CN100409646C - Method and device for realizing different IP data packet repeating using strategic flow - Google Patents
Method and device for realizing different IP data packet repeating using strategic flow Download PDFInfo
- Publication number
- CN100409646C CN100409646C CNB2003101017904A CN200310101790A CN100409646C CN 100409646 C CN100409646 C CN 100409646C CN B2003101017904 A CNB2003101017904 A CN B2003101017904A CN 200310101790 A CN200310101790 A CN 200310101790A CN 100409646 C CN100409646 C CN 100409646C
- Authority
- CN
- China
- Prior art keywords
- address
- ipv6
- ipv4
- stream
- tunnel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Abstract
The present invention provides a method and a device for realizing the simultaneous support for IPv4 and IPv6 forwarding in the same physical pipeline in the mode of strategic flow, wherein multiple attribute groups of corresponding data packets are taken according to data flow types of different target IP so as to calculate strategic flow identification ID, and thus, forwarding tables of the strategic flow are accurately searched; if a matched forwarding table of the strategic flow is discovered, correlated contents of a data packet are modified and forwarded according to the table, else, the data packet is correspondingly processed by a control plane according to the attributes of the data packet, and besides, a corresponding unique forwarding table of local strategic flow is generated by integrating information in the control plane and various kinds of tables correlated to application; finally, the forwarding table of the strategic flow is dispensed to the forwarding plane and used for a rearward data packet. As a result, the present invention avoids complicatedly searching IPv4 and IPv6 packet by packet, enhances forwarding efficiency and avoids introducing extra signal commands. Besides, the present invention can not influence the interconnection interoperability of the Internet.
Description
Technical field
The present invention relates to a kind of digital communication equipment, the method and apparatus that particularly a kind of packet is transmitted, it adopts strategic stream mode to support IPv4 and IPv6 to transmit in same physical pipeline simultaneously.
Background technology
Along with the continuous development of Internet technology, many defectives of existing IPv4 network come out gradually, and wherein the most outstanding is that IP address space is with the depleted and ever-increasing problem of backbone route table.According to present Internet development speed, the IPv4 address will assign in the period of 2005 to 2010.In order thoroughly to solve problem that present IPv4 runs into and to provide better support to the application in future, the IPng working group of Internet engineering group (IETF) (IPng Working Group) has proposed to revise the suggestion of IP agreement.New suggestion is the 6th version of IP, therefore is also referred to as IPv6.IPv6 has and reaches 128 address space, can thoroughly solve the problem of IPv4 address resource deficiency.In addition, IPv6 also adopts hierarchical address pattern, efficient IP packet header, service quality, host address to dispose automatically, authenticate and multinomial technology such as encryption.
In addition, 3G (Third Generation) Moblie 3GPP and 3GPP2 are defined as following mobile network's core technology to IPv6.Because the growth of mobile communication subscriber is more faster than fixed network user, particularly various fast developments with portable terminal of network savvy make that mobile network plays leading role in the process that advances IPv6, and fixed network can follow up subsequently.
Relevant expert's prediction, by 2010, the industry size relevant with IPv6 will reach 20,000,000,000 dollars at least.Because the IPv6 technology has huge market prospects, each country of the whole world has launched the research to IPv6, and obtains certain progress.In Japan, Europe and the U.S., each major company has released the router based on IPv6 respectively, and has begun building of experimental network.And increase rapidly owing to the Internet user in China, Chinese Government has also recognized the importance of IPv6, also research and development and the deployment to IPv6 carried out supporting energetically in the project of 863 Program and State Planning Commission.
The conversion of how finishing from existing IPv4 to IPv6 is first problem that the IPv6 development need solves.Existing almost each network and connection device thereof are all supported IPv4, and therefore wanting is unpractical with regard to the conversion of finishing from IPv4 to IPv6 in one night.IPv6 must be able to support the leftover problem with process IP v4 system.Therefore, in foreseeable future, can realize that IPv4 will be that can decision IPv6 the most critical commercial on a large scale problem to the level and smooth transition of IPv6.And the key that addresses this problem just is how to be implemented in the coexistence of carrying out IPv4 and IPv6 forwarding in the same physical pipeline on the network equipment.
The original solution and the deficiency that the present invention relates to content are described below.
Problem and shortage based on the IP forwarding of packet is at first described.
In prior network device, the overwhelming majority has adopted the IP based on bag to transmit.Traditional route querying is based on the longest matched and searched of purpose IP address, promptly finds one and the longest coupling clauses and subclauses of purpose IP address prefix in route.Following three clauses and subclauses are for example arranged in routing table (table 1):
Table 1 routing table
| Prefix(12bit) | Output Index |
| P1＝0101000000 | Ethernet 1/1 |
| P2＝0101110100 | |
| P3＝0101101011 | FDDI 3/1 |
If a packet is arranged, its preceding 12 bit are 01010110111, and then it will mate P1, sends from Ethernet 1/1.And if its preceding 12 bit are 0101101011, then it can mate prefix P3, sends from FDDI 3/1.In the most frequently used longest prefix match scheme, mainly be based on the lookup scheme of Radix Trie table, this algorithm implements comparatively simple, but efficient is lower, under the poorest situation, needs 32 or 128 internal storage access (corresponding respectively to IPv4 or IPv6).And in some improved algorithm, the efficient of searching greatly improves, but still needs O (log2W) inferior, and wherein W is the bit figure place of IP address.If also promptly under the IPv4 situation, need 5 times, and under the IPV6 situation, need 8 times.
This based on the longest coupling the mode of searching, can't adopt same table to come the environment of IPv4/IPv6 coexistence, must adopt many routing tables to support searching of IPv4 and IPv6 respectively.The difficulty that has so just caused the low and equipment development of search efficiency.In addition, the longest matching way can't be handled at each stream of user, the QoS that causes operator to be difficult to provide value-added service and to guarantee user's stream.
Next the problem and shortage that MPLS transmits is described.
The MPLS technology is present popular IP switching technology, and it also is used to solve IPv4, IPv6 at present and transmits and service quality problem.So-called MPLS, i.e. multi protocol label exchange.Label switched, the forwarding mechanism that refers to bottom has adopted simple label switched, and mark is incited somebody to action sign in return.When label switched with ATM or FR (frame relay) during as its link layer protocol, mark also corresponding employing VPI/VCI or DLCI.When the link layer of label switched is FDDI, Ethernet or PPP, because do not have label information in their original forms fully, must add extra encapsulation, what label switched adopted is the form of Shim.
Two layers of retransmission technique that MPLS adopts topology to drive.But when adopting MPLS to transmit, it need finish the mapping of forwarding equivalence class (FEC, Forwarding Equal Class) to mark (Label) at fringe node equally, then carries out quick forwarding based on mark at core node.Therefore, this means that the node at network edge need carry out the longest matched and searched and complicated strategy processing (as VPN, NAT and PPPOE) to each packet equally.In addition, for the Signalling exchange between support node, it has introduced complicated signaling protocol, as LDP (CR-LDP) or RSVP-TE, increase with regard to the software complexity that has caused equipment, and, the two big standards of present MPLS support that this has also caused the difficulty of network interoperability because being respectively different manufacturers.
Though the MPLS scheme is transmitted to compare at intermediate node with traditional IP and has been improved the efficient of transmitting, but still has the following disadvantages:
1, MPLS is in the complex process of fringe node to packet, transmit identical with IP based on bag, it still need carry out the longest matched and searched on LER, finish the mapping from FEC to Label, therefore long and equipment complexity of the construction cycle of the edge device that causes equally.
2, the realization of MPLS need be introduced complicated MPLS signaling, increases the development cost and the maintenance cost of equipment greatly.IETF RFC relevant with MPLS and draft are nearly individual surplus in the of 80, and the equipment development merchant implements very complicated, and operator also is difficult to safeguard.
3, owing to two kinds of signaling protocols of support in the standard of MPLS, LDP (CR-LDP) and RSVP, and manufacturer that this two big sign is respectively different supports.So the use mpls protocol can cause the interoperability of the network equipment to have problems.
When 4, among the MPLS assurance of service quality being adopted Differentiated Services, can't guarantee service quality at each stream.
5, MPLS can't network enabled address/protocol conversion (NAT-PT), therefore can't realize directly interconnected between IPv4 and the IPv6 network.
Summary of the invention
The object of the present invention is to provide a kind of strategic stream mode that adopts to support the method and apparatus that IPv4 and IPv6 transmit simultaneously in same physical pipeline, it can avoid complexity to search, and improves forward efficiency.
According to the present invention, provide a kind of and in the network equipment, realized the method that packet is transmitted with strategic stream mode, the described network equipment comprises Forwarding plane and control plane at least.This method may further comprise the steps: (a) packet in the receiving data stream comprises at least one packet in this data flow; (b) judge the type of purpose IP address at Forwarding plane; (c) in the data stream type of Forwarding plane according to different purpose IP, extract the polynary set of properties of corresponding packet, calculate local unique tactful traffic identifier ID of this data flow of sign; (d) Forwarding plane if discovery has matched strategy stream forwarding list order then carry out step (e), otherwise forwards step (f) according to the accurate search strategy stream forwarding list of described strategy stream ID order to; (e), then according to described tactful stream forwarding list order this packet is carried out related content and revise and transmit and operate, and forward step (i) at Forwarding plane if find to have matched strategy stream forwarding list order; (f) if find no matched strategy stream forwarding list order, illustrate that then first bag or tactful stream forwarding list order that this packet is a data flow are aging, Forwarding plane is delivered to control plane with this packet and is handled; (g) control plane the outbound port of this locality configuration relevant with application and pairing next jumping in IP address of this purpose according to type, the inbound port of the purpose IP address of this packet with use in the relevant configuration at least one, come jointly packet is handled accordingly with Forwarding plane, and the entry-information relevant with various application generates a corresponding unique tactful stream forwarding list order in the Comprehensive Control plane; (h) control plane should be distributed to described Forwarding plane by strategy stream forwarding list order, for the packet use of back; (i) handle next packet.
In the present invention, when finding no matched strategy stream forwarding list order, control plane does not need to carry out mutual signaling with other network equipment.In addition, when control plane generates new tactful stream forwarding list order, do not need this entry is announced to other network equipment yet.
According to the present invention, a kind of data transfer equipment of transmitting data with strategic stream mode also is provided, comprising an at least one Forwarding plane and a control plane, the packet in each Forwarding plane receiving data stream wherein, comprise at least one packet in this data flow, this data transfer equipment comprises: tactful stream forwarding list order storage area, be used for storage policy stream forwarding list order, wherein, described Forwarding plane comprises: purpose IP address class type judgment part is used to judge the data stream type of purpose IP address; Strategy stream ID computing unit is used for the type according to different purpose IP, selects the polynary set of properties of corresponding packet, calculates the local policy traffic identifier ID of this data flow of sign; Search the unit, be used for accurately searching the tactful stream forwarding list order that described tactful stream forwarding list order storage area is stored, see if there is described strategy stream ID matched strategy stream forwarding list order with this packet according to described strategy stream ID; Revise and retransmission unit,, then this packet is carried out related content and revise and transmit operation according to described tactful stream forwarding list order if at the tactful stream forwarding list order of searching unit discovery coupling; As not finding the tactful stream forwarding list order that mates searching the unit, then this packet is delivered to control plane, described control plane comprises: policy processing unit, according in the configuration of the outbound port of the configuration of the type of purpose IP address, inbound port and pairing next jumping in IP address of this purpose at least one, packet is handled accordingly; Strategy stream forwarding list order generation unit generates corresponding tactful stream forwarding list order according to the processing that packet is carried out, and should be distributed to described tactful stream forwarding list order memory cell by strategy stream forwarding list order, for the packet use of back.
The present invention is different from the mode that traditional packet-by-packet (Packet-based) transmits, and adopted the IP that handles based on data flow (Stream-based) to wrap the thought of transmitting, and, different policing types realized in same physical pipeline, carrying out simultaneously IPv4 and IPv6 forwarding of data by being set.The present invention has avoided packet-by-packet searching of complicated IPv4 and IPv6, has improved forward efficiency greatly, and has avoided introducing other extra signalings, can not influence the interconnecting property of the Internet.In addition, the present invention adopts same table to support multiple application, handles at each stream, makes operator provide business and service quality to become possibility at each stream of user.
Description of drawings
Fig. 1 to Fig. 4 is from IPv4 different phase network diagram to the process of IPv6 evolution;
Fig. 5 supports IPv4 and IPv6 forwarding strategy stream list data structure
Fig. 6 demonstrates the logical view of application of the present invention;
Fig. 7 demonstrates the ply-yarn drill structure of application of the present invention;
Fig. 8 is the basic procedure of processing data packets;
Fig. 9 has shown the ply-yarn drill that adopts network processing unit;
Figure 10 is the internal structure block diagram of network processing unit;
Figure 11 has shown the bag handling process;
Figure 12 schematic presentation the principle handled of scheduling;
Figure 13 has shown the exchange model that adopts the N port input rank of VOQ;
Figure 14 is the system configuration that adopts the Crossbar switching technology;
Figure 15 is the internal structure of queue management device;
Figure 16 has shown the block diagram of Crossbar Chip;
Figure 17 has showed MSR equipment cabinets front view;
Figure 18 has provided the building-block of logic of application apparatus of the present invention
Figure 19 is the theory diagram of the power rectifier system of band remote monitoring function;
Figure 20 schematically illustrates the structure chart of embedded ethernet system
Figure 21 is the R8002 devices exchange systematic schematic diagram that adopts the double crossing over structure;
Figure 22 has provided the general line card structure figure of application apparatus of the present invention;
Figure 23 has provided POS general line card structure figure;
Figure 24 has provided the software hierarchy structure of application apparatus of the present invention;
Figure 25 is system's adaptation layer schematic diagram of application apparatus of the present invention;
Figure 26 is system's adaptation layer schematic diagram of application apparatus of the present invention;
Figure 27 is the interface mappings graph of a relation;
Figure 28 shows the block diagram of transmitting the data transfer equipment of data with strategic stream mode of the present invention;
Figure 29 is the block diagram of the strategy stream ID computing unit in the Forwarding plane of data transfer equipment of the present invention;
Figure 30 is the modification in the Forwarding plane of data transfer equipment of the present invention and the block diagram of retransmission unit;
Figure 31 is the block diagram of the classification transmission unit among Figure 30;
Figure 32 is the block diagram of the policy processing unit among Figure 28;
Figure 33 is the flow chart of data forwarding method of the present invention;
Figure 34 is when purpose IP is the IPv4 address, and the data flow at this packet place of definite reply is carried out the flow chart of the method for which kind of business operation; And
Figure 35 is when purpose IP is the IPv6 address of IPv4 compatibility, and the data flow at this packet place of definite reply is carried out the flow chart of the method for which kind of business operation.
Embodiment
At first introduce three kinds of technical schemes and four kinds of typical network topology under IPv4 and the IPv6 coexistence situation, introduce the algorithm that a kind of IPv4 that can satisfy above three kinds of technical schemes and four kinds of network topologies and IPv6 transmit then, introduce how to realize this algorithm at last with strategic stream mode.
How IEFT coexists about IPv4 and IPv6 has proposed three kinds of technical schemes: IPv4/IPv6 dual stack mode, tunnel style and the network address/protocol conversion mode.They specify as follows:
The IPv4/IPv6 dual stack mode: the solution of dual stack is safeguarded IPv6 and IPv4 two cover route protocol stacks in fact exactly in a router device, make the main frame in the network can support IPv6 and IPv4 agreement respectively, also can support these two kinds of agreements simultaneously, router can with the IPv4 main frame also can with the IPv6 main-machine communication, support independently IPv6 and IPv4 Routing Protocol respectively, IPv4 and IPv6 routing iinformation calculate according to Routing Protocol separately, safeguard different routing tables.IPv6 datagram (comprising the IPv6 datagram with IPv4 address compatibility) is transmitted according to the routing table that the IPv6 Routing Protocol obtains, and the IPv4 datagram is transmitted according to the routing table that the IPv4 Routing Protocol obtains.
The network address/protocol conversion mode: this scheme is generally used for communicating by letter between pure IPv4 node and the IPv6 node, does not advise adopting this scheme for the IPv4 protocol communication in pure IPv6 node and the dual stack node.Straightforward transform mode has been adopted in address/protocol conversion, and need not revise upper-layer protocol can mutual communication.The central apparatus of this scheme is called the NAT-PT gateway again, can realize the mutual conversion of IPv4 and IPv6 protocol stack, comprises the mutual conversion between network layer protocol, transport layer protocol and some application layer protocols.
