is a node of supporting HMIPv6, then should select to use the HMIPv6 agreement to carry out mobile management.
The operation of MN:
When MN moves to a new MAP territory, just need two Care-of Address of configuration: one is the RCoA that generates according to the MAP address prefix, and another is LCoA.RCoA generates with the stateless address auto configuration mode, and LCoA both can generate by the stateless address auto configuration mode, also can pass through DHCP (DHCP-Dynamic Host Configuration Protocol) and distribute.
After generating RCoA, MN need send LBU to MAP, and wherein showing must set with the A flag bit that needs binding acknowledgement (BA-Binding Acknowledgement) to the M flag bit of MAP registration.That the home address option among the LBU (HAO-Home Address Option) comprises is the RCoA of MN.Do not include alternative Care-of Address option in this information.The source address of LBU is LCoA.MAP will carry out duplicate address detection (DAD-Duplicate AddressDetection) for the RCoA of MN, and return a local binding acknowledgement message (LBA-Local BindingAcknowledgement) to MN.MN must ignore the RCoA that comprises MN but the not BA of containing type 2 route heads grouping.
After MN successfully registers, will between MN and MAP, set up a bidirectional tunnel.All groupings that MN sends will be sent to the MAP place by the tunnel.Source address in the grouping external head is the LCoA of MN, and destination address is the address of MAP.In the internal head, source address is the RCoA of MN, and destination address is the address of CN.Similarly, all destination address is MN was was all intercepted and captured and was forwarded to by the tunnel in the grouping of the RCoA of MN by MAP LCoA.
If when MN receives the RA that comprises a plurality of MAP option informations, can dispose a plurality of RCoA.In this case, MN must carry out the Binding Update program with each RCoA.
After the MAP registration, MN must be to its new RCoA of HA registration.Wherein, among the HAO HoA of MN, CoA is RCoA.Simultaneously, MN also needs to send BU to its CN.
The LBA that MN should receive MAP by the time registers to HA afterwards again.To life cycle of HA and CN binding can not greater than with life cycle of MAP binding, and obtain the life cycle of this binding in LBA.
In order to be increased in the speed of switching between the MAP, and reduce losing of grouping, MN should send LBU to its original MAP, registers its new LCoA.Like this, arrive the grouping of original MAP, also can be forwarded to new LCoA place immediately.
It should be noted that if CN and MN on same link the time, the address of MN execution this moment BU can be its LCoA, like this, just can direct communication between MN and the CN, no longer need the forwarding of MAP.
The operation of MAP:
MAP plays the part of the role of HA: it intercepts and captures the grouping that all destination addresses are the RCoA of MN, and is forwarded to the current LCoA place of MN by the tunnel.
MAP is ignorant for the HoA of MN.MN sends a LBU who contains M, A position to MAP.The purpose of this binding is to show that to MAP MN has formed RCoA.If bind successfully, then MAP must return the success that a LBA shows this registration.This is identical with the operation of HA.In HMIPv6, do not introduce new error coded.The route head that must comprise a type 2 among the BA comprises the RCoA of MN therein.
MAP must be able to intercept and capture and transmit grouping from MN by the tunnel.Wherein MN is the entrance in tunnel, and MAP is the exit point in tunnel.
MAP is as the HA of LCoA, and destination address is that all groupings of RCoA use proxy neighbor advertisements to be intercepted and captured by MAP, and encapsulation is routed to the LCoA place of MN then.
As for the operation of HA and CN, HMIPv6 does not do any modification.
The defective of known technology one is only to be applicable to the IPv6 network, does not support the IPv4 node.
Known technology two: promptly the MIPv6 agreement is to the support of dual stack node.
The MIPv6 agreement is supported IPv6 node certain operations, for the node of not supporting IPv6, also needs to revise this agreement, thereby satisfies the requirement of dual stack nodes such as IPv4 and IPv6.In order to make dual stack node can use the MIPv6 agreement, MN should use the HoA of an IPv4 and an IPv6 simultaneously, and its HA is registered this two addresses respectively; And MN will be appreciated that IPv4 and the IPv6 address of HA, finds but do not need to carry out prefix for the IPv4 address.In order to make dual stack node can use the MIPv6 agreement to carry out mobile management, need to revise its BU and corresponding BA information to the local.Specify as follows:
The MN of two stacks when moving to field network, need register to its HA, binds the CoA at its current place.In order to make MN keep communication in a field network of not supporting IPv6, MN needs the HoA of an IPv4, and for all setting up corresponding banding cache item in each address.At this moment the form of IP packet that carries BU, BA is with different.This depends on whether field network supports IPv6.At this moment three kinds of different situations will be arranged:
Field network is supported IPv6, and disposes the CoA of an IPv6 to MN;
MN has only disposed a globally unique IPv4 address at field network;
MN has disposed a privately owned IPv4 address at field network.
