CN1972295A - IPv4 over IPv6 extension method of BGP protocol software - Google Patents

Abstract

This invention relates to router transfer technique field by IPv4 over IPv6 extending layer based on BGP agreement, which is characterized by the following: filling the relationship between user network IPv4 address and IPv6 address into expend property of the Update message and sending the each boundary router between them; otherwise storing the relationship into local router sealing list.

Description

The IPv4 over IPv6 extended method of bgp protocol software

Technical field

The IPv4 over IPv6 expansion of bgp protocol software belongs to Internet technical field, relates in particular to the router retransmission technique of network layer.

Background technology

Along with the fast-developing of IPv6 technology and progressively ripe and extensively be used, the IPv6 network size progressively enlarges.Some countries on policy in order to promote the transition of network, quicken the IPv6 networks development, promote IPv6 and development of technologies thereof, built up pure IPv6 backbone network, as CNGI demonstrative project (CNGI:China Next GenerationInternet).Yet, existing a large amount of internet, applications and service still are present in the IPv4 network, a large amount of application and service like this is transplanted on the IPv6, never be the work that to finish by short-term, therefore, IPv4 and IPv6 network are with long-term co-existence, and the transition of its IPv4/v6 will have chronicity and complexity.

The IPv4/v6 transit mechanism is as a hot issue of research field, and researchers have proposed a large amount of schemes.Existing achievement in research can be divided into dual stack, tunnel and protocol translation three classes.Yet present most of IPv4/v6 network transition scheme mainly is to solve based on interconnect problem between the small-scale IPv6 access network of IPv4 backbone network and the interconnecting property problem between the IPv4/IPv6 network, and need realize interconnected demand by the IPv6 backbone network at a large amount of IPv4 networks that occurred, correlative study is very few.Though the general tunnel of IPv6/generic route encapsulation GRE over IPv6 tunnel, can realize the IPv4 network to a certain extent by the IPv6 network interconnection based on the tunnel of multiprotocol label switching MPLS and dynamic synchronous transfer mode, DTM DSTM technology, yet because have poor expandability, protocol is complicated or need be to network problem such as make amendment.In the large-scale complex network, pass through between a large amount of IPv4 networks to dispose and realize that these mechanism are difficult under the interconnected situation of IPv6 core net that its management maintenance burden has exceeded controllable scope.

A kind of IPv4 network interconnection transit mechanism framework based on IPv6 (being called for short 4over6 mechanism) has solved the IPv4 network need realize interconnected problem by the IPv6 backbone network.4over6 mechanism is by the acting in conjunction of 4over6 map addresses, 4over6 route and 4over6 packet transaction, realize that the IPv4 network is interconnected by the IPv6 backbone network, and do not need to set up explicit tunnel, but and possess network and end system transparent characteristic, light-duty self adaptation dynamic routing characteristic, have less network management maintenance load, can be adapted to the interconnected of large scale network.Yet there is potential route loop problem in the 4over6 fundamental mechanism under interconnected environment such as multiple spot access and multi-path transmission.In addition, have the semantically transparent characteristics of upper-layer protocol, so the grouping translating mechanism in the 4over6 mechanism packet transaction possesses the possibility of transparent exchanging visit between the IPv4/v6 network because 4over6 grouping translation is transmitted.

The present invention is directed in the IPv6 backbone network process of construction, transparent exchanging visit demand between IPv4 over IPv6 transmission problem that faces between the multi self-governing system under the complicated interconnected environment of being badly in need of solving and the IPv4/v6 network, on basic 4over6 manufacturing basis, in conjunction with bgp protocol and AS Path attribute thereof, designed and Implemented the 4over6 expansion of BGP software, not only well solved the route loop problem, and towards the interconnected environment of complexity, solved the IPv4 network by interconnect problem between the territory of large-scale ip v6 backbone network, the problem of the transparent intercommunication of IPv4/v6 possesses good extensibility.

Summary of the invention

The objective of the invention is to utilize the multi-protocols MP attribute of bgp protocol to transmit the routing iinformation of IPv4 over IPv6, the corresponding relation of the IPv6 address of the 4over6 virtual interface of PE (Provider Edge, the border router of IPv4 and IPv6) and the IPv4 network that links to each other with it just.