Tunnel style: so-called tunneling technique is exactly to utilize the IPv4 agreement of moving in the existing network facility to set up the communication mechanism of IPv6 for carrier, the transmission of datagram is undertaken by IPv4 mechanism between the node at two, tunnel, the tunnel is seen as a direct-connected passage, and tunneling technique is the technology that IPv4 was easy to adopt most to the initial stage of IPv6 transition.The tunnel can be set up between router and the main frame and between main frame and the main frame between router and router, and the tunnel can also can be set up automatically in manual configuration foundation.The thinking of tunnel strategy briefly is exactly in fact, router is encapsulated into IPv4 with the packet of IPv6, the IPv4 address of the source address of IPv4 grouping and destination address corresponding tunnel portal of difference and outlet, the exit in the tunnel is taken out the IPv6 grouping and is transmitted to the purpose website.Tunneling technique is only required at the entrance and exit place in tunnel and is made amendment, and other parts is not required, thereby realize than being easier to.
More than be three kinds of technical schemes that solve IPv4 and IPv6 coexistence, and following this is four kinds of typical network topology that adopt these three kinds of modes.
As Fig. 1, for IPv4 in the process of IPv6 evolution, the backbone node of network has become the IPv6 node, and edge network still is the network of IPv4.At this moment, need be when the IPv4 network is visited mutually among the figure by middle IPv6 network, will adopt the tunnel style of IPv4 in IPv6 this moment, and the terminal node address in tunnel is to be the address, tunnel of manual configuration usually.
As Fig. 2, for IPv4 in the process of IPv6 evolution, the isolated island of two IPv6 need be visited mutually by the network of IPv4.Need to adopt the tunnel style of IPv4 In IPv6 this moment, wherein according to the type difference of purpose IPv6 address, can be divided into automatic configured tunneling technique and manual configuration tunnel again.If promptly purpose IPv6 address is the IPv6 address of IPv4 compatibility, then is automatic configured tunneling technique, otherwise is the manual configuration tunnel.
As Fig. 3, be the final stage of IPv4 in the IPv6 evolution process, this moment, all networks were the IPv6 network, and therefore whole network only need adopt pure IPv6 forwarding to get final product.
As Fig. 4, for IPv4 in the IPv6 process, the directly mutually visit between IPv4 and the IPv6 network.At this moment, must adopt the network address/protocol conversion mode, realize both intercommunications.
Next, at first introduce the algorithm IPFA (Ipv4 and IPv6compatible Forwarding Algorithm or IPFA) of the forwarding of all operable IPv4 that in same physical pipeline, realizes of a kind of IPv4/IPv6 node provided by the invention and IPv6 compatibility.When this algorithm will adopt IPv6 to transmit with deciding the network equipment when to adopt IPv4 to transmit, and when adopt automatic tunnel, when will adopt the manual configuration tunnel, and when adopt address transition mode (NAT-PT).This algorithm mainly is based on following consideration: the network equipment will depend on the purpose IP address class type of packet, the configuration of inbound port and the outbound port (OutputPort) of searching pairing next jumping in this purpose IP address (Next Hop) place to the processing mode of packet.This algorithm can be supported above-mentioned three kinds of modes and four kinds of topologys effectively.
This arthmetic statement is as follows:
1. if purpose IP address is the IPv4 address:
1) judge at first whether whether inbound port has enabled NAT-PT and purpose IPv4 address is address in the address pool, if not would turn to step 2), otherwise turn to step 4);
2) carry out the relevant forwarding information of IPv4 and search and transmit, judge the interface type of next jumping,, then finish the IPv4 forwarding, otherwise carry out step 3) if be IPv4;
3) if next jumps corresponding interface is IPv6, carry out the tunnel operations of IPv4 in IPv6, wherein destination address can also can be the address, tunnel of manual configuration for the IPv6 address of the IPv4 compatibility of this IPv4 address correspondence;
4) carry out the NAT-PT operation, the IPv4 bag is converted to the bag of IPv6;
2. if destination address is the IPv6 address of IPv4 compatibility
1) judge at first whether whether inbound port has enabled NAT-PT and destination address is address in the address pool, if not would turn to step 2), otherwise turn to step 5)
2) carry out the relevant forwarding message of IPv6 and search and transmit, judge the outgoing interface type of next jumping,, then carry out the IPv6 forwarding, otherwise carry out step 3) if be IPv6
3) if the type of next jumping is IPv4, then at first judges port and whether enabled NAT-PT,, otherwise carry out step 4) if the sign of NAT-PT operation is arranged then carry out step 5)
4) carry out IPv6 in IPv4 operation, wherein destination address is low 32 of original IPv6 destination address.
5) carry out the NAT-PT operation, wherein conversion back source address is the address in the nat address pool, and the purpose IP address after the conversion is low 32 of original IPv6 destination address.
3. if destination address is the IPv6 address
1) searches according to the relevant forwarding message table of IPv6 and transmit, and judge that next jumps pairing interface type.If be IPv6, then carry out IPv6 and transmit.
2) otherwise be IPv4, then need to carry out the tunnel operations of IPv6 in IPv4, wherein the destination address of IPv4 is the IPv4 address in the tunnel of configuration.
Next, the present invention realizes IPFA with strategy stream method and apparatus is described.
Strategy circulation is sent out and has been adopted fully the thinking handled based on stream rather than based on the thought of processing data packets, the main practice is: for the different pieces of information bag in the same data flow, the network equipment should be identical to its behavior (Action).In addition, for Forwarding plane, it does not need to know which kind of oneself should adopt search algorithm (clean culture, multicast, IPv4 or Ipv6) to packet, and is opposite, it only need know its outbound port and next hop information, and which kind of strategy of this employing is made amendment to packet.Therefore, the strategy circulation is sent out and adopted unified forwarding entry and the single algorithm (Exact Match) of searching efficiently in Forwarding plane, has so just simplified the processing of Forwarding plane, has improved search efficiency.
Specifically, the exchange of so-called strategy stream, it is not to search exchange at some concrete IP packets, but adopted specific I P data flow is transmitted the thought of processing.The exchange of strategy stream has been adopted according to polynary group of specific next unique definite stream (Stream).At specific Stream, first bag that the network equipment only needs to flow send control plane to handle, and is the unique stream ID of this distribution of flows (Stream ID), gets final product and this stream bag subsequently only needs Forwarding plane simply to transmit operation.Therefore, strategy circulation is sent out and only need be carried out complicated tabling look-up to first packet of this data flow and search, and by the specific transactions operation of customization to it.Control plane finish search and customized type of service after, generation strategy stream forwarding list order (Policy Stream Forwarding Table), and by routing engine with this entry distribution (Distribute) to Forwarding plane.Entry includes the outgoing interface information of next jumping and indicates the policing type field of class of business.And packet will be done unified processing to this Stream according to the policing type of this business operation when transmitting.Because for Forwarding plane, it will be indifferent to the strategy of searching, and be concerned about just whether this data flow can be forwarded and should be forwarded to that outgoing interface, so just the processing procedure of the Forwarding plane of simplification greatly.
The main distinction such as the following table 2 of algorithm and MPLS sent out in strategy circulation of the present invention:
The main distinction of algorithm and MPLS is sent out in the circulation of table 2-strategy
| The strategy circulation is sent out | ||
| Forward efficiency to the Ipv4/v6 bag | LER adopts the longest coupling, low; LSR adopts and accurately searches, height | Employing is accurately searched, height |
| With the interface protocol of other vendor equipment, interconnecting property and interface | Introduce extra complicated signaling (be a whole set of control plane agreement of MPLS, comprise CR-LDP and RSVP two big classes), bringing interconnects becomes complicated, a lot of and signaling of protocol option | Only need the logic flow of the data flow of change device interior, need not increase any agreement, and only need according to unified route table items, |
| Complexity | Problems such as the consistency operation of agreement | The upgrading in house software |
| Can adopt Differentiated Services or RSVP mode, but Differentiated Services can't guarantee the bandwidth of each stream | In individual node, can guarantee bandwidth at each stream, granularity can arrive 64Kbps | |
| Support to use | Can't directly support ACL (Access Control List (ACL)) or NAT application such as (network address translation), when operator proposed to increase other value-added service, needing increased other list item, and the kind of list item increases, and complexity increases | Same list item supported all application (comprising ACL and NAT etc.), is convenient to provide value-added functionality |
Fig. 5 supports IPv4 and IPv6 forwarding strategy stream list data structure.
The Data Structures of the strategy stream table (PSFB) that the present invention adopts:
Source IPv4 Address (4 byte): the IPv4 address, source of data flow.
Destination IPv4 Address (4 byte): the purpose IPv4 address of data flow
Source IPv6 Address (16 byte): the IPv6 address, source of data flow
Destination IPv6 Address:(16 byte): the purpose IPv6 address of data flow
Protocol Type (2 byte): protocol type
Source Protocol Port (2 byte): the source protocol port, decide port into which kind of agreement by protocol type
Destination Protocol Port (2 byte): the purpose protocol port, decide port into which kind of agreement by protocol type
Flow Label (2 byte): the header field among the IPv6, represent the various flows stream in the same source address, this field be for can select field for use, is 0 to be to represent to ignore its use.
Stream ID (4 byte): unique ID of a stream of expression
Alias Port (2 byte): pseudo-port, the protocol port after the conversion is used for network address translation
Alias IPv4 Address (4 byte): pseudo-IP address, the IPv4 address after the conversion is used for network address translation
Alias IPv6 Address:(16 byte): pseudo-IPv6 address, the IPv6 address after the conversion is used for network address translation
Tunnel ID (2 byte): tunnel ID is used for IPv6 to IPv4 or IPv4 to IPv6 tunnel
Policy Type (2 byte): policing type, show which kind of business operation type that this carries out to data flow, can carry out business customizing by webmaster flexible configuration or user.
QoS (2 byte): service quality shows the QoS parameter to this stream.
Expired Timer (1 byte): overtime timer, judge whether this stream forwarding list order is overtime.
TCP Flag (1 byte): the TCP flag bit is used for judging whether TCP stream finishes.
Output Port Index (2 byte): the outbound port index is used for specifying the port that sends out of this packet
Next Hop IPv4 Address (4 byte): next-hop IP v4 address
Next Hop IPv6 Address (16 byte): next-hop IP v6 address
Tunnel end IPv4 Address (4 byte): termination point of a tunnel IPv4 address and pseudo-IPv4 address multiplex.
Tunnel end IPv6 Address (16 byte): termination point of a tunnel IPv6 address and pseudo-IPv6 address multiplex.
Annotate: above data structure only is a Data Structures of the present invention, and the user can expand this data structure according to the application of reality.
The present invention is based on the method and apparatus that strategy stream realizes that IPv4 and the two stack packets of IPv6 are transmitted below with reference to description of drawings.
At first the data transfer equipment of transmitting data with strategic stream mode of the present invention is described with reference to figure 28-32.Figure 28 shows the block diagram of transmitting the data transfer equipment of data with strategic stream mode of the present invention.As shown in figure 28, this data transfer equipment comprises an at least one Forwarding plane and a control plane, and wherein the packet in each Forwarding plane receiving data stream comprises at least one packet in this data flow, this data transfer equipment also comprises: PSFB entry memory cell is used to store the PSFB entry.
Forwarding plane comprises: purpose IP address class type judging unit is used to judge the type of purpose IP address; Strategy stream ID computing unit is used for the type according to different purpose IP, selects the polynary set of properties of corresponding packet, calculates the strategy stream ID of this data flow of sign; Search the unit, be used for accurately searching the PSFB entry that PSFB entry memory cell is stored, see if there is the PSFB entry that the strategy stream ID with this packet is complementary according to strategy stream ID; Revise and retransmission unit,, then this packet is carried out related content and revise and transmit operation according to the PSFB entry if at the PSFB entry of searching unit discovery coupling; As not finding the PSFB entry that mates searching the unit, then this packet is delivered to control plane.
Control plane comprises: policy processing unit, according in the configuration of the outbound port of the configuration of the type of purpose IP address, inbound port and pairing next jumping in IP address of this purpose at least one, packet is handled accordingly; PSFB entry generation unit generates corresponding PSFB entry according to the processing that packet is carried out, and this PSFB entry is distributed to PSFB entry memory cell, for the packet use of back.
The type of the purpose IP address of the handled packet of data transfer equipment of the present invention can be the IPv6 address of IPv4 address, IPv4 compatibility or the IPv6 address of non-IPv4 compatibility.
Figure 29 is the block diagram of the strategy stream ID computing unit in the Forwarding plane of data transfer equipment of the present invention.As shown in figure 29, strategy stream ID computing unit comprises: first calculating section, when being used for being the IPv4 address, calculate strategy stream ID according to the five meta-attribute batch totals of forming by IPv4 address, source, purpose IPv4 address, source protocol port, purpose protocol port, protocol type in purpose IP address; IPv6 flow label Flow_Label detecting unit, when being used in purpose IP address being the IPv6 address of the IPv6 address of IPv4 compatibility or non-IPv4 compatibility, whether the IPv6 flow label Flow_Label in the judgment data packet header is 0; Second calculating section, being used for judging IPv6 flow label Flow_Label is 0 o'clock, calculates strategy stream ID according to the five meta-attribute batch totals of being made up of IPv6 address, source, purpose IPv6 address, source protocol port, purpose protocol port, protocol type; And the 3rd calculating section, being used for judging IPv6 flow label Flow_Label is not 0 o'clock, according to the binary set of properties calculative strategy stream ID that is made up of IPv6 address, source, IPv6 flow label Flow_Label.
Record in the PSFB entry of in PSFB entry memory cell, storing show should to data stream carry out which kind of business operation type policing type, be used for specifying the outbound port index and the next-hop ip address that send out port of this packet.
Figure 30 is the modification in the Forwarding plane of data transfer equipment of the present invention and the block diagram of retransmission unit.As shown in figure 30, revise and retransmission unit comprises: the policing type acquiring unit is used for from obtaining policing type corresponding to this data flow with data packet matched PSFB entry; And classification transmission unit, according to policing type the packet basis is transmitted operation accordingly corresponding to this data flow, draw outbound port and link corresponding layer information according to outbound port index list item in the PSFB entry and next-hop ip address adjacency list list item pointed, this packet is sent from outbound port.
Figure 31 is the block diagram of the classification transmission unit among Figure 30.As shown in figure 31, the classification transmission unit comprises: IPv4 transmits operating unit, when policing type is pure IPv4 when transmitting IPv4_FORWARD, carries out the forwarding operation of IPv4, revises TTL, and again calculation check and; IPv6 transmits operating unit, when policing type is pure IPv6 when transmitting IPv6_FORWARD, carries out the forwarding operation of IPv6, revises TTL; The NAT-PT operating unit when policing type is network address translation NAT_PT, is carried out the NAT-PT operation, according to pseudo-IP address in the PSFB entry and pseudo-port, the content of packet is made amendment; IPv4 to IPv6 automatic tunnel operating unit, when policing type is the automatic tunnel pattern IPv4_IN_IPv6_AUTO of IPv4 in IPv6, carry out the automatic tunnel operation of IPv4 to IPv6, according to the tunnel ID in the PSFB entry (Tunnel ID) encapsulated data packet, the purpose IPv6 address in tunnel is the IPv6 compatible address of purpose IPv4 address correspondence; The manual tunnel operations of IPv4 to IPv6 unit, when policing type is the manual tunnel mode IPv4_IN_IPv6_MANU of IPv4 in IPv6, carry out the manual tunnel operations of IPv4 to IPv6, according to the Tunnel ID encapsulated data packet in the PSFB entry, the purpose IPv6 address in tunnel is the address in the PSFB entry; IPv6 to IPv4 automatic tunnel operating unit, when policing type is the automatic tunnel pattern IPv6_IN_IPv4_AUTO of IPv6 in IPv4, carry out the automatic tunnel operation of IPv6 to IPv4, according to the Tunnel ID encapsulated data packet in the PSFB entry, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility; And the manual tunnel operations of IPv6 to IPv4 unit, when policing type is the manual tunnel mode IPv6_IN_IPv4_MANU of IPv6 in IPv4, carry out the tunnel operations of IPv6 to IPv4, according to the Tunnel ID encapsulated data packet in the PSFB entry, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility.