Respectively these three kinds of situations are analyzed below:
Field network is supported IPv6:
At this moment, MN can dispose a globally unique IPv6 address.MN will send BU to the IPv6 address of HA.The HAO that can include IPv4 among this BU.After receiving BU, HA will create two banding cache items.A HoA who is used for depositing the IPv4 of MN, another is used for depositing the HoA of the IPv6 of MN.These two cache entry are all pointed to the CoA of the IPv6 of MN.Therefore destination address is the CoA place that the grouping of the HoA of the IPv4 of MN or IPv6 can be sent to the IPv6 of MN by the tunnel.In order to raise the efficiency, MN can set up two different tunnels, and one is used for the IPv4 business, and another is used for the IPv6 business.
In this case, the HAO that only need increase an IPv4 in BU gets final product.
After receiving BU and finishing corresponding banding cache item, HA must send BA to MN.In addition, if include the HAO of IPv4 among the BU, then also must comprise this among the BA.Show acceptance to the HoA of the IPv4 of MN.
When MN obtains the CoA of IPv4 and IPv6 simultaneously in field network after, preferentially use the CoA of IPv6 to carry out the MIPv6 binding registration.
Only support IPv4 for field network, following two kinds of situations will be arranged:
Field network is only supported the IPv4 public address:
At this moment the MN IPv6 grouping that needs to comprise BU is sent to the place, IPv4 address of HA by the tunnel.Wherein the source address of external head is the CoA of MN from the IPv4 of field network acquisition, and HAO comprises the HoA of the IPv6 of MN among the BU.Yet because CoA is the IPv4 address of MN, MN must be included in the CoA of IPv4 in the IPv6 grouping.
After receiving BU, the HoA that HA is necessary for the IPv6 of MN creates the banding cache item.If include the HAO of IPv4 among the BU, then be necessary for another cache entry of this address creation, and all point to the CoA place of the IPv4 of MN.Therefore, so destination address is the grouping of the HoA of MN, all will be encapsulated in the IPv4 head.Wherein, source address is the HoA of the IPv4 of HA, and destination address is the CoA of the IPv4 of MN.
Field network is only supported the privately owned address of IPv4:
At this moment, MN need send the place, IPv4 address that the IPv6 that comprises BU is grouped into HA by the tunnel.In order to pass through network address translation apparatus (NAT-Network Address Translation), the IPv6 grouping is encapsulated in the IPv4 grouping to transmit by the tunnel with User Datagram Protoco (UDP) (UDP-User Datagram Protocol) form.In the IPv4 head, the IPv4 address that MN uses field network to obtain is source address, the HoA of the IPv6 of the MN that comprises among the HAO of BU.The content of IPv6 grouping is identical with the public address occasion.
After receiving BU, the HoA that HA is necessary for the IPv6 of MN creates the banding cache item.If include the HAO of IPv4, then must create in addition.All all must point to the CoA place of the IPv4 of MN in the IPv6 source of packets address field.Therefore, destination address is that the grouping of the HoA of MN all encapsulates with UDP, and then is encapsulated in the IPv4 head, and the source address in this is the IPv4 address of HA, and destination address is the CoA of the IPv4 of MN.
In two stack situations, the routing optimization manner between MN and the CN is no longer suitable, and all groupings all will be transmitted through HA.
What the defective of known technology two was to move in this scheme is the MIPv6 agreement, therefore has all deficiencies of MIPv6 agreement, and especially the business for delay sensitive can't be suitable for.