Method proposed by the invention is characterised in that: it is characterized in that making between each IPv4 network and the IPv6 backbone network and all set the 4over6 border router, contain following steps successively:

Step (1), the routing protocol software of modification border router;

For the Open message of the usefulness that connects, set following parameter successively: the parameter type territory of 1 byte, the expression parameter is a Capabilities Advertisement type; The parameter length territory of 1 byte, the byte number of expression subsequent parameter value; The parameter value of 6 bytes is followed successively by: the multi-protocols extended coding Capability Code of 1 byte, the Capability Length field of 1 byte; The Address-Family Identifier symbol AFI field of two bytes, the reserved field of 1 byte; The identifier SAFI of the subsequent address family field of 1 byte,

Update message for renewal usefulness:

Form for the Network Layer Reachable Information attribute MP_REACH_NLRI attribute of multi-protocols expansions is followed successively by: the AFI of 2 bytes, and the SAFI of 1 byte, next of 1 byte jumped length; The IPv6 address of the 4over6 virtual interface of next variable jumping and variable purpose IPv4 network address NLRI;

Form for the unreachable information attribute MP_UNREACH_NLRI of the network attribute of multi-protocols expansions, be followed successively by: the AFI of 2 bytes, the SAFI of 1 byte cancels the set that routed domain extends this as the prefix content of the IPv4 prefix length of 1 byte and variable-length to variable quilt;

Step (2), the following order of configuration 4over6 on border router;

The order of the interface of the IPv4 of configure boundaries router and virtual interface address,

The order of the interface of the IPv6 of configure boundaries router and virtual interface address,

Dispose the order of following bgp protocol: the interface IP address of each reciprocity node peer and autonomous system AS number; The starting command of redirect command and 4over6;

Step (3), local border router is set up reciprocity peer relation by the Open message that sends Open messages to other reciprocity node peer or accept other reciprocity node peer;

Step (4), each border router, comprise local border router, corresponding relation between the IPv6 address of IPv4 address in the customer network and IPv4 overIPv6 virtual interface is filled up in the extended attribute of Update message of bgp protocol, and between each border router, transmit, so that the information of the border router of a corresponding customer network and correspondence thereof propagates into other border router;

Step (5), local border router receives and resolves the Update message that other routers are sent, and obtains new 4over6 route, just the corresponding relation of the IPv6 address of IPv4 address in other customer network and IPv4 over IPv6 virtual interface;

Step (6) is given the routing management module of the bgp protocol of local border router new 4over6 route, through route metric value relatively, selects optimum route, carries out following steps:

Step (6.1) is propagated resulting new route to other reciprocity node peer;

Step (6.2) is given the encapsulating sheet of the storing IP v4 network address and 4over6 virtual interface corresponding relation new route;

Step (6.3) is given IPv4 routing table the IPv4 network address in the new route and the IPv4 address of local 4over6 virtual interface.

The present invention has been implemented and has been deployed on the Chinese scientific research and education net Cernet2 of future generation.

Description of drawings

Fig. 1 .4over6 system topological schematic diagram;

Fig. 2 .MP_REACH_NLRI attribute fills in form;

Fig. 3 .MP_UNREACH_NLRI attribute fills in form, and wherein arrow is represented being cancelled the explanation of routed domain;

The IPv4 over IPv6 expansion flow chart of Fig. 4 .BGP protocol software.

Embodiment

The implementation of this module is to add the code relevant with 4over6 on the basis of the bgp protocol software of Zebra.Reason is: 1) running environment of Zebra is same with the environment facies that present 4over6 will realize, all is Linux; 2) realization of Zebra0.94 version is more comprehensive, has comprised the basic code of BGP-MP expansion.

The secondary module of BGP-MP among the 4over6 can be divided into three modified modules and two interface modules:

● revise the packet parsing constructor: the parsing and the construction process of the new ident value relevant with 4over6 of the BGP-MP attribute of open message and update message are joined in original code.Concrete message is filled in a form trifle as follows.

● add an OAM order: the 4over6 function that starts BGP gets final product.4over6 started at 4over6vif and its mandatum cassatorium (no 4over6 started at 4over6vif).

● the interface of routing management and bgp protocol: utilize the Zebra agreement, the IPv4 network that bgp protocol is received joins local routing table with the IPv4 address of local virtual interface, like this when the IPv4 of this PE interface receives that purpose is the message of the IPv4 network that links to each other with another PE, can directly message be given the IPv4overIPv6 virtual interface of this machine and handle, virtual interface is transmitted its encapsulation back in the IPv6 network.

● the interface of encapsulating sheet and bgp protocol: bgp protocol receives that whenever a new route offers bgp protocol by encapsulating sheet particular interface passes to encapsulating sheet with this route, two kinds of situations are arranged, a kind of is to insert, but the IPv4 network that does not have before comprising or had in the past the IPv6 address change of the virtual interface of its corresponding PE, another kind is deletion.