Figure 32 is the block diagram of the policy processing unit among Figure 28.Shown in figure 32, policy processing unit comprises: the policing type determining unit, and the data flow that is used for this packet place of definite reply is carried out the policing type of which kind of business operation; And policing type information returns the unit, is used for the information of relevant policing type is sent it back Forwarding plane, thereby corresponding to the operating unit of this policing type packet transmitted operation in the classification transmission unit by Forwarding plane.And the policing type determining unit comprises: first judgment part is used to judge whether inbound port has enabled NAT-PT; Second judgment part is used to judge whether purpose IPv4 address is address in the address pool; Forwarding information is searched the unit, is used to search transmit, and judges the interface type of next jumping.
Wherein, when purpose IP address is the IPv4 address, at first judge by first judgment part whether inbound port has enabled NAT-PT, and judge by second judgment part whether purpose IPv4 address is address in the address pool, if at least one judged result is for being, then determine and to carry out the NAT-PT operation at Forwarding plane, the IPv4 bag is converted to the bag of IPv6; If judged result all is not, then searching the unit by forwarding information carries out the relevant forwarding information of IPv4 and searches and transmit, judge the interface type of next jumping, if the interface type of next jumping is IPv4, determine that then should finish IPv4 at Forwarding plane transmits, if next jumps corresponding interface is IPv6, then determine to carry out at Forwarding plane the tunnel operations of IPv4 in IPv6, wherein destination address can also can be the address, tunnel of manual configuration for the IPv6 address of the IPv4 compatibility of this IPv4 address correspondence.
When purpose IP address is the IPv6 address of IPv4 compatibility, at first judge by first judgment part whether inbound port has enabled NAT-PT, and judge by second judgment part whether purpose IPv4 address is address in the address pool, if at least one judged result is for being, then determine and should carry out the NAT-PT operation at Forwarding plane, the purpose IP address after wherein changing is low 32 of original IPv6 destination address; If judged result all is not, then searching the unit by forwarding information carries out the relevant forwarding information of IPv6 and searches and transmit, judge the outgoing interface type of next jumping, if be IPv6, determine that then should carry out IPv6 at Forwarding plane transmits, if the type of next jumping is IPv4, then at first judge port and whether enabled NAT-PT by first judgment part, if have then determine and to carry out the NAT-PT operation at Forwarding plane, wherein source address is the address in the address pool, and destination address is the low 32 of original IPv6 destination address, otherwise determines and should carry out IPv6 in IPv4 operation at Forwarding plane, and wherein destination address is low 32 of original IPv6 destination address.
When purpose IP address is the IPv6 address of non-IPv4 compatibility, searching the unit by forwarding information searches according to the relevant forwarding information of IPv6 and transmits, and judge that next jumps pairing interface type, if be IPv6, determine that then should carry out IPv6 at Forwarding plane transmits; Otherwise be IPv4, then should carry out the tunnel operations of IPv6 in IPv4 at Forwarding plane, wherein the destination address of IPv4 is the IPv4 address in the tunnel of configuration.
PSFB entry generation unit comprises: the first entry generating portion, and when packet being carried out the IPv4 forwarding, policing type is IPv4_FORWARD, the content of the PSFB entry of generation comprises policing type, stream ID, outbound port index, next-hop IP v6 address; The second entry generating portion, when packet being carried out the IPv6 forwarding, policing type is IPv6_FORWARD, the content of the PSFB entry of generation comprises policing type, stream ID, outbound port index, next-hop IP v6 address; The 3rd entry generating portion, when packet being carried out the conversion of the network address/protocol address, policing type is NAT_PT, the content of the PSFB entry that generates comprises policing type, stream ID, Alias IPv6 Address, Source IPv4 Address, Destination IPv4 Address, Source Port, Destination Port, Alias Port, Protocol Type etc., outbound port index and next-hop IP v6 address etc.; The 4th entry generating portion, when packet being carried out automatic IPv4 in IPv6 tunnel operations, policing type is IPv4_IN_IPv6_AUTO, and the content of the PSFB entry of generation comprises policing type, stream ID, outbound port index, next-hop IP v6 address, tunnel ID; The 5th entry generating portion, when packet being carried out the IPv4 in IPv6 tunnel operations of manual configuration, policing type is IPv4_IN_IPv6_MANU, and the content of the PSFB entry of generation comprises the end address in policing type, stream ID, outbound port index, next-hop IP v6 address, tunnel ID and tunnel; The 6th entry generating portion, when packet being carried out the operation of IPv6 in IPv4 automatic tunnel, policing type is IPv6_IN_IPv4_AUTO, the content of the PSFB entry of generation comprises policing type, stream ID, outbound port index and tunnel ID; The 7th entry generating portion, when packet being carried out IPv6 in IPv4 craft tunnel operations, policing type is IPv6_IN_IPv4_MANU, the content of the PSFB entry of generation comprises policing type, stream ID, outbound port index and tunnel ID.
Describe the present invention below with reference to accompanying drawing 33-35 and realize the method that packet is transmitted with strategic stream mode in the network equipment, this network equipment comprises Forwarding plane and control plane at least.
Figure 33 is the flow chart of data forwarding method of the present invention.As shown in figure 33, at step S1, the packet in the receiving data stream comprises at least one packet in this data flow.At step S2, Forwarding plane is judged the type of purpose IP address then.The type of purpose IP address can be the IPv6 address of IPv4 address, IPv4 compatibility or the IPv6 address of non-IPv4 compatibility among the present invention.At step S3, Forwarding plane is selected the polynary set of properties of corresponding packet according to the type of different purpose IP, calculates the strategy stream ID of this data flow of sign.Wherein when purpose IP address was the IPv4 address, polynary set of properties comprised IPv4 address, source, purpose IPv4 address, source protocol port, purpose protocol port, protocol type; When purpose IP address is the IPv6 address of the IPv6 address of IPv4 compatibility or non-IPv4 compatibility, whether the IPv6 flow label Flow_Label in the judgment data packet header is 0, if the IPv6 flow label Flow_Label in the data packet head is 0, then polynary set of properties comprises IPv6 address, source, purpose IPv6 address, source protocol port, purpose protocol port, protocol type, if the IPv6 flow label Flow_Label in the data packet head is not 0, then polynary set of properties comprises IPv6 address, source, IPv6 flow label Flow_Label.
At step S4, Forwarding plane is accurately searched the PSFB entry according to strategy stream ID.If discovery is mated then is entered step S8, according to the PSFB entry this packet is carried out related content at Forwarding plane and revise and transmit and operate, forward step S9 then to and handle next packet; If do not find the PSFB entry that mates at step S4 then illustrate that this is that first bag of data flow or PSFB entry are aging, then forward step S5 to, Forwarding plane will be delivered to control plane to this packet.
The front points out, record in the PSFB entry show should to data stream carry out which kind of business operation type policing type, be used for specifying the outbound port index and the next-hop ip address that send out port of this packet.The operation of step S8 particularly, comprise the policing type of from the PSFB entry, knowing this data flow, and transmit operation accordingly, draw outbound port and link corresponding layer information according to outbound port index list item in the PSFB entry and next-hop ip address adjacency list list item pointed, this packet is sent from outbound port.
More particularly, when purpose IP address is the IPv4 address, transmit IPv4_FORWARD, then carry out the forwarding operation of IPv4, revise TTL if policing type is pure IPv4, and again calculation check and; If policing type is the automatic tunnel pattern IPv4_IN_IPv6_AUTO of IPv4 in IPv6, then carry out the automatic tunnel operation of IPv4 to IPv6, according to the Tunnel ID encapsulated data packet in the PSFB entry, the purpose IPv6 address in tunnel is the IPv6 compatible address of purpose IPv4 address correspondence; If policing type is the manual tunnel mode IPv4_IN_IPv6_MANU of IPv4 in IPv6, then carry out the manual tunnel operations of IPv4 toIPv6, according to the Tunnel ID encapsulated data packet in the PSFB entry, the purpose IPv6 address in tunnel is the address in the PSFB entry; If policing type is network address translation NAT_PT, then carry out the NAT-PT operation, according to pseudo-IP address in the PSFB entry and pseudo-port, the content of packet is made amendment.
When purpose IP address is the IPv6 address of IPv4 compatibility, if being pure IPv6, policing type transmits IPv6_FORWARD, then carry out the forwarding operation of IPv6, revise TTL; If policing type is the automatic tunnel pattern IPv6_IN_IPv4_AUTO of IPv6in IPv4, then carry out the automatic tunnel operation of IPv6 to IPv4, according to the Tunnel ID encapsulated data packet in the PSFB entry, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility; If policing type is network address translation NAT_PT, then carry out the NAT-PT operation, according to pseudo-IP address in the PSFB entry and pseudo-port, the content of packet is made amendment.
When purpose IP address is the IPv6 address of non-IPv4 compatibility, if being pure IPv6, policing type transmits IPv6_FORWARD, then carry out the forwarding operation of IPv6, revise TTL; If policing type is the manual tunnel mode IPv6_IN_IPv4_MANU of IPv6 in IPv4, then carry out the tunnel operations of IPv6 to IPv4, according to the Tunnel ID encapsulated data packet in the PSFB entry, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility.
At step S6, control plane according in the configuration of the outbound port of the configuration of the type of purpose IP address, inbound port and pairing next jumping in IP address of this purpose at least one, come packet is handled accordingly with the Forwarding plane cooperation, and generate corresponding PSFB entry.Determine that particularly the data flow at this packet place of reply carries out the policing type of which kind of business operation, and the information of relevant policing type sent it back Forwarding plane, so as by Forwarding plane according to determined policing type handle packet.Hereinafter the concrete mode which kind of business operation is relevant definite data flow of tackling this packet place carry out and generate the operation of corresponding PSFB entry will be specifically described.
At step S7, control plane is distributed to Forwarding plane with this PSFB entry then, for the packet use of back.Enter step S9 at last, handle next packet.
Describe respectively below when when purpose IP address is the IPv6 address of the IPv6 address of IPv4 address, IPv4 compatibility, non-IPv4 compatibility, the data flow at this packet place of definite reply is carried out the concrete grammar of which kind of business operation.
Figure 34 is when purpose IP is the IPv4 address, and the data flow at this packet place of definite reply is carried out the flow chart of the method for which kind of business operation.As shown in figure 34, when purpose IP address is the IPv4 address, at step S611, judge at first whether inbound port has enabled NAT-PT and whether purpose IPv4 address is address in the address pool, if not then turn to step S612, otherwise turn to step S613, carry out the NAT-PT operation, the IPv4 bag is converted to the bag of IPv6 at Forwarding plane.At step S612, carrying out the relevant forwarding information of IPv4 searches and transmits, judge the interface type of next jumping, if be IPv4, then at step S614, Forwarding plane is finished IPv4 and is transmitted, if next jumps corresponding interface is IPv6, then enter step S615, carry out the tunnel operations of IPv4 in IPv6 at Forwarding plane, wherein destination address can also can be the address, tunnel of manual configuration for the IPv6 address of the IPv4 compatibility of this IPv4 address correspondence.
Figure 35 is when purpose IP is the IPv6 address of IPv4 compatibility, and the data flow at this packet place of definite reply is carried out the flow chart of the method for which kind of business operation.As shown in figure 35, when purpose IP address is the IPv6 address of IPv4 compatibility, at step S621, judge at first whether inbound port has enabled NAT-PT and whether destination address is address in the address pool, if not then turn to step S622, otherwise turn to step S626, carry out the NAT-PT operation at Forwarding plane, the purpose IP address after wherein changing is low 32 of original IPv6 destination address.At step S622, carry out the relevant forwarding information of IPv6 and search and transmit, judge the outgoing interface type of next jumping, if be IPv6, then turn to S624, Forwarding plane carries out the IPv6 forwarding.If the type of next jumping is IPv4, then turn to step S623, at first judge port and whether enabled NAT-PT.If, then turn to step S626, carry out the NAT-PT operation at Forwarding plane, wherein source address is the address in the address pool, and destination address is low 32 of original IPv6 destination address; Otherwise turn to step S625, carry out IPv6 in IPv4 operation at Forwarding plane, wherein destination address is low 32 of original IPv6 destination address.
When purpose IP address is the IPv6 address of non-IPv4 compatibility, search according to the relevant forwarding information of IPv6 and to transmit, and judge that next jumps pairing interface type, if be IPv6, then carry out the IPv6 forwarding at Forwarding plane.Otherwise be IPv4, then need carry out the tunnel operations of IPv6 in IPv4 at Forwarding plane, wherein the destination address of IPv4 is the IPv4 address in the tunnel of configuration.
The following describes according to the different processing that packet carried out, correspondingly generate the concrete mode of PSFB entry.
If when packet was carried out the IPv4 forwarding, policing type was IPv4_FORWARD, the content of the PSFB entry of generation comprises policing type, strategy stream ID, outbound port index, next-hop IP v6 address.
If when packet was carried out the IPv6 forwarding, policing type was IPv6_FORWARD, the content of the PSFB entry of generation comprises policing type, strategy stream ID, outbound port index, next-hop IP v6 address.
If packet is carried out the NAT/PT conversion, then policing type is NAT_PT, the content of the PSFB entry that generates comprises policing type, strategy stream ID, Alias IPv6 Address, Source IPv4Address, Destination IPv4 Address, Source Port, Destination Port, Alias Port, Protocol Type etc., outbound port index and next-hop IP v6 address etc.
If packet is carried out automatic IPv4 in IPv6 tunnel operations, then policing type is IPv4_IN_IPv6_AUTO, and the content of the PSFB entry of generation comprises policing type, strategy stream ID, outbound port index, next-hop IP v6 address, tunnel ID.
If packet is carried out the IPv4 in IPv6 tunnel operations of manual configuration, then policing type is IPv4_IN_IPv6_MANU, and the content of the PSFB entry of generation comprises the end address in policing type, strategy stream ID, outbound port index, next-hop IP v6 address, tunnel ID and tunnel.
If packet is carried out the operation of IPv6 in IPv4 automatic tunnel, then policing type is IPv6_IN_IPv4_AUTO, and the content of the PSFB entry of generation comprises policing type, strategy stream ID, outbound port index and tunnel ID.
If packet is carried out the manual tunnel operations of IPv6 in IPv4, then policing type is IPv6_IN_IPv4_MANU, and the content of the PSFB entry of generation comprises policing type, strategy stream ID, outbound port index and tunnel ID.
In addition, record the overtime sign that shows that this PSFB entry is whether overtime in the PSFB entry, the packet use of being mated whenever this PSFB entry once promptly refreshes this overtime sign, control plane goes to check this sign at regular intervals, if this sign is not refreshed within a certain period of time, illustrate that this PSFB entry is aging, then deletes this PSFB entry.
Concrete processing under the situation of IPv6 address of IPv6 address that purpose IP address is IPv4 address, IPv4 compatibility, non-IPv4 address compatibility is described respectively below.
1. if purpose IP address is the IPv4 address:
(1) after packet entered network node, Forwarding plane is basis＜source IPv4 address, purpose IPv4 address, source protocol port, purpose protocol port, protocol type at first〉calculate corresponding Stream ID.
Wherein, the algorithm that calculates strategy stream ID by five-tuple can have multiple, only needs to realize that tuple and stream ID shine upon and be convenient to hardware one by one or microcode is realized getting final product.Below be wherein a kind of algorithmic descriptions:
Concrete computational methods are as follows:
If the size in all available strategy stream ID spaces is P.
IPv4 address, five-tuple source, purpose IPv4 address, source protocol port, purpose protocol port, protocol type are arranged from the low level to a high position with binary mode according to said sequence, and calculated permutations go out the pairing CRC32 verification of binary combination and.Suppose this verification and be M.Then by M to P delivery (being M delivery P), just can obtain a strategy stream ID L.
(2) Forwarding plane removes to search stream forwarding list (PSFB) according to the stream ID that (1) is drawn according to accurately searching (Exact Match).If discovery is mated then carried out (3), otherwise the not generation or aging of this entry is described then, need send control plane to handle, carry out (9).
(3) if the coupling of discovery, then begin the determination strategy type, if being pure IPv4, policing type transmits IPv4_FORWARD, then turn to (4),, then turn to (5) if be the automatic tunnel pattern IPv4_IN_IPv6_AUTO of IPv4 in IPv6, if the manual tunnel mode IPv4_IN_IPv6_MANU of IPv4 in IPv6, then turn to (6),, then turn to (7) if be network address translation NAT_PT.
(4) carry out the forwarding operation of IPv4, revise TTL, and again calculation check and, turn to (8).
(5) automatic tunnel of carrying out IPv4 to IPv6 is operated, and according to the Tunnel ID encapsulated data packet among the PSFB, the purpose IPv6 address in tunnel is the IPv6 compatible address of purpose IPv4 address correspondence, turns to (8).