Summary of the invention
Main purpose of the present invention is to address the aforementioned drawbacks, and avoids defective to exist, and a kind of double-stack support extension method of layer mobile IPv 6 protocol is provided, and when MN moved to field network, this method comprised:
MN receives the RA of the IPv6 address that includes MAP;
MN configuration RCoA and LCoA;
MN and MAP bind;
MN and HA bind, and MN comprises the HoA that a HAO deposits the IPv6 of MN in mailing to the BU of HA, the HA response, and the BA that sends the HAO that includes correspondence is to MN.
A kind of double-stack support extension method of layer mobile IPv 6 protocol further comprises: MN receives the RA of the IPv4 address that includes MAP in the external region.
A kind of double-stack support extension method of layer mobile IPv 6 protocol in the process that MN and HA bind, also comprises the HoA that a HAO deposits the IPv4 of MN in mailing to the BU of HA.
A kind of double-stack support extension method of layer mobile IPv 6 protocol, the RCoA of MN configuration generates with the stateless address auto configuration mode according to the MAP address prefix.
The present invention compares with known technology, and its advantage is: in a dual-stack network, move two cover mobility management protocols respectively, not only increased the complexity of each node configuration, and increased the complexity of network management.In case node is moved, will produce a large amount of information interactions, the performance of whole network is caused very big influence.The present invention can reduce the complexity that network entity disposes so that only move once the cover mobility management protocol and can realize management to mobile node by expansion in dual-stack network, improved the performance of network management.
Description of drawings
Fig. 1 is the basic topological diagram of layer mobile IPv 6 protocol;
Fig. 2 is the information interactive process figure of dual-stack network.
Embodiment
Relevant technology contents of the present invention and detailed description, existing conjunction with figs. is described as follows:
In order to make the HMIPv6 agreement can support the dual-stack network node, the present invention proposes a kind of double-stack support extension method of HMIPv6 agreement, be used for realizing support to IPv4, two kinds of business of IPv6.In HMIPv6, all groupings between MN and the CN all are to be undertaken by the forwarding of HA and MAP.When CN when MN sends grouping, at first intercept and capture, and be transmitted to the MAP in territory, the current place of MN by HA, send to MN by MAP then.
In dual-stack network, be that all nodes are all supported IPv4 and IPv6 in the network, the present invention need revise RA, increase the IPv4 address notification of MAP in the external region, also needing increases the HoA that a home address option HAO deposits the IPv4 of MN in mailing to the BU of HA, guarantee the dual stack node communication continuity.When MN leaves the local and moves to field network, the support situation of IPv6 is disposed the CoA of needs according to field network.Because the mobility management protocol that moves in the network is the HMIPv6 agreement, so necessarily there is a MAP who supports IPv6, the present invention supposes that said MAP all supports IPv6.And MAP also has an IPv4 address and an IPv6 address simultaneously.
Fig. 1 is the basic topological diagram of layer mobile IPv 6 protocol.When field network was supported IPv6, the operating process of whole HMIPv6 agreement was with just the same under the general case, and current RCoA and LCoA are IPv6's.Only need at MN in the Binding Update of HA, increase a HAO about the IPv4 of MN, with the banding cache clauses and subclauses of generation IPv4 on HA and MAP, be used for guaranteeing MN can with the node communication of IPv4.When field network is not supported IPv6, at this moment will there be three kinds of situations to discuss respectively.Below MN place field network branch situation is described.Fig. 2 is the information interactive process figure of dual-stack network.
1, field network is supported IPv6:
AR1 among the present invention below the MAP and AR2 support IPv6.When MN moves to such field network, receive the RA that includes the MAP address, in this announcement, both included the IPv6 address of MAP, also include the IPv4 address of MAP, and Care-of Address.Here will preferentially use the IPv6 address.
According to the HMIPv6 agreement, MN can dispose RCoA and LCoA, because current AR supports IPv6, so these two addresses are the address of IPv6.After completing successfully address configuration, MN need send LBU to MAP, and the source address of this information is LCoA, and destination address is the IPv6 address of MAP.Wherein fill the RCoA of MN among the HAO.LCoA status in the MAP territory in HMIPv6 is the same with CoA status in the HA territory in MIPv6, need set up the banding cache clauses and subclauses with it, that fill among the HAO should be the HoA of MN, that correspondence is RCoA in HMIPv6, since in MIPv6 MN communicate by letter with CN the use the address be the HoA of MN, but be grouped in the CoA that the source address of using in the network routing procedure is MN, should be filled in the HoA of MN among the HAO this moment, when being submitted to upper level applications after CN receives, upper level applications can know whether it is the grouping that the MN with this interprogram communication sends by HAO.At this moment the form of LBU is just the same with what stipulate among the HMIPv6.