The Open message will increase an optional parameters, at first is the parameter type territory of 1 byte, is 2 in the 4over6 intermediate value, and the expression parameter is a capability advertisement Capabilities Advertisement type; Next being the parameter length territory of a byte, is 6 in the 4over6 intermediate value, and the parameter value of expression back has 6 bytes; Next fill in parameter value, the 1st byte of parameter value is capability code Capability code, value is 1, the expansion of expression multi-protocols, the 2nd byte is Capability length field, extend this as 4,3rd, 4 bytes are AFI (address family identifier) fields, the BGP message is propagated in IPv6, so extend this as 2, the 5th byte is a reserved field, and the 6th byte is SAFI (Subsequent address family identifier) field, extends this as 67 in 4over6.(another kind of arguement is to fill in into 128 according to VPN.)

The Update message utilizes the multi-protocols extended attribute to propagate the corresponding relation of the virtual interface of IPv4 and PE, be MP_REACH_NLRI attribute, its types value is 14, utilize MP_UNREACH_NLRI attribute to cancel route in addition, and can not cancel with the field of the path attribute front of BGP update message.Wherein the length of NLRI does not have explicitly to provide, but can come out by the total length and the preceding two-part length computation of UPDATE frame.MP_REACH_NLRIattribute fills in form such as Fig. 2.

The MP_UNREACH_NLRI attribute fill in form such as Fig. 3, wherein the total length of the Withdrawn routes of back does not show and provides, length that it can subtract two fixed fields in front by the length of attribute is calculated.Every route cancelling is identical with the form of common IPv4 prefix, all is to provide prefix length earlier, is concrete prefix content then.

Claims (1)

1, the IPv4 over IPv6 extended method of Border Gateway Protocol software is characterized in that making between each IPv4 network and the IPv6 backbone network and all sets the 4over6 border router, contains following steps successively:

Step (1), the routing protocol software of modification border router;

For the Open message of the usefulness that connects, set following parameter successively: the parameter type territory of 1 byte, the expression parameter is a Capabilities Advertisement type; The parameter length territory of 1 byte, the byte number of expression subsequent parameter value; The parameter value of 6 bytes is followed successively by: the multi-protocols extended coding Capability Code of 1 byte, the Capability Length field of 1 byte; The Address-Family Identifier symbol AFI field of two bytes, the reserved field of 1 byte; The identifier SAFI of the subsequent address family field of 1 byte,

Update message for renewal usefulness:

Form for the Network Layer Reachable Information attribute MP_REACH_NLRI attribute of multi-protocols expansions is followed successively by: the AFI of 2 bytes, and the SAFI of 1 byte, next of 1 byte jumped length; The IPv6 address of the 4over6 virtual interface of next variable jumping and variable purpose IPv4 network address NLRI;

Form for the unreachable information attribute MP_UNREACH_NLRI of the network attribute of multi-protocols expansions, be followed successively by: the AFI of 2 bytes, the SAFI of 1 byte cancels the set that routed domain extends this as the prefix content of the IPv4 prefix length of 1 byte and variable-length to variable quilt;

Step (2), the following order of configuration 4over6 on border router;

The order of the interface of the IPv4 of configure boundaries router and virtual interface address,

The order of the interface of the IPv6 of configure boundaries router and virtual interface address,

Dispose the order of following bgp protocol: the interface IP address of each reciprocity node peer and autonomous system AS number; The starting command of redirect command and 4over6;

Step (3), local border router is set up reciprocity peer relation by the Open message that sends Open messages to other reciprocity node peer or accept other reciprocity node peer;

Step (4), each border router, comprise local border router, corresponding relation between the IPv6 address of IPv4 address in the customer network and IPv4 overIPv6 virtual interface is filled up in the extended attribute of Update message of bgp protocol, and between each border router, transmit, so that the information of the border router of a corresponding customer network and correspondence thereof propagates into other border router;

Step (5), local border router receives and resolves the Update message that other routers are sent, and obtains new 4over6 route, just the corresponding relation of the IPv6 address of IPv4 address in other customer network and IPv4 over IPv6 virtual interface;

Step (6) is given the routing management module of the bgp protocol of local border router new 4over6 route, through route metric value relatively, selects optimum route, carries out following steps:

Step (6.1) is propagated resulting new route to other reciprocity node peer;

Step (6.2) is given the encapsulating sheet of the storing IP v4 network address and 4over6 virtual interface corresponding relation new route;

Step (6.3) is given IPv4 routing table the IPv4 network address in the new route and the IPv4 address of local 4over6 virtual interface.

Images