(6) the manual tunnel operations of execution IPv4 to IPv6, according to the Tunnel ID encapsulated data packet among the PSFB, the purpose IPv6 address in tunnel is the address in the PSFB entry, turns to (8).
(7) carry out the NAT-PT operation,, the content of packet is made amendment, and continue execution (8) according to pseudo-IP address among the PSFB and pseudo-port.
(8) draw outbound port (Output Port) and link corresponding layer message (as MAC Address) according to Output Port Index list item among the PSFB and next-hop ip address adjacency list (Adjacency Table) list item pointed then, this packet is sent from outbound port, and with overtime flag bit Expired Flag position set again, and turn to (12)
(9) if there is not entry coupling, illustrate that then this is that first bag or the PSFB stream forwarding list order of data flow is aging, Forwarding plane need then be delivered to control plane to first packet of data flow by data interaction message (PEM).
(10) control plane is handled packet according to the first of the Processing Algorithm IPFA of described IPv4 and IPv6 above, and according to different processing modes, generates different entries, and is specific as follows:
If IPv4 transmits, then policing type is IPv4_FORWARD, and the content of the PSFB table of generation comprises stream ID, outbound port index, next-hop IP v4 address;
If NAT-PT, then policing type is NAT_PT, the content of the PSFB table that generates comprises stream ID, Alias IPv6 Address, Source IPv4 Address, Destination IPv4 Address, SourcePort, Destination Port, Alias Port, Protocol Type etc., outbound port index and next-hop IP v6 address etc.;
If automatic IPv4 in IPv6 Tunnel, then policing type is IPv4_IN_IPv6_AUTO, and the entry of generation comprises stream ID, outbound port index, next-hop IP v6 address, tunnel ID.
If the IPv4 in IPv6 Tunnel of manual configuration, then policing type is IPv4_IN_IPv6_MANU, and the entry of generation comprises the end address in stream ID, outbound port index, next-hop IP v6 address, tunnel ID and tunnel.
(11) control plane is distributed to Forwarding plane by stream entry interpolation message with this entry.
(12) end operation is handled next packet.
2. if purpose IP address is the IPv6 address of IPv4 compatibility:
(1) after packet enters network node, Forwarding plane judges at first whether the IPv6 flow label Flow_Label in the IPv6 packet header is 0, and if be 0 basis＜IPv6 address, source, purpose IPv6 address, source protocol port, purpose protocol port, protocol type calculate corresponding Stream ID; Otherwise, then basis＜source IPv6 address, IPv6 flow label Flow_Label〉and calculate corresponding Stream ID.
Wherein, the algorithm that calculates strategy stream ID by tuple can have multiple, only needs to realize that tuple and stream ID shine upon and be convenient to hardware one by one or microcode is realized getting final product.Below be wherein a kind of algorithmic descriptions:
Concrete computational methods are as follows:
If the size in all available strategy stream ID spaces is P.
IPv6 address, five-tuple source, purpose IPv6 address, source protocol port, purpose protocol port, protocol type or IPv6 address, two tuple sources, IPv6 flow label Flow_Label are arranged with binary mode from the low level to a high position according to said sequence, and the verification of the pairing CRC32 of binary combination that goes out of calculated permutations and.Suppose this verification and be M.Then by M to P delivery (being M delivery P), just can obtain a strategy stream ID L.
(2) Forwarding plane removes to search stream forwarding list (PSFB) according to the stream ID that (1) is drawn according to accurately searching (Exact Match).If discovery is mated then carried out (3), otherwise the not generation or aging of this entry is described then, need send control plane to handle, carry out (8).
(3) if the coupling of discovery, then begin the determination strategy type, if being pure IPv6, policing type transmits IPv6_FORWARD, then turn to (4), if be the automatic tunnel pattern IPv6_IN_IPv4_AUTO of IPv6 in IPv4, then turn to (5),, then turn to (6) if be network address translation NAT_PT.
(4) forwarding of carrying out IPv6 is operated, and revises TTL, turns to (7).
(5) automatic tunnel of carrying out IPv6 to IPv4 is operated, and according to the Tunnel ID encapsulated data packet among the PSFB, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility, turns to (7).
(6) carry out the NAT-PT operation,, the content of packet is made amendment, continue (7) according to pseudo-IP address among the PSFB and pseudo-port.
(7) draw outbound port (Output Port) and link corresponding layer message (as MAC Address) according to Output Port Index list item among the PSFB and next-hop ip address adjacency list (Adjacency Table) list item pointed then, this packet is sent from outbound port, and with overtime flag bit Expired Flag position set again, and turn to (11)
(8) if there is not entry coupling, illustrate that then this is that first bag or the PSFB entry of data flow is aging, Forwarding plane need then be delivered to control plane to first packet of data flow by data interaction message (PEM).
(9) control plane is handled packet according to the second portion of the Processing Algorithm IPFA of described IPv4 and IPv6 above, and according to different processing modes, generates different entries, and is specific as follows:
If IPv6 transmits, then policing type is IPv6_FORWARD, and the content of the PSFB table of generation comprises stream ID, outbound port index, next-hop IP v6 address;
If NAT-PT, then policing type is NAT_PT, the content of the PSFB table that generates comprises stream ID, Alias IPv6 Address, Source IPv4 Address, Destination IPv4 Address, SourcePort, Destination Port, Alias Port, Protocol Type etc., outbound port index and next-hop IP v6 address etc.;
If IPv6 in IPv4 is automatic Tunnel, then policing type is IPv6_IN_IPv4_AUTO, and the entry of generation comprises stream ID, outbound port index and tunnel ID.
(10) control plane is distributed to Forwarding plane by stream entry interpolation message with this entry.
(11) end operation is handled next packet.
3. if the IPv6 address of the non-IPv4 in purpose IP address address compatibility:
(1) after packet enters network node, Forwarding plane judges at first whether the IPv6 flow label Flow_Label in the packet header is 0, and if be 0 basis＜IPv6 address, source, purpose IPv6 address, source protocol port, purpose protocol port, protocol type calculate corresponding Stream ID; Otherwise, then basis＜source IPv6 address, IPv6 flow label Flow_Label〉and calculate corresponding Stream ID.
Wherein, the algorithm that calculates strategy stream ID by tuple can have multiple, only needs to realize that tuple and stream ID shine upon and be convenient to hardware one by one or microcode is realized getting final product.Below be wherein a kind of algorithmic descriptions:
Concrete computational methods are as follows:
If the size in all available strategy stream ID spaces is P.
With IPv6 address, five-tuple source, purpose IPv6 address, source protocol port, purpose protocol port, protocol type or IPv6 address, two tuple sources, IPv6 flow label Flow_Label according to from the low level to a high position, arranging, and calculate the pairing CRC32 of the binary combination of arranging out verification and.If this verification and be M.Then by M to the P delivery, just can obtain a strategy stream ID L.
(2) Forwarding plane removes to search stream forwarding list (PSFB) according to the stream ID that (1) is drawn according to accurately searching (Exact Match).If discovery is mated then carried out (3), otherwise the not generation or aging of this entry is described then, need send control plane to handle, carry out (7).
(3) if the coupling of discovery then begins the determination strategy type, if being pure IPv6, policing type transmits IPv6_FORWARD, then turn to (4), if be the manual tunnel mode IPv6_IN_IPv4_MANU of IPv6 in IPv4, then turn to (5).
(4) forwarding of carrying out IPv6 is operated, and revises TTL, turns to (6).
(5) tunnel operations of execution IPv6 to IPv4, according to the Tunnel ID encapsulated data packet among the PSFB, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility, continues (6).
(6) draw outbound port (Output Port) and link corresponding layer message (as MAC Address) according to Output Port Index list item among the PSFB and next-hop ip address adjacency list (Adjacency Table) list item pointed then, this packet is sent from outbound port, and with overtime flag bit Expired Flag position set again, and turn to (10)
(7) if there is not entry coupling, illustrate that then this is that first bag or the PSFB entry of data flow is aging, Forwarding plane need then be delivered to control plane to first packet of data flow by data interaction message (PEM).
(8) control plane is handled packet according to the third part of the Processing Algorithm IPFA of described IPv4 and IPv6 above, and according to different processing modes, generates different entries, and is specific as follows:
If IPv6 transmits, then policing type is IPv6_FORWARD, and the content of the PSFB table of generation comprises stream ID, outbound port index, next-hop IP v6 address;
If IPv6 in IPv4 is manual Tunnel, then policing type is IPv6_IN_IPv4_MANU, and the entry of generation comprises stream ID, outbound port index and tunnel ID.
(9) control plane is distributed to Forwarding plane by stream entry interpolation message with this entry.
(10) end operation is handled next packet.
About the interpolation of entry in the strategy stream, deletion, aging and safeguard, and the mode that control plane communicate by letter with Forwarding plane is with reference to patent " with the method for strategic stream mode raising routing table lookup speed " among the present invention.
The application example of the said equipment of the present invention is described below, i.e. concrete application on the R8002 of Wuhan Firberhome Networks company.
The R8002 of Wuhan Firberhome Networks company development is positioned the IPv4 of metropolitan area network convergence-level and trunk layer and the route exchange device of IPv6 compatibility, and it is supported the multiple interfaces kind and has business generation ability flexibly.When R8002 is positioned the network equipment of metropolitan area network convergence-level, it is mainly finished the tandem of Access Layer upper level link in the metropolitan area network (Metro Aggregation), can insert signals such as Fast Ethernet, Gigabit Ethernet and low speed ATM at user side, and provide IN service to generate (Service Creation) function, for operator provides various value-added functionalitys, and link to each other with metropolitan area network trunk layer equipment by GE or POS in network layer.In addition, R8002 also can be connected with the local ring of SDH by pos interface, perhaps forms annular or star network by GE, forms the trunk of metropolitan area network, and links to each other with backbone network equipment by OC-48POS.
From the demand of networking, need to realize agreement and inter-domain protocols such as BGP-4, BGP4+ in RIP, RIPv6 and the territories such as OSPF, OSPFv6 on the R8002, support agreements such as PPP, Ethernet, LAPS and HDLC at link layer.From application point of view, R8002 can provide and realize that clean culture, multicast and MPLS transmit, and application such as NAT, Firewall, VPN, Virtual Router and mobile IP are provided.In addition, consider the networking mode of present domestic Access Layer, last this of R8002 can provide the support of two layers of application (VLAN).As the value-added functionality that offers operator, R8002 can provide at present based on the bandwidth constraints of port and PPPoE Session and QoS and guarantee.From concerning user's management view, R8002 can provide the authentication mode based on PPPoE at present, and can realize charging to customer flow by Radius.In addition, bind and web authentication for three grades that go back supported vlans+IP+MAC.
The frame of R8002 adopts 19 inches cabinets of industrial standard, dish column pitch 25.4mm, and 16 groove positions altogether, wherein master cpu and swapdisk take No. 7 and No. 8 groove positions, are the backup of 1+1, offer the ply-yarn drill use and remain 14 groove positions, and ply-yarn drill is 9U.
Fig. 6 demonstrates the logical view of application of the present invention.Among the figure, black arrow is represented high speed data bus, and green arrow is represented the High-speed Control bus.Wherein, the large-capacity data passage that data/address bus at a high speed provides, and the passage of administrative messag is provided in the control bus, and signals such as Health#, the Present# of monitoring hardware state and Alarm# are provided.Whole system adopts 3: 1 fan backup and 1: 1 power supply backup, and redundant support of high availability of hardware is provided.
In addition, R8002 has adopted control and has transmitted the architecture of separating, wherein control and management function operate on the master cpu, moved above it fault detection module (FDM), fault-tolerant module (FTM) and and each protocol-dependent agreement correlation unit (PSE), realize fault detect and recovery between primary and backup.And the architectural framework that ply-yarn drill has adopted processor Network Based to transmit, as Fig. 7, core at each ply-yarn drill is a high performance network processing unit, it safeguards the forwarding message library (FIB) that the overall situation is arranged in SRAM and SDRAM, therefore in active and standby handoff procedure, as long as there is flow to enter, still can transmit normally.
When adopting the strategy circulation to send out, first packet of each data flow will be delivered on the risc processor of interface card by network processing unit, finish searching and other business operation (as mobile IP) of routing table by it.Then, it will generate a PSFB entry, and it is downloaded among SRAM and the SDRAM.Owing to adopted accurately and searched, and the mode of two-stage storage has been adopted in the storage of the entry of PSFB, the index that at first on SRAM, has corresponding each Stream ID, and by the actual entry among this index point SDRAM, and the index of this list item has adopted the algorithm of Hash.Under the present circumstances, every the stream forwarding list order is 128Bytes, and on R8002, each interface card is supported 64K Stream, therefore the shared SDRAM space of PSFB is the 128*64K=8M space, and the space that the present invention uses is the 8MBytes space that begins from SDRAM initial address 0x3ff0000.And SRAM be owing to only need deposit correlation table purpose index, therefore only needs the initial address among the SRAM that 64K*4=256KBytes space, the present invention use to be 0xc120000.And because each stream all is unidirectional, so each stream exists only in each ply-yarn drill, so whole system can be supported 64K*14=896K Stream.If considering on average has 20 Stream when each user surfs the Net, then R8002 can support 45K altogether, and promptly 45,000 users are simultaneously online.
Below be in the typical R 8002 to the handling process of IPv6 packet:
(1) after packet enters R8002 by physical chip and framer, microcode in the network processing unit judges at first whether the IPv6 flow label Flow_Label in the packet header is 0, and if be 0 basis＜IPv6 address, source, purpose IPv6 address, source protocol port, purpose protocol port, protocol type calculate corresponding Stream ID; Otherwise, then basis＜source IPv6 address, IPv6 flow label Flow_Label〉and calculate corresponding Stream ID.
(2) microcode removes to search stream forwarding list (PSFB) according to the stream ID that (1) is drawn according to accurately searching (Exact Match).If discovery is mated then is carried out (3), otherwise carries out (5).
(3) if the coupling of discovery then begins the determination strategy type,, then turn to (4),, then turn to (5) if be IPv6_IN_IPv4_MANU if policing type is IPv6_FORWARD.
(4) microcode is carried out the forwarding operation of IPv6, revises TTL, turns to (6).
(5) microcode is carried out the tunnel operations of IPv6 to IPv4, and according to the Tunnel ID encapsulated data packet among the PSFB, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility, turns to (6).
(6) microcode draws outbound port (Output Port) and link corresponding layer message (as MAC Address) according to Output Port Index list item among the PSFB and next-hop ip address adjacency list (Adjacency Table) list item pointed, this packet is sent from outbound port, and with the set again of Expired Flag position, and turn to (10)
(7) if there is not entry coupling, illustrate that then this is that first bag or the PSFB entry of data flow is aging, microcode needs the inner high speed bus first packet of data flow to be delivered on the RISC of master control RSP to handle.
(8) control plane that operates on the risc processor is handled packet according to the third part of the Processing Algorithm IPFA of IPv4 and IPv6, and according to different processing modes, generates different entries, and is specific as follows:
If IPv6 transmits, then policing type is IPv6_FORWARD, and the content of PSFB table comprises stream ID, outbound port index, next-hop IP v6 address;
If IPv6 in IPv4 is manual Tunnel, then policing type is IPv6_IN_IPv4_MANU, and entry comprises stream ID, outbound port index and tunnel ID.
(9) control plane is distributed to Forwarding plane by high-speed internal bus with this entry.
(10) end operation is handled next packet.
In a word, owing to adopted based on the IPv4 of strategy stream and the forwarding algorithm of IPv6 compatibility, the R8002 of Firberhome Networks company has following characteristics:
1. be supported in IPv4 in the same physical pipeline and the forwarding of IPv6, make IPv4 become possibility to seamlessly transitting of IPv6.
2. different with the MPLS technology, do not changing the device hardware structure and do not introducing under the prerequisite of signaling between other networks, just, improved forwarding performance, and supported multiple application by changing the logic flow of device interior.
3. transmit differently based on bag IP with traditional, adopted based on stream and handled rather than, only need do a complex process to a data flow and get final product, avoided the longest matched and searched of complexity, improved forwarding performance greatly based on the thought of each processing data packets.
4. adopt same tactful stream forwarding list to support the multiple mode of IPv4 and IPv6 coexistence simultaneously, avoided adopting the complex data structures of many tables.
5. adopt polynary group to identify an IP operation stream, adopt unified tactful stream forwarding list PSFB at Forwarding plane, and provide abundant extendible list item, can decide processing according to the business operation type to packet, can be used for increasing various needs of users uses, and the interface that provides custom service for the user, make customized service become possibility, and make operator to provide different business and service quality according to each stream of user.