Behind the LBU that receives MN, MAP sends the LBA information of response, and the source address of this information is the IPv6 address of MAP, and destination address is the LCoA of MN.That HoA wherein confirms the option filling is the RCoA of MN.
Afterwards, MN sends BU to HA.The source address of this information is the RCoA of MN, and destination address is the IPv6 address of HA.The IPv6 address of HA is meant the consistent address of prefix with the HoA of MN, so that MN mails to the normal route of grouping energy of HA, when MN does not roam into field network as yet, be to insert network that HA connected by the interface that has disposed this address on the HA, i.e. all groupings of mailing to MN all will arrive MN by this interface of HA, and vice versa.Here, the same with MIPv6 to the dual stack node support, in this BU, increase the HoA that a HAO is used for depositing the IPv4 of MN.So fill the IPv4 of MN and the HoA of IPv6 respectively among two HAO among this BU.Like this, HA need set up two banding cache items for MN, a HoA who deposits IPv6, and another deposits the HoA of IPv4, but these two RCoA that all point to MN.Wherein the HoA of IPv4 is the HoA of the IPv4 of MN.This address is the IPv4 address of a clean culture, may be disclosed, also may be privately owned.This address is a disclosed IPv4 address among the present invention.For the situation of privately owned address, only need to get final product through NAT.
Before MN does not move to other network as yet, be direct communication between HA and the MN, do not need to transmit, when MN moves to field network, the grouping that CN mails to MN still will be mail to HA, because CN does not also know whether MN moves, but MN moves the back receives BA by above-mentioned transmission BU process, the banding cache clauses and subclauses have been set up at HA, the current address CoA and corresponding HoA of MN have wherein been indicated, if HA exists these clauses and subclauses can be according to the HoA of these clauses and subclauses and the mapping relations between the CoA receiving after CN mails to the grouping of MN like this, find grouping purpose MN the current address and transmit this and be grouped into MN.
Receive the BU of MN as HA after, send the BA information of response, the source address of this information is the IPv6 address of HA, and destination address is the RCoA of MN.In this BA, two HoA of response confirm to fill the IPv4 of MN and the HoA of IPv6 respectively in the option.Wherein the HoA of IPv4 is the HoA of the IPv4 that HA will use in banding cache.This address is the IPv4 address of a clean culture, may be disclosed, also may be privately owned.This address is a disclosed IPv4 address among the present invention.For the situation of privately owned address, only need to get final product through NAT.Automatically distribute if this address is DHCP, HA notifies this address to MN.Otherwise if this address is that static configuration is given MN, then HA will duplicate this address from BU information.
At this moment, when the CN of two stacks and MN communicate, this grouping will at first be intercepted and captured by HA, be transmitted to the MAP in territory, MN place then, be transmitted to the position at the current place of MN at last by MAP, wherein the service traffics of IPv4 are encapsulated in the tunnel of IPv6, thereby have guaranteed MN communication continuity in dual-stack network.
2, field network is not supported IPv6
When AR only supported IPv4, the LCoA of MN configuration was the IPv4 address, and at this moment at first MN need send LBU to MAP, registers the LCoA of its IPv4.After waiting to receive the LBA of MAP, MN just can send BU to HA, registers its RCoA, at this moment, necessarily also needs to comprise a HAO who carries the IPv6 of MN.The business that can guarantee IPv4 and IPv6 is transmitted simultaneously like this.
Do not support IPv6 for field network, have following three kinds of situations when moving the triggering switching:
When MN moves to a field network, receive the RA that includes the MAP address, in this announcement, both included the IPv6 address of MAP, also include the IPv4 address of MAP, and Care-of Address.