One, the hardware systems design of R8002
Wuhan Firberhome Networks company is the exchange of multi-service light and the transmission platform of a new generation based on the metropolitan area network Multiple Service Ring R8002 equipment of RPR, it is simplified exchange and transmission, also carried out organic integrated simultaneously to exchanging and transmitting these two technology, make it into as a whole, X.87 R8002 equipment adopt up-to-date ITU-T suggestion, and processor Network Based, big capacity Switch-Fabric, the high-performance network protocol processing engine, the FPGA of independent intellectual property right and ASIC, advanced technologies such as optical wavelength-division multiplex, on same platform, realized the fusion of multiple business, improved the bandwidth efficiency in metropolitan area and the Local Area Network, greatly reduced the telecom operation cost, can obtain the hugest business revenue potentiality, make operator carry out low cost to its infrastructure with the network architecture flexibility of maximum, upgrading expeditiously is for telecom operators provide abundant service combination.
R8002 equipment can provide Ethernet, gigabit Ethernet, DVB, ATM, POS, branch service X.85 and X.86, can also work in the mode of dynamic data grouping ring simultaneously, transmit on ring as router and to comprise the Packet Service of IP in wrapping in, the business of moving on ring can provide clean culture, multicast and broadcast mode.Adopt the networking mode of R8002 that network charges and maintenance workload are significantly reduced.
1. the distributed forwarding of processor Network Based (NP)
Network processing unit is to carry out packet filtering under a kind of Gbps of aiming at or the Tbps speed and transmit the programmable specialized hardware ASIC that handles and optimize.Forwarding engine can be classified to agreement and analyzes with linear speed, has reduced the complexity of system, has improved the performance of system greatly.The Premium Features that network processing unit is finished comprise: branch group categories, differentiation priority, adjustment, traffic shaping, conversion, service dispatching, encrypt/decrypt, forwarding, ordering and route, the network interconnection, statistics collection information etc.Network processing unit is also supported DiffServ service, QOS, MPLS etc.For the function of lower level, network processing unit can be analyzed and classifies packet header and load; In the table of predefined, do content-based searching; Grouping is assigned to suitable output port or formation; Adjust packet domain with supported vlans and QOS.
The realization of forwarding engine can have dual mode: a kind of is that forwarding engine and line interface separate, interconnected by exchange mechanism, the advantage of this design is to make full use of forwarding engine, improve the utilization ratio of forwarding engine, shortcoming is load and the load balancing algorithm more complicated that has increased switching fabric; Another kind of mode is that every line interface unit is equipped with a forwarding engine, and the benefit of making is to have reduced the load of switching fabric and do not needed balancing dynamic load like this.Therefore we adopt second kind of design in actual design, promptly based on the distributed packet forward of NP.
In order to improve the R8002 equipment efficient of tabling look-up, two key technologies have been adopted: distributed routing table and hardware searching.All have the part of a and overall routing table same item to transmit in each forwarding engine, forwarding engine is independently searched this locality and is transmitted the transmit path of determining grouping.For realize the part transmit with overall routing table synchronously and some management functions, universal cpu and local storage in addition in the forwarding engine.
Firberhome Networks's programmable network processing unit of R8002 device height provides IN service generative capacity and abundant safety function flexibly, can realize aaa functionality, address transition (NAT) and distribution function, service quality guarantee QOS and bandwidth controlled function, PBR, WEB authentication, PPPoE, based on the security mechanism of ACL, virtual private network function or the like, and can provide Diffsev service for the user according to different strategy stream.
NP is the core part of ply-yarn drill, it at first receives packet from the MAC layer, pass through resolver then, take out the key information in the packet header, and in transmitting, find relevant forwarding entry according to special algorithm (as hash algorithm) according to this key, rewrite packet header, recast verification by this entry again, priority is set, transmits packet at last to exchcange core.
Fig. 8 is the basic procedure of processing data packets.At first forwarding engine parses the packet rs destination MAC Address, and judges whether to carry out the 3rd layer of exchange thus; If target MAC (Media Access Control) address is not the pairing port mac of forwarding engine address, then explanation need be carried out the exchange processing (bridge joint) of the second layer, and at this moment, purpose MAC target is exactly the final purpose address of packet.If target MAC (Media Access Control) address is the MAC Address of this port, judge then whether its purpose IP address is the IP address of router, if think that then this is a management packets (as the communications packets between the master-slave cpu), and it is mail to host CPU by exchcange core.If the IP address is not the IP address of router, so just use special algorithm (as hash algorithm) in SDRAM, to find out the MAC Address of this IP address correspondence, and in transmitting, find corresponding output port, source MAC, destination-mac address and TTL counter in revising bag, recomputate IP packet header verification and after, be sent to exchcange core.If there is not this IP address in transmitting, then search its overall routing table by host CPU, if find this address, then finish subsequent treatment, and will will transmit renewal according to new routing iinformation by host CPU.If host CPU can not find this address, then it is passed to default router or it is abandoned.(in edge router,, will pass to default router to it if host CPU still can not find the entry of IP address correspondence, and in core router, owing in the host CPU the overall routing table of whole network is arranged, if still can not find, to think that the IP address is wrong, and it will be abandoned.)
NP in the R8002 equipment is the integrated forwarding engine able to programme of height that adopts the CMOS technology, supports the line interface of 2.5Gbps, can be upgraded to the interface rate of 10Gbps later on.NP is connected with MAC/Framer by the data/address bus of high bandwidth, can form the ply-yarn drill of multiple different interface type flexibly, and for example FE, GE, POS, Ethernet over SDH, R8002 etc. can satisfy application such as Optical Transmission Network OTN, WAN, MAN, 3G.The typical case of NP on ply-yarn drill uses as Fig. 9:
NP is based on programmable micro engine structure, adopted pipelining, the function of forwarding engine (Forwarding Engine) can be divided into five too partly: bag is resolved, key is searched, revise in packet header, bag is lined up and dispatch, and adopts independent engine to realize this five functional.In the process that bag is handled, these engines will carry out individual processing to bag according to fixing order.The process that this each packet must pass through is wrapped streamline exactly.Owing to adopted pipelining, each engine just can send to this bag the next engine on the streamline, and can begin to handle new packet after finishing the task of oneself.This just means that each engine does not need to wait for that NP finishes the complete process to a bag, just can handle next new packet, thereby improve execution efficient greatly.Figure 10 is the internal structure block diagram of network processing unit.
Two inner streamlines are arranged among the NP, are respectively Incoming bag streamline and Outcoming bag streamline.The Incoming streamline is that each bag from the MAC layer to exchcange core must travel through.The Outcoming streamline then is the reverse overall process from exchcange core to the MAC layer.To wrap streamline with Incoming below is example, and the workflow of NP is described.
Go into the streamline workflow:
When a bag enters NP and sends to the process of exchcange core from Incoming Stream Interface, it must be by permanent order through following engine: the inlet flow scheduler, receives stream analyzer, searches and upgrades engine, receives editing machine and input rank manager, as Figure 11:
Wherein, three engines are programmable AFP, and they realize function by carrying out the microcode instruction of storing on the engine.These three engines are: receive resolver, search and upgrade engine, receive editing machine.Each engine can be downloaded microcode by the cpu i/f of NP, thereby can be according to later needs expanded function.And two other engine that can not programme is the state machine of highly integrated special use.So just when having flexibility, guaranteed the high efficiency of carrying out again.
It is as follows to go into the streamline functional description:
Scheduling
When a bag arrives Input Stream Interface, scheduler will store it corresponding to one among 32 inner FIFO of 32 Stream into.Then, scheduler will be adjudicated data stream.Scheduler at first will be weighted each physical port, dispatches based on wheel commentaries on classics method (Round Robin) then.So-called wheel commentaries on classics method allows the time of the stand-by period of each process in ready queue and enjoyment service be directly proportional exactly.The basic conception of wheel commentaries on classics method is the time with CPU to be divided into the regular time sheet.If process has been used up system's official hour sheet after choosing being scheduled, but does not finish the work yet, it discharges the cpu resource that occupy with itself and is discharged to the end of ready queue, waits for scheduling next time.Process dispatcher goes to dispatch first process or the operation in the current ready queue again simultaneously.Its principle is as shown in figure 12:
In addition, scheduler also will be with the bag segmentation.Wherein, first section is called SOP (Start of Packet).In the SOP section, contain the complete information of whole bag segmentation situation, and comprise whole key information.Therefore the next engine of streamline (search and upgrade engine) can only need the pack processing correctly just by treatment S OP.
Resolve
Resolver mainly is parsing and an abstraction function of carrying out bag.Owing to included the key information of whole bags among the SOP, resolver only needs by just can finish the resolving of bag to the processing of SOP.Like this, the efficient of bag processing just improves greatly.The core of NP is resolver and Cha Zhao ﹠amp; Upgrade engine, they have realized the major function of forwarding engine.The major function of resolver is to provide the key of searching to searching and upgrading engine.Because it is programmable AFP, be IP destination address, MAC Address or UCP/TCP port numbers so can decide key by the microcode of downloading from CPU, or even high layer information more.After obtaining these information, and after adding some out of Memory (as the MPLS label) as required, resolver will generate transmits head (a Forwarding header), and it is sent to searches and upgrade engine.
Search and upgrade
Search and upgrade the core of engine, it search algorithm and carry out the Packet Forwarding Rate that efficient has determined forwarding engine as forwarding engine.Search engine and at first from analytics engine, to transmit head, carry out then and search algorithm, and find required entry in the transmitting among the SDRAM externally.Find the relevant entry of key if search engine, it will mail to the reception editing engine to this entry and be for further processing.
Owing to transmit the storage mode that is based on Hash among the SDRAM externally, therefore, in search procedure, search the Hash that engine at first is treated to key a uniqueness, and search in transmitting according to longest prefix match with this Hash.If these clauses and subclauses exist, search engine and will be sent to editing engine to its Data payload territory.
On structure, the engine of searching of NP is partly formed by two, and by the MS master-slave mode operation.Leading engine partly is based on the AFP of microcode, and partly is a series of finite state machine from engine, and wherein each state machine is carried out some specific search operations.Because leading engine is based on microcode, thus suitable flexibility had, and can be expanded at any time as required.Fixing then can finish fixing search operation fast from engine.
Search engine and support the generation and the interpolation of DML (Data Manipulation Layer) head equally, can just carry out traffic classification, classification of service and queuing easily like this.
In addition, according to number of clicks and aging counting, upgrade engine and also will the forwarding-table item among the SDRAM be upgraded.Like this, make the list item of outside in transmitting be consistent with the routing table of the overall situation all the time, and higher search efficiency is arranged.
Editor
Editing engine obtains relevant information from search engine, and in view of the above the domain of dependence of packet is made amendment.At first, it will subtract one with TTL (Time to live) territory, and regenerate IP bag verification and, finish to the verification of TCP/UDP packet header and renewal and the source and destination MAC Address rewritten, wherein, target MAC (Media Access Control) address is updated to the MAC Address of transmitting next Hop in the entry, and source MAC is updated to the MAC Address of the pairing router port of forwarding engine.In addition, editing machine also will insert or delete the VLAN sign, and produces clean culture or multicast head.
Queue management
The input rank manager is the final step of Incoming streamline.The function that input manager is finished receives bag from editing engine exactly, with their formationizations, and mails to exchcange core according to scheduling strategy.Input manager is managed the packet of input by input bag memory interface and these two external interfaces of Fabric interface.Wherein, input bag memory interface is a sdram interface, and it can support 128Mbytes and 264 logic input ranks at most.
When the formation device receives a packet, at first judge according to the local reception queue heads (LQRHeader) before the packet header.In the LQR head, there is the port id of Crossbar, priority ID and multicast information.It is by resolver, search after engine and the editing engine acting in conjunction generate.After judging, the formation device will store bag in 264 input ranks one.As long as 264 formations have formation not empty, the formation device is dispatched formation according to wheel commentaries on classics method algorithm, repeatedly with in the formation etc. bag to be sent mail to exchcange core.Like this, just finished the forwarding from NP to the exchcange core.
2. big capacity exchcange core Switch-Fabric
Switching fabric is the key component in the R8002 equipment, is to solve the key node that high-speed message is transmitted, and its performance has directly determined machine performance.The switching fabric of R8002 equipment adopts the mode of cross bar switch to realize.Cross bar switch can reach very high speed, and favorable expandability is easy to the expansion and the upgrading of system, but needs the perfect dispatching algorithm of design and realize scheduler with high-speed hardware.Along with people's going deep into to the cross bar switch Study of Scheduling, designed and Implemented many functional, realize simple dispatching algorithm, therefore, we have adopted high performance cross bar switch structure in return in R8002 equipment, below we will discuss the design and the realization of cross bar switch.
Use cross bar switch mainly based on following consideration: at first, cross bar switch can be set up the connection of point-to-point between network interface unit, and this just means can the high-speed transfer data between the network interface unit.At present, in commercial product, chip can reach 3.6Gb/s to the line speed of chip, can reach 4～10Gb/s in the laboratory.Secondly, cross bar switch can provide multiplex simultaneously, as long as source and destination are not conflicted, just can transmit simultaneously, can greatly increase bandwidth like this.
When using cross bar switch, need become fixed length packet to message.When using fixed length packet, can handle length by length according to timeslice according to the big or small time division sheet of grouping.If use elongated grouping, the time of dividing into groups so by cross bar switch is exactly at random, and scheduler just must be known the state of all input and output, and this makes the design of scheduler quite complicated, and is difficult to accomplish equity dispatching.
When using cross bar switch, need to solve following subject matter.
When using cross bar switch in return during structure, may run into following 3 kinds of obstructions.
The 1st kind is that head of line is blocked.If input port is waited for the grouping of exchange and is all used same formation queuing, the head of line obstructing problem will occur.For instance, if the port that the grouping of formation head will be gone is just busy, this grouping can only be waited in formation so.Even the port that the grouping of its back at this moment will be gone is idle, also have no chance to send.This obstruction can greatly reduce the flow of cross bar switch.Adopt the mechanism of VOQ (virtual output queuing is called for short VOQ) can solve the head of line obstructing problem.Thinking is: n input port and n output port are arranged in the supposing the system, and each input port all has respectively corresponding each output port of n formation so.Like this, the output port that takes place to block other as if an output port can not exert an influence.Certainly, when reality was exported, therefore each port was referred to as VOQ only corresponding to a line in the cross bar switch.
The 2nd kind is that input is blocked.Because the line of the cross bar switch of VOQ correspondence has only 1, therefore, can only exchange a grouping at every turn.When a plurality of grouping was arranged in the VOQ, the formation header packet that can not get exchange opportunity just was in the input blocked state.The flow that does not influence cross bar switch is blocked in input, only can increase the delay of the grouping that gets clogged.
The 3rd kind is that output is blocked.If the grouping of two input ports all will be removed same output port, output will take place block.This obstruction is the same with the input obstruction, can not influence the flow of cross bar switch, only increases the delay of the grouping that gets clogged.Input is blocked and output is blocked in order to solve, and has proposed two kinds of solutions: the first, and the VOQ of band priority.Divide 4 priority for each output queue, corresponding output queue also just becomes 4, and the packet priority of high priority sends.This scheme can not solve the input obstructing problem fully, and such as between the identical grouping of priority obstruction being arranged still, but it can guarantee that the retardation ratio of grouping of high priority is less; The second, the exchange velocity (Speedup) of quickening cross bar switch.If the exchange velocity of cross bar switch is the twice of port speed, so, with respect to port, cross bar switch can once exchange two groupings.In theory, if n input port and n output port are arranged, the speed of cross bar switch must be that the n of port speed doubly could guarantee not export obstruction so.In actual use, as long as hand over and the speed of switch is the twice of port speed, just can guarantee not occur output substantially and block.
Another major issue of cross bar switch is exactly a dispatching algorithm, and dispatching algorithm can be divided into the dispatching algorithm of input queue and the dispatching algorithm of output work queue.The basic demand of dispatching algorithm design is:
(1) efficient height.High efficiency dispatching algorithm should be able to be mated input rank as much as possible simultaneously.In general, use hardware to be difficult to calculate optimum Match fast, therefore, generally when the design dispatching algorithm, always seek the algorithm of suboptimum as possible.
(2) stability.The what state of input rank no matter, dispatching algorithm all should find feasible scheduling rapidly.
(3) some formation can not occur and can not get situation about responding forever.
(4) quick.Dispatching algorithm must be carried out fast, otherwise will offset the increase of the bandwidth that cross bar switch brings.
(5) be easy to realize.Dispatching algorithm should be easy to realize with hardware.The complexity that realizes comprises the amount of state that scheduler is safeguarded, the communication overhead the when complexity of the logic that makes decisions based on these states and the state of modification.
The R8002 of Firberhome Networks equipment adopts the high-capacity intelligent switching matrix (Crossbar) of 80Gbps, a high performance route, switching system have been constructed jointly in conjunction with the queue management engine, support perfect QOS algorithm able to programme and clean culture and multicast service, the service priority classification is provided, adopts the performance of VOQ formation and Speedup factor optimizing switching fabric.Switching system is supported function of redundancy protection, supports the hot plug operations of port, adopts a plurality of crossbar fabrics dynode system exchange capacity easily, is convenient to the upgrading of system.
Figure 13 has shown the exchange model that adopts the N port input rank of VOQ.
Sensu lato switching fabric is made of three parts: the switch type backboard of the queue management device on the interface card, high bandwidth, Crossbar exchcange core.Figure 14 is the system configuration that adopts the Crossbar switching technology:
Queue management device on the ply-yarn drill is mainly finished distribution, the serial/parallel translation function of scheduling, multicast and unicast bandwidth of management, the service priority of VOQ formation, the parallel data that it sends network processing unit here encapsulates according to certain data format, then parallel data stream is converted to serial data stream at a high speed, be sent to the exchange ply-yarn drill by backboard.Otherwise it receives from the serial data of exchange ply-yarn drill, carries out the data decapsulation according to the indication of overhead-bits, then serial data at a high speed is converted to the parallel data of low speed.Figure 15 is the internal structure of queue management device.
The switch type backboard is that interface card sticks into the physical channel that row is connected with switched line.The exchange chip group is determined the availability of data channel on the backboard by the clock synchronization process, and the signal quality of sense channel at any time.Connection from ply-yarn drill to central switching fabric is the connection of simple point-to-point, this means can adopt at a high speed to connect.Each circuit uses independent transceiver, thereby can control signal reflex, thereby allows signal that arranged shorter settling time.Being connected with of short point-to-point is beneficial to control clock Skew, signal integrity and minimizing electromagnetic interference.And, in the high speed serialization circuit, extensively adopt differential technique at present, both improved the antijamming capability of circuit when high-speed transfer, lower again power consumption has reduced EMI/EMC greatly.
The Crossbar exchcange core is the key component (seeing Figure 16) among the whole Switch-Fabric.Exchange ply-yarn drill in the R8002 equipment adopts the synchronous serial packet-switch technology from route.At first, receiving on the ply-yarn drill, to be retimed from the data that exchange chip is sent, this has eliminated the variation of any duty ratio or the influence of data dependence, exchange chip must provide the ply-yarn drill that clock desynchronizes and has more when resetting, and can adopt the method in exchange chip clock source when resetting.Clock frequency information is embedded in the transceiver that sends in the serial data stream on all ply-yarn drills.The word clock of exchange chip is removed to produce main bit clock BCLK by clock multiplier unit (CMU) frequency multiplication, and bit clock is used for the data that serialization sends to transceiver.Transceiver utilizes a clock recovery unit (CRU) to go to recover the frequency and the phase information of bit clock, utilizes clock recovered to send serial data to exchange chip then.At the exchange chip place, only need utilize data recovery unit (DRU) to go to recover phase information, because the serial data that the serial data that exchange chip recovers and it send to ply-yarn drill is a same frequency, can design that DRU goes to absorb factors vary such as Yin Wendu or power supply and the phase change that produces.In a single day set up the bit sequential between exchange chip and the transceiver for a long time, after each exchange chip is reshuffled, no longer needed the extra phase sample time.Another advantage of the method for synchronization is except bit synchronous, between exchange chip and transceiver, also can carry out word synchronously and cell synchronous, command word and information just can transmit between them then.
Traditional serial chi structure has only single control port, thereby becomes the bottleneck of route variable-length packets.In the scheme of asynchronous exchange, exchange chip can not receive and produce data, because it only provides the passage service for serial data stream.Different with it, the synchronous serial switching backplane of word can transmit complete data or command word between transceiver and exchange chip.Utilize the simple Round-Robin arbitration unit in inside of exchange chip, backboard can be realized from route.
Except the design of hardware, the dispatching algorithm of crosspoint also is to realize a critical aspects of high-throughput, execute word synchronously after, when sending data,---arbitration---transmission mechanism of affirmation that carries out connection request between exchange chip and the transceiver carries out the scheduling of formation on ply-yarn drill, change the arbitration algorithm of priority at the enterprising road wheel of exchange chip, and adopt fiduciary dispatching algorithm, and solve the data jamming problem of exchange ply-yarn drill, finally realize high efficiency crosspoint.
3. high-performance route processing engine
R8002 equipment adopts high performance embedded microprocessor as the route processing engine, supports abundant protocol type, and the collection network topology information generates overall routing table.In addition, it supports general SNMP NMP, supports TELNET to be used for management and the configuration to router, also supports the MIBII information database in addition.The system maintenance and the management function of R8002 equipment are also finished by microprocessor.
Two, the system configuration of R8002 equipment
R8002 equipment adopts 19 inches frames of standard, can be installed in easily in the telecommunications room, adopts the backboard and the machine frame system of front plug-in card.Always have 14 groove positions among Figure 17, wherein system disk takies two groove positions 7 and 8, and redundancy protecting is provided; The mixed insertion of various ply-yarn drills is supported in other 12 groove positions, makes things convenient for the user to match different ply-yarn drills as required, and ply-yarn drill can provide FE, GE, POS, MSR/RPR, CWDM interface.
For satisfying the requirement of carrier class high reliability, hardware system possesses high reliability, supports the hot plug of ply-yarn drill, and the heat that cooperates software to finish active/standby parts switches, can interrupting service during system maintenance.Power supply and cooling system all possess the N+1 function of redundancy protection, and can include webmaster in and manage.Hardware system provides perfect system state-detection and fault location ability, for system maintenance provides means efficiently and effectively.
R8002 equipment has adopted the mode of system's adaptation layer and operating system adaptation layer to solve the mapping of physical resource to logical resource on software, and it is from being divided in logic with the lower part:
1. control plane: provide route and signaling capability and other network equipments mutual and processing protocol bag, with management plane together for datum plane generate when transmitting according to various entries and strategy, handle unusually or other option bag.
2. datum plane: finish per-packet and handle, the processing that realization is arrived layer 7 to the layer 2 that transmits packet, provide multiple pass-through mode (layer 2, IP, MPLS etc.) and security strategy to handle (PBR, ACL and VPN), finish counting statistical variable.
3. management plane: finish to agreement and system configuration and management,, way to manage mainly comprises order line (CLI), serial ports graphical interfaces, SNMP network management workstation and based on the graphics management interface of WEB.The function that it is mainly finished has: to the configuration management of router, and the checking of performance, the record of alarm and daily record, the authentication of user security, the maintenance and management of database information.
4. operating system adaptation layer: for the control plane and the management plane on upper strata provides unified interface, the realization details of masking operation system; Provide general system call, as interpolation, deletion and the foundation of single linked list, multilinked list and Hash table; Provide the management of Installed System Memory and buffer area is called.
5. system's adaptation layer: finish control plane adaptive to datum plane,, provide general upper strata api interface to control plane shielding harness hardware configuration details.Finish the mutual of datum plane and control plane information, comprise and transmit entry, protocol package and network management data etc.
Wherein, the core of R8002 device software is multiprocessor communication agreement (MPCP).It finishes the adaptive of datum plane and control plane, management plane, makes that whole router can externally be a complete organism.For protocol package and network management data provide data channel, and realize key data structure on ply-yarn drill and the main control processor synchronously.
In sum, R8002 apparatus logic structure such as Figure 18:
1, the basic demand of sub-shelf structure and circuit board and connector meets following standard:
IEC STANDARD Publication 297 Dimensions of mechanical structure of482.6mm(19inch)series
IEEE(1101.1，1101.10，1101.11)Equipment Practice Eurocard packaging
IEC-61076-4-101 Specification for 2mm Connector Systems
2, casing structure:
Mechanical structure adopts software designs such as AutoCAD, Pro/E, adopts methods such as Theoretical Calculation, software emulation, Meter Test to carry out the cooling system analysis, guarantees that cooling system satisfies the requirement of circuit design, guarantees whole system operation reliably steady in a long-term.
The fundamental characteristics of R8002 equipment cabinets is as follows:
Size:19 inch (482.6mm) industrial standard frame width (comprising the convex edge that usefulness is installed), 13U (578mm) highly (does not comprise the AC rectification unit), the 585mm degree of depth (convex edge is installed from the front portion to begin to calculate)
19 inches width that the user can match, the 110/220V of 3U height exchanges the self adaptation rectifier power source
Mounting: use 19 inches frames of EIA RS-310-C of standard that the convex edge is installed
Slot:14 9U (400mm) slot that possesses hot-swap capabilities
Power Supplies:3～4 hot-swappable, a redundant ac rectifying power source of N+1 (110/220V AC self adaptation) is from the frame previous action.The cabinet rear portion optionally-48V DC input socket.
Fans: hot-swappable blower module.
Air Flow: both sides and front air intake, rear portion air-out.
ESD Grounding: two ESD earth points, forward and backward each one.
Metal：Aluminum alloy，T5052-H32
Metal Plating：Chemical film per MIL-C-5541，clear
There is the fan unit of individual 1U height the top of cabinet, is the whole system heat radiation.Fan unit always has 7 fans and forms; constitute the fan unit of an integral body; lay fan control circuitry and hot plug control circuit on the fan unit; system is the information such as rotating speed, fan-status of monitoring machine case temperature inside, each fan at any time; if monitored amount surpasses the thresholding of setting, then system starts alarm and error protection mechanism automatically.In porch, system air channel dustproof filter screen is installed, is prevented that dust is adsorbed on the circuit board.
Cabinet provides-48V DC and two kinds of external power source input modes of 110/220V AC.If need using, whole system exchanges input, then need be at 19 inches communication power supplies of a 3U height of the below of whole sub-frame (comprising fan unit) apolegamy, these are 19 years old " power rectifier the interchange input capability of wide region partly is provided; can adapt to 110V automatically and exchange input with 220V; output 48VDC, the about 1200W of gross output.This power rectifier partly provides perfect power protection and monitoring function, possesses redundancy and hot-swap capabilities, and can be managed by system board by RS-232 interface.All ply-yarn drills adopt decentralized power-supply methods, and power unit has perfect protection mechanisms such as overvoltage, under-voltage, overcurrent, lightning protection.
Figure 19 is the theory diagram of the power rectifier system of band remote monitoring function:
R8002 equipment is used for reference the thought about the system configuration related specifications of PICMG association issue, in conjunction with R8002
The system demand of equipment has designed the embedded ethernet control bus, supports 10/100M Base-Tx self adaptation, can be upgraded to the speed of Gigabit Ethernet later on.The Ethernet structure of employing star topology has promoted the performance of system greatly, has simplified the control structure of system, has strengthened system's control ability, makes system configuration more flexible.
For the high reliability of enhanced system, the R8002 system adopts two Ethernet power boards that redundancy protecting is provided.The equipment embedded Ethernet of R8002 is as the channel of " Handshake Protocol " between two system disks, and MPCP multiprocessor communication agreement and HA software protocol information are all uploaded carry information at ethernet channel.In addition, Ethernet is also as the passage of transmitting renewal and on-line software updating.
Figure 20 schematically illustrates the structure chart of embedded ethernet system.
Switching fabric is the packet data exchanged center of entire equipment.The switching fabric of R8002 equipment is made of the queue management device on the interface card, backboard, three links of Crossbar chipset.On the ply-yarn drill the schedule level one that divides into groups of queue management device; safeguard the VOQ formation; utilize SeDes the parallel bus of network processing unit to be converted to a plurality of high-speed serial bus of the 2.6Gbps of serial, a plurality of serial-ports have the function of redundancy protecting, accelerated factor, load balancing.On backboard, adopt stripline design, support the high-speed differential signal transmission of 2.6Gbps.Crossbar is the synchronous switch matrix structure of a 80Gbps exchange capacity, carries out two step cones and changes the priority scheduling algorithm, monitors the state of high-speed link simultaneously.R8002 equipment has two swapdisks, and redundancy protecting is provided, and can carry out fast automatic protection under the control of HA software and switch, and has improved the high reliability of system.
Figure 21 is the R8002 devices exchange systematic schematic diagram that adopts the double crossing over structure:
For the gigabit router of high reliability, require system to possess 99.999% reliability.Reach this requirement, must adopt corresponding error protection and parts redundant technology and HA software.Transceiver in the Crossbar chipset has a plurality of high speed serial ports, realizes active/standby redundancy feature, can pass through simple configuration switch operating channel, improves reliability.And system master dish, Switch-Fabric, Ethernet switching node have all adopted the method for 1: 1 Hot Spare, communicate monitoring state each other between two CPU by the Hearteat agreement; The state of two CPU keeps fully synchronously, and when host CPU breaks down, backup CPU will take over the control of system immediately, realize APS.In addition, the power pack of system, radiator fan and user's ply-yarn drill all will adopt the method for redundancy backup.The state information that system master CPU is responsible for monitoring and safeguard whole system and collects each several part, in case certain part of system breaks down, CPU can monitor at once, and enables the work that redundant parts are taken over inoperative component immediately, and CPU will report fault message to the system manager simultaneously.Adopt after the error protection measure, the reliability of whole system increases significantly, and any inoperative component can online easily (on-line) be changed under the condition of not interrupting regular traffic.
All state informations of system can be checked by the gui interface of webmaster.The status indicator LED of several keys is set on each ply-yarn drill.The operating state that on the panel of master cpu, shows system by status indicator LED.If desired, can add the part index signal more in addition.System board can partly be monitored fan unit and alternating current input power supplying, can also the monitoring system temperature.By combining of self-defining system maintenance bus and network management system, R8002 equipment can reflect the running status of system efficiently, for subscriber's local or remote monitoring.
Three, R8002 equipment ply-yarn drill structure
Wuhan Firberhome Networks company is the exchange of multi-service light and the transmission platform of a new generation based on the metropolitan area network Multiple Service Ring R8002 equipment of RPR, and it is to exchange and transmit these two technology and carried out organic integratedly, makes it into as a whole.The R8002 equipment first step at first provides the circuit board of following function:
②2Port Gigbit Ethernet
3. 2port MSR/RPR (flank speed STM-16/OC-48 or STM-4/OC12)
4. system master dish, the Switch-Fabric of 80Gbps exchange capacity
The ply-yarn drill of R8002 equipment adopts unified structure, and except that the CWDM ply-yarn drill, other ply-yarn drills have all adopted the common techniques of network processing unit and queue management device, and difference mainly is positioned at the network processing unit front end.Figure be R8002 Device-General ply-yarn drill structure as shown in figure 22.
In OSI7 layer network model, PHY is positioned at physical layer, it be responsible for and a certain particular medium type between physics, electric interfaces, guarantee the correct transmission of physics bit.
Medium accesses is controlled (MAC) sublayer data stream is carried out the framing processing, guarantees the error free transmission of frame, has the function of medium distribution, conflict processing and addressing simultaneously.
Network processing unit is finished the parsing to the data grouping, functions such as classification, differentiation priority, adjustment, traffic shaping, conversion, service dispatching, forwarding, ordering.
It is adaptive to the interface of queue management device that adaptive FPGA mainly finishes the network processing unit interface.
Queue management device is safeguarded VOQ (VOQ), carries out scheduling by service priority, carries out allocated bandwidth according to algorithm, adopts the SerDes interface interconnected by backboard and Switch-Fabric.
For the POS ply-yarn drill, ply-yarn drill structure slightly change following (as Figure 23):
Transceiver is finished the circuit coding/decoding, carries out the conversion of light/signal of telecommunication.Framer is finished the mapping of IP to the Sonet/SDH frame.
Four, the general structure of software
The software of R8002 is from being divided into following five parts (Figure 24-software hierarchy structure) in logic:
1. control plane: provide route and signaling capability and other network equipments mutual and processing protocol bag, with management plane together for datum plane generate when transmitting according to various entries and strategy, handle unusually or other option bag.
2. datum plane: finish per-packet and handle, realize processing, provide multiple forwarding (layer 2, IP, MPLS and PPPoE) mode and safety (NAT and VPN) strategy to handle to the layer 2 to layer7 that transmits packet.Finish counting to statistical variable.