(1) AR1 supports IPv6, and AR2 supports IPv4;
This situation represents that MN supports the AR1 of IPv6 to move to an AR2 who only supports IPv4 from one.Because what take place is the mobile MAP territory in,, and have only LCoA to change so the RCoA of MN does not change.At this moment, because current AR2 is the node of an IPv4, so the LCoA that configuration produces is exactly an IPv4 address.MN need send LBU to MAP, and source address is the LCoA of MN in the IPv4 head of this information, and destination address is the IPv4 address of MAP.Source address is the RCoA of MN in the internal layer IPv6 head, and destination address is the IPv6 address of MAP.Because AR2 only supports IPv4, therefore MN can only use the LCoA of IPv4, the LCoA of IPv6 before can't use, therefore can only be encapsulated in internal layer packet header to the RCoA of IPv6, because need set up the banding cache clauses and subclauses with RCoA, so as MAP intercept and capture the IPv6 mail to MN RCoA bag and be encapsulated in the IPv4 bag and issue MN.That HAO wherein fills is the RCoA of MN.At this moment the form of LBU is just the same with what stipulate among the HMIPv6.
Behind the LBU that receives MN, MAP sends the LBA information of response, and source address is the IPv4 address of MAP in the IPv4 head of this information, and destination address is the LCoA of the IPv4 of MN.Source address is the IPv6 address of MAP in the IPv6 head, and destination address is the RCoA of the IPv6 of MN.That HoA wherein confirms the option filling is the RCoA of MN.
At this moment, because the MN generation is moving in the MAP territory, according to the HMIPv6 agreement, MN does not need to send binding to HA.MAP intercepts and captures all groupings that HA mails to the RCoA of MN, then the flow of IPv6 is encapsulated in the LCoA that the IPv4 tunnel mails to the IPv4 of MN.
(2) AR1 supports IPv4, and AR2 also supports IPv4;
When MN was mobile between such two AR, LCoA changed, and MN is identical with (1) with the binding procedure of MAP.
(3) AR1 supports IPv4, and AR2 supports IPv6.
This situation represents that MN supports the AR1 of IPv4 to move to an AR2 who only supports IPv6 from one.Because what take place is the mobile MAP territory in,, and have only LCoA to change so the RCoA of MN does not change.At this moment, because current AR2 is the node of an IPv6, so the LCoA that configuration produces is exactly an IPv6 address.MN need send LBU to MAP, and source address is the LCoA of MN in the IPv6 head of this information, and destination address is the IPv6 address of MAP.That HAO wherein fills is the RCoA of MN.At this moment the form of LBU is just the same with what stipulate among the HMIPv6.
Behind the LBU that receives MN, MAP sends the LBA information of response, and source address is the IPv4 address of MAP in the IPv4 head of this information, and destination address is the LCoA of MN.Source address is the IPv6 address of MAP in the IPv6 head, and destination address is the LCoA of MN.That HAO wherein fills is the RCoA of MN.
At this moment, because the MN generation is moving in the MAP territory, according to the HMIPv6 agreement, MN does not need to send binding to HA.Other transmitted in packets is similar with the IPv6 situation.
(1), the essential distinction of (3) two kinds of situations is, source address in the packet header of the LBU that sends is different with the destination address agreement: the former is IPv4, the latter is IPv6, its basic reason is whether support IPv6, after switching is finished in addition, the MAP that sets up is also different to the tunnel type of MN, and the former is IPv4-in-IPv4, and the latter is IPv6-in-IPv6.But which kind of situation no matter, that register on HA all is the RCoA of the IPv6 of MN, and the former RCoA on MAP points to the LCoA of the IPv4 of MN, that the latter points to then is the LCoA of IPv6.
When MN was positioned at a network of only supporting IPv4, will send the IPv6 grouping that is encapsulated in the IPv4 tunnel to HA according to following form: in outer IPv4 head, source address was the LCoA of the IPv4 of MN, and destination address is the IPv4 address of HA.In internal layer IPv6 head, source address is the HoA of the IPv6 of MN, and destination address is the IPv6 address of CN.Similarly, MN will send the IPv4 tunnel packet to HA with following form: in outer IPv4 head, source address is the LCoA of the IPv4 of MN, and destination address is the IPv4 address of HA.In internal layer IPv4 head, source address is the HoA of the IPv4 of MN, and destination address is the IPv4 address of CN.
The present invention mainly is the angle from IPv6, realizes by modification HMIPv6 agreement.And in fact for the support of dual stack node, also can realize by revising MIPv4 agreement and MIPv4 regional registration.
Above-mentioned is preferred embodiment of the present invention only, is not to be used for limiting the scope of the present invention.Be that all equalizations of being done according to the present patent application claim change and modification, be all claim of the present invention and contain.