3. management plane: finish to agreement and system configuration and management,, way to manage mainly comprises order line (CLI), serial ports graphical interfaces, SNMP network management workstation and based on the graphics management interface of WEB.The function that it is mainly finished has: to the configuration management of router, and the checking of performance, the record of alarm and daily record, the authentication of user security, the maintenance and management of database information.
4. operating system adaptation layer: for the control plane and the management plane on upper strata provides unified interface, the realization details of masking operation system; Provide general system call, as interpolation, deletion and the foundation of single linked list, multilinked list and Hash table; Provide the management of Installed System Memory and buffer area is called.
5. system's adaptation layer: finish control plane adaptive to datum plane,, provide general upper strata api interface to control plane shielding harness hardware configuration details.Finish the mutual of datum plane and control plane information, comprise and transmit entry, protocol package and network management data etc.
Five, system's generic adaptation layer (UAL)
Figure 25 is seen in system's adaptation layer signal.
System's adaptation layer on the R8002 is divided into three parts:
(1) the general upper strata api interface that provides for agreement.
(2) the multiprocessor communication agreement of main control card and Line cards.
(3) network processing unit api interface privately owned and that oneself develop.
(4) position of system's adaptation layer in protocol stack
The target of the UAL module of R8002 is the tie (as shown in figure 26) as contact physical layer and upper strata, operation to upper strata shielding physical layer, provide the unified function call of extracting and be provided with the every characteristic of physical layer to the upper strata, and provide unified packet receiving function and physical layer state notifying function call to the upper strata to physical layer.
Generic adaptation layer is positioned on the physical layer, and for layer identification code on all provides general call function to calling of physical layer, effect is the influence of the different physical layer of shielding to last layer identification code, the platform-neutral of layer identification code and portability in the realization.
(5) relation of the interface mappings in the adaptation layer
Interface mappings relation be in the whole R8002 code in very important relation.It concerns as shown in figure 27:
(6) system's adaptation layer Interface design explanation
(6.1) bottom api interface definition
System's adaptation layer general (UAL) api interface comprises following content: routing table modification, the modification of NAT table, the bag that transmits and receive data, MPLS transmit entry modification, two layers of bridging table modification, network management information, interface management and QoS management.
The present invention can be used for all network equipment and chips of supporting IPv4, IPv6 and IPv4/IPv6, comprises the following equipment that is not limited to: router, switch, portable terminal, intelligent terminal, RNC, NODE B, GGSN, SGSN, PDSN, information household appliances, two or three layers of exchange chip and network processing unit etc.
The scope of application of the present invention is mainly at the router kind equipment based on Ipv4/v6, the layer 3 ethernet switch kind equipment, the broadband integrated access kind equipment, have three layers of exchange, the multi-service transport platform of routing function (MSTP) kind equipment and by the Ipv4/V6 equipment under two fine lopps equipment of Resilient Packet Ring (RPR) metropolitan area network that (medium access control) realized as MAC and the wireless domain also comprises and searching relevant thus forwarding engine (Network Processor) and three layers of first time that Ethernet switching chip realized, repeatedly transmit the function of (pressing the sign ID that IPv6 flow label Flow_Label generates) subsequently.
The present invention is different from the mode that traditional packet-by-packet (Packet-based) transmits, and adopted the IP that handles based on data flow (Stream-based) to wrap the thought of transmitting, and, different policing types realized in same physical pipeline, carrying out simultaneously IPv4 and IPv6 forwarding of data by being set.This method can both be supported two stack modes, tunnel style and the address transition mode under IPv4 and the IPv6 environment, also can be used for the forwarding of simple IPv4 or IPv6.Owing to adopted strategic stream mode, this method has avoided adopting different list structures and the IPv4 of complexity and packet-by-packet searching of IPv6, but adopt same table to support multiple application, decide which kind of concrete operation of employing by the policing type in the strategy stream table, cycle of the equipment development of Jian Huaing greatly, improve forward efficiency (improving more than 5 times than common mode under the IPv6 situation), effectively realized the coexistence that IPv4 in same physical pipeline and IPv6 transmit.In addition, this method need not introduced any extra signaling, under the prerequisite that does not change IETF and any standard recommendation of IEEE, has realized the raising of forwarding performance, can not influence the interconnecting property of the Internet simultaneously.At last, the present invention can handle at user's each stream, and making operator provide value added service and guaranteeing service quality becomes possibility.The present invention can be used for IPv4/IPv6 dual stack node and pure IPv4 or the IPv6 node of IPv4 under the IPv6 transitional environment.
Although specifically show and described the present invention with reference to the preferred embodiments of the present invention, but persons skilled in the art should be understood that, under the situation that does not break away from the spirit and scope of the present invention that claims limit, can carry out various modifications on form and the details to it.
Claims (20)
1. realize the method that packet is transmitted with strategic stream mode for one kind in the network equipment, the described network equipment comprises Forwarding plane and control plane at least, and this method may further comprise the steps:
(a) packet in the receiving data stream comprises at least one packet in this data flow;
(b) judge the type of purpose IP address at Forwarding plane;
(c) in the type of Forwarding plane according to different purpose IP addresses, extract the polynary set of properties of corresponding packet, calculate local unique tactful traffic identifier ID of this data flow of sign;
(d) Forwarding plane if discovery has matched strategy stream forwarding list order then carry out step (e), otherwise forwards step (f) according to the accurate search strategy stream forwarding list of described tactful traffic identifier ID order to;
(e), then according to described tactful stream forwarding list order this packet is carried out related content and revise and transmit and operate, and forward step (i) at Forwarding plane if find to have matched strategy stream forwarding list order;
(f) if find no matched strategy stream forwarding list order, illustrate that then first bag or tactful stream forwarding list order that this packet is a data flow are aging, Forwarding plane is delivered to control plane with this packet and is handled;
(g) control plane the outbound port of this locality configuration relevant with application and pairing next jumping in IP address of this purpose according to type, the inbound port of the purpose IP address of this packet with use in the relevant configuration at least one, packet is handled accordingly, and the entry-information relevant with various application generates a corresponding unique tactful stream forwarding list order in the Comprehensive Control plane;
(h) control plane should be distributed to described Forwarding plane by strategy stream forwarding list order, for the packet use of back;
(i) handle next packet.
2. the method for claim 1, wherein the type of purpose IP address is the IPv6 address of IPv4 address, IPv4 compatibility or the IPv6 address of non-IPv4 compatibility.
3. method as claimed in claim 2, wherein when purpose IP address was the IPv4 address, described polynary set of properties comprised IPv4 address, source, purpose IPv4 address, source protocol port, purpose protocol port, protocol type at least; When purpose IP address is the IPv6 address of the IPv6 address of IPv4 compatibility or non-IPv4 compatibility, whether the IPv6 flow label Flow_Label in the judgment data packet header is 0, if the IPv6 flow label Flow_Label in the described data packet head is 0, then described polynary set of properties comprises IPv6 address, source, purpose IPv6 address, source protocol port, purpose protocol port, protocol type at least, if the IPv6 flow label Flow_Label in the described data packet head is not 0, then described polynary set of properties comprises IPv6 address, source, IPv6 flow label Flow_Label at least.
4. the method for claim 1, record in the wherein said tactful stream forwarding list order show should to data stream carry out which kind of business operation type policing type, be used for specifying the outbound port index and the next-hop ip address that send out port of this packet, described step (e) comprising:
(e1) from described tactful stream forwarding list order, know the policing type of this data flow;
(e2) transmit operation accordingly, draw outbound port and link corresponding layer information, this packet is sent from outbound port according to outbound port index list item in the tactful stream forwarding list order and next-hop ip address adjacency list list item pointed.
5. method as claimed in claim 4, wherein the type of purpose IP address is the IPv6 address of IPv4 address, IPv4 compatibility or the IPv6 address of non-IPv4 compatibility, and wherein step (e2) comprising:
When purpose IP address is the IPv4 address, transmit IPv4_FORWARD if policing type is pure IPv4, then carry out the forwarding operation of IPv4, revise TTL, and again calculation check and; If policing type is the automatic tunnel pattern IPv4_IN_IPv6_AUTO of IPv4 in IPv6, then carry out the automatic tunnel operation of IPv4 to IPv6, according to the tunnel ID encapsulated data packet in the tactful stream forwarding list order, the purpose IPv6 address in tunnel is the IPv6 compatible address of purpose IPv4 address correspondence; If policing type is the manual tunnel mode IPv4_IN_IPv6_MANU of IPv4 in IPv6, then carry out the manual tunnel operations of IPv4 to IPv6, according to the tunnel ID encapsulated data packet in the tactful stream forwarding list order, the purpose IPv6 address in tunnel is the address in the tactful stream forwarding list order; If policing type is network address translation NAT_PT, then carry out the NAT-PT operation, according to pseudo-IP address in the tactful stream forwarding list order and pseudo-port, the content of packet is made amendment;
When purpose IP address is the IPv6 address of IPv4 compatibility, if being pure IPv6, policing type transmits IPv6_FORWARD, then carry out the forwarding operation of IPv6, revise TTL; If policing type is the automatic tunnel pattern IPv6_IN_IPv_4AUTO of IPv6 in IPv4, then carry out the automatic tunnel operation of IPv6 to IPv4, according to the tunnel ID encapsulated data packet in the tactful stream forwarding list order, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility; If policing type is network address translation NAT_PT, then carry out the NAT-PT operation, according to pseudo-IP address in the tactful stream forwarding list order and pseudo-port, the content of packet is made amendment;
When purpose IP address is the IPv6 address of non-IPv4 compatibility, if being pure IPv6, policing type transmits IPv6_FORWARD, then carry out the forwarding operation of IPv6, revise TTL; If policing type is the manual tunnel mode IPv6_IN_IPv4_MANU of IPv6 in IPv4, then carry out the tunnel operations of IPv6 to IPv4, according to the tunnel ID encapsulated data packet in the tactful stream forwarding list order, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility.
6. method as claimed in claim 2, in the wherein said step (g) packet handled accordingly and comprise:
(g1) definite data flow of tackling this packet place is carried out the policing type of which kind of business operation; And
(g2) information with relevant policing type sends it back Forwarding plane, so as by Forwarding plane according to determined policing type handle packet.
7. method as claimed in claim 6, wherein,
(A) when purpose IP address is the IPv4 address,
(ga1) judge that at first whether whether inbound port has enabled NAT-PT and purpose IPv4 address is address in the address pool,, otherwise turn to step (ga4) if not then turn to step (ga2);
(ga2) carry out the relevant forwarding information of IPv4 and search and transmit, judge the interface type of next jumping,, then finish the IPv4 forwarding, otherwise carry out step (ga3) at Forwarding plane if be IPv4;
(ga3) if next jumps corresponding interface is IPv6, carry out the tunnel operations of IPv4 in IPv6 at Forwarding plane, wherein destination address be this IPv4 address correspondence the IPv4 compatibility the IPv6 address or be the address, tunnel of manual configuration;
(ga4) carry out the NAT-PT operation at Forwarding plane, the IPv4 bag is converted to the bag of IPv6;
(B) when purpose IP address is the IPv6 address of IPv4 compatibility,
(gb1) judge that at first whether whether inbound port has enabled NAT-PT and destination address is address in the address pool,, otherwise turn to step (gb5) if not then turn to step (gb2);
(gb2) carry out the relevant forwarding information of IPv6 and search and transmit, judge the outgoing interface type of next jumping,, then carry out the IPv6 forwarding, otherwise carry out step (gb3) at Forwarding plane if be IPv6;
(gb3) if the type of next jumping is IPv4, then at first judges port and whether enabled NAT-PT,, otherwise carry out step (gb4) if the sign of NAT-PT operation is arranged then carry out step (gb5);
(gb4) carry out IPv6 in IPv4 operation at Forwarding plane, the purpose IP address after wherein changing is low 32 of original IPv6 destination address;
(gb5) carry out the NAT-PT operation at Forwarding plane, wherein source address is the address in the nat address pool, and the purpose IP address after the conversion is low 32 of original IPv6 destination address;
(C) when purpose IP address is the IPv6 address of non-IPv4 compatibility,
(gc1) search according to the relevant forwarding information of IPv6 and transmit, and judge that next jumps pairing interface type,, then carry out the IPv6 forwarding at Forwarding plane if be IPv6;
(gc2) otherwise be IPv4, then need carry out the tunnel operations of IPv6 in IPv4 at Forwarding plane, wherein the destination address of IPv4 is the IPv4 address in the tunnel of configuration.
8. method as claimed in claim 7, wherein,
If when packet was carried out the IPv4 forwarding, policing type was IPv4_FORWARD, the tactful stream forwarding list purpose content of generation comprises policing type, strategy stream ID, outbound port index, next-hop IP v6 address at least;
If when packet was carried out the IPv6 forwarding, policing type was IPv6_FORWARD, the tactful stream forwarding list purpose content of generation comprises policing type, strategy stream ID, outbound port index, next-hop IP v6 address at least;
If packet is carried out the network address/protocol address conversion, then policing type is NAT_PT, and the tactful stream forwarding list purpose content of generation comprises policing type, strategy stream ID, pseudo-IPv6 address, IPv4 address, source, purpose IPv4 address, source port, destination interface, pseudo-port, protocol type, outbound port index and next-hop IP v6 address at least;
If packet is carried out automatic IPv4 in IPv6 tunnel operations, then policing type is IPv4_IN_IPv6_AUTO, and the tactful stream forwarding list purpose content of generation comprises policing type, strategy stream ID, outbound port index, next-hop IP v6 address, tunnel ID;
If packet is carried out the IPv4 in IPv6 tunnel operations of manual configuration, then policing type is IPv4_IN_IPv6_MANU, and the tactful stream forwarding list purpose content of generation comprises the end address in policing type, strategy stream ID, outbound port index, next-hop IP v6 address, tunnel ID and tunnel;
If packet is carried out the operation of IPv6 in IPv4 automatic tunnel, then policing type is IPv6_IN_IPv4_AUTO, and the tactful stream forwarding list purpose content of generation comprises policing type, strategy stream ID, outbound port index and tunnel ID;
If packet is carried out the manual tunnel operations of IPv6 in IPv4, then policing type is IPv6_IN_IPv4_MANU, and the tactful stream forwarding list purpose content of generation comprises policing type, strategy stream ID, outbound port index and tunnel ID.
9. the method for claim 1, record the overtime sign that shows that this strategy stream forwarding list order is whether overtime in the wherein said tactful stream forwarding list order, the packet use of being mated whenever this strategy stream forwarding list order once promptly refreshes this overtime sign, described control plane goes to check this sign at regular intervals, if this sign is not refreshed within a certain period of time, illustrate that this strategy stream forwarding list order is aging, then deletes this strategy stream forwarding list order.
10. the method for claim 1 records in the wherein said tactful stream forwarding list order:
Source IPv4 Address, 4 bytes: the IPv4 address, source of data flow;
Destination IPv4 Address, 4 bytes: the purpose IPv4 address of data flow;
Source IPv6 Address, 16 bytes: the IPv6 address, source of data flow;
Destination IPv6 Address, 16 bytes: the purpose IPv6 address of data flow;
Protocol Type, 2 bytes: protocol type;
Source Protocol Port, 2 bytes: the source protocol port, decide port into which kind of agreement by protocol type;
Destination Protocol Port, 2 bytes: the purpose protocol port, decide port into which kind of agreement by protocol type;
Flow Label, 2 bytes: the flow label Flow_Label among the IPv6;
Stream ID, 4 bytes: unique ID of a stream of expression;
Alias Port, 2 bytes: pseudo-port, the protocol port after the conversion is used for network address translation;
Alias IPv4 Address, 4 bytes: pseudo-IP address, the IPv4 address after the conversion is used for network address translation;
Alias IPv6 Address, 16 bytes: pseudo-IPv6 address, the IPv6 address after the conversion is used for network address translation;
Tunnel ID, 2 bytes: tunnel ID is used for IPv6 to IPv4 or IPv4 to IPv6 tunnel;
Policy Type, 2 bytes: policing type, show which kind of business operation type that this carries out to data flow, can carry out business customizing by webmaster flexible configuration or user;
QoS, 2 bytes: service quality shows the QoS parameter to this data flow;
Expired Timer, 1 byte: overtime timer, judge whether this data flow is transmitted entry overtime;
TCP Flag, 1 byte: the TCP flag bit is used for judging whether TCP stream finishes;
Output Port Index, 2 bytes: the outbound port index is used for specifying the port that sends out of this packet;
Next Hop IPv4 Address, 4 bytes: next-hop IP v4 address;
Next Hop IPv6 Address, 16 bytes: next-hop IP v6 address;
Tunnel end IPv4 Address, 4 bytes: termination point of a tunnel IPv4 address and pseudo-IPv4 address multiplex; And
Tunnel end IPv6 Addddress, 16 bytes: termination point of a tunnel IPv6 address and pseudo-IPv6 address multiplex.
11. data transfer equipment of transmitting data with strategic stream mode, comprising an at least one Forwarding plane and a control plane, wherein the packet in each Forwarding plane receiving data stream comprises at least one packet in this data flow, and this data transfer equipment comprises:
Strategy stream forwarding list order storage area is used for storage policy stream forwarding list order,
Wherein, described Forwarding plane comprises:
Purpose IP address class type judgment part is used to judge the type of purpose IP address;
Strategy stream ID computing unit is used for the type according to different purpose IP addresses, selects the polynary set of properties of corresponding packet, calculates the tactful traffic identifier ID of this data flow of sign;
Search the unit, be used for accurately searching the tactful stream forwarding list order that described tactful stream forwarding list order storage area is stored, see if there is described strategy stream ID matched strategy stream forwarding list order with this packet according to described strategy stream ID;
Revise and retransmission unit,, then this packet is carried out related content and revise and transmit operation according to described tactful stream forwarding list order if at the tactful stream forwarding list order of searching unit discovery coupling; As not finding the tactful stream forwarding list order that mates searching the unit, then this packet is delivered to control plane,
Described control plane comprises:
Policy processing unit, according in the configuration of the outbound port of the configuration of the type of purpose IP address, inbound port and pairing next jumping in IP address of this purpose at least one, packet is handled accordingly;
Strategy stream forwarding list order generation unit generates corresponding tactful stream forwarding list order according to the processing that packet is carried out, and should be distributed to described tactful stream forwarding list order storage area by strategy stream forwarding list order, for the packet use of back.
12. data transfer equipment as claimed in claim 11, wherein the type of purpose IP address is the IPv6 address of IPv4 address, IPv4 compatibility or the IPv6 address of non-IPv4 compatibility.
13. data transfer equipment as claimed in claim 12, wherein said strategy stream ID computing unit comprises:
First calculating section when being used for being the IPv4 address in purpose IP address, is calculated described strategy stream ID according to the five meta-attribute batch totals of being made up of source IPv4 address, purpose IPv4 address, source protocol port, purpose protocol port, protocol type at least;
IPv6 flow label Flow_Label detecting unit, when being used in purpose IP address being the IPv6 address of the IPv6 address of IPv4 compatibility or non-IPv4 compatibility, whether the IPv6 flow label Flow_Label in the judgment data packet header is 0;
Second calculating section is used for when judging described IPv6 flow label Flow_Label power 0, calculates described strategy stream ID according to the five meta-attribute batch totals of being made up of source IPv6 address, purpose IPv6 address, source protocol port, purpose protocol port, protocol type at least; And
The 3rd calculating section, being used for judging described IPv6 flow label Flow_Label is not 0 o'clock, calculates described strategy stream ID according to the binary set of properties of being made up of source IPv6 address, IPv6 flow label Flow_Label at least.
14. data transfer equipment as claimed in claim 11, record in the tactful stream forwarding list order of storing in the wherein said tactful stream forwarding list order memory cell show to data stream carry out the policing type of which kind of business operation type, with not specifying the outbound port index and the next-hop ip address that send out port of this packet, and described modification and retransmission unit comprise:
Policing type obtains part, is used for from obtaining policing type corresponding to this data flow with described data packet matched tactful stream forwarding list order; And
The classification transmission unit, according to policing type the packet basis is transmitted operation accordingly corresponding to this data flow, draw outbound port and link corresponding layer information according to outbound port index list item in the tactful stream forwarding list order and next-hop ip address adjacency list list item pointed, this packet is sent from outbound port.
15. data transfer equipment as claimed in claim 14, wherein the type of purpose IP address is the IPv6 address of IPv4 address, IPv4 compatibility or the IPv6 address of non-IPv4 compatibility, and described classification transmission unit comprises:
IPv4 transmits operating unit, when policing type is pure IPv4 when transmitting IPv4_FORWARD, carries out the forwarding operation of IPv4, revises TTL, and again calculation check and;
IPv6 transmits operating unit, when policing type is pure IPv6 when transmitting IPv6_FORWARD, carries out the forwarding operation of IPv6, revises TTL;
The NAT-PT operating unit when policing type is network address translation NAT_PT, is carried out the NAT-PT operation, according to pseudo-IP address in the tactful stream forwarding list order and pseudo-port, the content of packet is made amendment;
IPv4 to IPv6 automatic tunnel operating unit, when policing type is the automatic tunnel pattern IPv4_IN_IPv6_AUTO of IPv4 in IPv6, carry out the automatic tunnel operation of IPv4 to IPv6, according to the tunnel ID encapsulated data packet in the tactful stream forwarding list order, the purpose IPv6 address in tunnel is the IPv6 compatible address of purpose IPv4 address correspondence;
The manual tunnel operations of IPv4 to IPv6 unit, when policing type is the manual tunnel mode IPv4_IN_IPv6_MANU of IPv4 in IPv6, carry out the manual tunnel operations of IPv4 to IPv6, according to the tunnel ID encapsulated data packet in the tactful stream forwarding list order, the purpose IPv6 address in tunnel is the address in the tactful stream forwarding list order;
IPv6 to IPv4 automatic tunnel operating unit, when policing type is the automatic tunnel pattern IPv6_IN_IPv4_AUTO of IPv6 in IPv4, carry out the automatic tunnel operation of IPv6 to IPv4, according to the tunnel ID encapsulated data packet in the tactful stream forwarding list order, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility; And
The manual tunnel operations of IPv6 to IPv4 unit, when policing type is the manual tunnel mode IPv6_IN_IPv4_MANU of IPv6 in IPv4, carry out the tunnel operations of IPv6 to IPv4, according to the tunnel ID encapsulated data packet in the tactful stream forwarding list order, the purpose IPv4 address in tunnel is the pairing IPv4 address, IPv6 address of IPv4 compatibility.
16. data transfer equipment as claimed in claim 15, wherein said policy processing unit comprises:
The policing type determining unit, the data flow that is used for this packet place of definite reply is carried out the policing type of which kind of business operation; And
Policing type information is returned the unit, is used for the information of relevant policing type is sent it back Forwarding plane, thereby corresponding to the operation part of this policing type packet is transmitted operation in the classification transmission unit by Forwarding plane.
17. data transfer equipment as claimed in claim 16, wherein said policing type determining unit comprises:
First judgment part is used to judge whether inbound port has enabled NAT-PT;
Second judgment part is used to judge whether purpose IPv4 address is address in the address pool;
Forwarding information is searched the unit, is used to search transmit, and judges the interface type of next jumping,
Wherein, when purpose IP address is the IPv4 address, at first judge by first judgment part whether inbound port has enabled NAT-PT, and judge by second judgment part whether purpose IPv4 address is address in the address pool, if at least one judged result is for being, then determine and to carry out the NAT-PT operation at Forwarding plane, the IPv4 bag is converted to the bag of IPv6; If judged result all is not, then searching the unit by forwarding information carries out the relevant forwarding information of IPv4 and searches and transmit, judge the interface type of next jumping, if the interface type of next jumping is IPv4, determining that then should finish IPv4 at Forwarding plane transmits, is IPv6 if next jumps corresponding interface, then determines to carry out at Forwarding plane the tunnel operations of IPv4 in IPv6, wherein destination address be this IPv4 address correspondence the IPv4 compatibility the IPv6 address or be the address, tunnel of manual configuration
When purpose IP address is the IPv6 address of IPv4 compatibility, at first judge by first judgment part whether inbound port has enabled NAT-PT, and judge by second judgment part whether purpose IPv4 address is address in the address pool, if at least one judged result is for being, then determine and should carry out the NAT-PT operation at Forwarding plane, the purpose IP address after wherein changing is low 32 of original IPv6 destination address; If judged result all is not, then searching the unit by forwarding information carries out the relevant forwarding information of IPv6 and searches and transmit, judge the outgoing interface type of next jumping, if be IPv6, determine that then should carry out IPv6 at Forwarding plane transmits, if the type of next jumping is IPv4, then at first judge port and whether enabled NAT-PT by first judgment part, if have then determine and to carry out the NAT-PT operation at Forwarding plane, the address in the source address power address pool wherein, and destination address is the low 32 of original IPv6 destination address, otherwise determines and should carry out IPv6 in IPv4 operation at Forwarding plane, and wherein destination address is low 32 of original IPv6 destination address;
When purpose IP address is the IPv6 address of non-IPv4 compatibility, searching the unit by forwarding information searches according to the relevant forwarding information of IPv6 and transmits, and judge that next jumps pairing interface type, if be IPv6, determine that then should carry out IPv6 at Forwarding plane transmits; Otherwise be IPv4, then should carry out the tunnel operations of IPv6 in IPv4 at Forwarding plane, wherein the destination address of IPv4 is the IPv4 address in the tunnel of configuration.
18. data transfer equipment as claimed in claim 17, wherein said tactful stream forwarding list order generation unit comprises:
The first entry generating portion, when packet being carried out the IPv4 forwarding, policing type is IPv4_FORWARD, the tactful stream forwarding list purpose content of generation comprises policing type, stream ID, outbound port index, next-hop IP v6 address at least;
The second entry generating portion, when packet being carried out the IPv6 forwarding, policing type is IPv6_FORWARD, the tactful stream forwarding list purpose content of generation comprises policing type, stream ID, outbound port index, next-hop IP v6 address at least;
The 3rd entry generating portion, when packet being carried out the conversion of the network address/protocol address, policing type is NAT_PT, and the tactful stream forwarding list purpose content of generation comprises policing type, stream ID, pseudo-IPv6 address, IPv4 address, source, purpose IPv4 address, source port, destination interface, pseudo-port, protocol type, outbound port index and next-hop IP v6 address at least;
The 4th entry generating portion, when packet being carried out automatic IPv4 in IPv6 tunnel operations, policing type is IPv4_IN_IPv6_AUTO, and the tactful stream forwarding list purpose content of generation comprises policing type, stream ID, outbound port index, next-hop IP v6 address, tunnel ID at least;
The 5th entry generating portion, when packet being carried out the IPv4 in IPv6 tunnel operations of manual configuration, policing type is IPv4_IN_IPv6_MANU, and the tactful stream forwarding list purpose content of generation comprises the end address in policing type, stream ID, outbound port index, next-hop IP v6 address, tunnel ID and tunnel at least;
The 6th entry generating portion, when packet being carried out the operation of IPv6 in IPv4 automatic tunnel, policing type is IPv6 IN IPv4 AUTO, the tactful stream forwarding list purpose content of generation comprises policing type, stream ID, outbound port index and tunnel ID at least;
The 7th entry generating portion, when packet being carried out IPv6 in IPv4 craft tunnel operations, policing type is IPv6_IN_IPv4_MANU, the tactful stream forwarding list purpose content of generation comprises policing type, stream ID, outbound port index and tunnel ID at least.
19. data transfer equipment as claimed in claim 11, record the overtime sign that shows that this strategy stream forwarding list order is whether overtime in the tactful stream forwarding list order of storing in the wherein said tactful stream forwarding list order memory cell, the packet use of being mated whenever this strategy stream forwarding list order once promptly refreshes this overtime sign, described control plane goes to check this sign at regular intervals, if this sign is not refreshed within a certain period of time, illustrate that this strategy stream forwarding list order is aging, then deletes this strategy stream forwarding list order.
20. data transfer equipment as claimed in claim 11 records in the tactful stream forwarding list order of storing in the wherein said tactful stream forwarding list order memory cell:
Source IPv4 Address, 4 bytes: the IPv4 address, source of data flow;
Destination IPv4 Address, 4 bytes: the purpose IPv4 address of data flow;
Source IPv6 Address, 16 bytes: the IPv6 address, source of data flow;
Destination IPv6 Address, 16 bytes: the purpose IPv6 address of data flow;
Protocol Type, 2 bytes: protocol type;
Source Protocol Port, 2 bytes: the source protocol port, decide port into which kind of agreement by protocol type;
Destination Protocol Port, 2 bytes: the purpose protocol port, decide port into which kind of agreement by protocol type;
Flow Label, 2 bytes: the flow label Flow_label among the IPv6;
Stream ID, 4 bytes: unique ID of a stream of expression;
Alias Port, 2 bytes: pseudo-port, the protocol port after the conversion is used for network address translation;
Alias IPv4 Address, 4 bytes: pseudo-IP address, the IPv4 address after the conversion is used for network address translation;
Alias IPv6 Address, 16 bytes: pseudo-IPv6 address, the IPv6 address after the conversion is used for network address translation;
Tunnel ID, 2 bytes: tunnel ID is used for IPv6 to IPv4 or IPv4 to IPv6 tunnel;
Policy Type, 2 bytes: policing type, show which kind of business operation type that this carries out to data flow, can carry out business customizing by webmaster flexible configuration or user;
QoS, 2 bytes: service quality shows the QoS parameter to this data flow;
Expired Timer, 1 byte: overtime timer, judge whether this data flow is transmitted entry overtime;
TCP Flag, 1 byte: the TCP flag bit is used for judging whether TCP stream finishes;
Output Port Index, 2 bytes: the outbound port index is used for specifying the port that sends out of this packet;
Next Hop IPv4 Address, 4 bytes: next-hop IP v4 address;
Next Hop IPv6 Address, 16 bytes: next-hop IP v6 address;
Tunnel end IPv4 Address, 4 bytes: termination point of a tunnel IPv4 address and pseudo-IPv4 address multiplex; And
Tunnel end IPv6 Addddress, 16 bytes: termination point of a tunnel IPv6 address and pseudo-IPv6 address multiplex.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2003101017904A CN100409646C (en) | 2003-10-28 | 2003-10-28 | Method and device for realizing different IP data packet repeating using strategic flow |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2003101017904A CN100409646C (en) | 2003-10-28 | 2003-10-28 | Method and device for realizing different IP data packet repeating using strategic flow |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1612562A CN1612562A (en) | 2005-05-04 |
| CN100409646C true CN100409646C (en) | 2008-08-06 |
Family
ID=34756249
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2003101017904A Expired - Fee Related CN100409646C (en) | 2003-10-28 | 2003-10-28 | Method and device for realizing different IP data packet repeating using strategic flow |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100409646C (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11886325B2 (en) * | 2022-06-30 | 2024-01-30 | Browserstack Limited | Network status simulation for remote device infrastructure |
Families Citing this family (43)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7508829B2 (en) * | 2005-11-30 | 2009-03-24 | Cisco Technology, Inc. | Method and apparatus providing prioritized recursion resolution of border gateway protocol forwarding information bases |
| CN101273590B (en) * | 2005-12-29 | 2012-01-18 | 中兴通讯股份有限公司 | Packet fast forwarding method and system |
| CN100461732C (en) * | 2006-06-16 | 2009-02-11 | 华为技术有限公司 | Ethernet technology switching and forwarding method, system and equipment |
| CN101155196B (en) * | 2006-09-27 | 2011-05-11 | 中国电信股份有限公司 | Service-oriented IPv6 address specification and distribution method, terminal and system for implementing the same |
| CN101335689B (en) * | 2007-06-26 | 2011-11-02 | 华为技术有限公司 | Traceroute implementing method and equipment |
| CN101110772B (en) * | 2007-08-13 | 2010-06-09 | 杭州华三通信技术有限公司 | Device and method for handling message |
| CN101136851A (en) * | 2007-09-29 | 2008-03-05 | 华为技术有限公司 | Stream forwarding method and equipment |
| US7987290B2 (en) * | 2007-12-21 | 2011-07-26 | Microsoft Corporation | Security modes for a routing table distributed across multiple mesh nodes |
| CN101304383B (en) * | 2008-07-07 | 2010-10-27 | 杭州华三通信技术有限公司 | Method for switching message of switching network and switching system |
| CN101325597B (en) * | 2008-07-30 | 2011-04-06 | 北京星网锐捷网络技术有限公司 | Method, apparatus and system for processing data |
| US8295279B2 (en) | 2008-12-02 | 2012-10-23 | Electronics And Telecommunications Research Institute | Routing method and apparatus for providing different path by service |
| CN101582851B (en) * | 2009-06-12 | 2011-11-30 | 中兴通讯股份有限公司 | Method and system for realizing sharing route capacity on dual-stacker router |
| CN101938452B (en) * | 2009-07-01 | 2013-01-09 | 大唐移动通信设备有限公司 | Communication device |
| CN102036334A (en) * | 2009-09-30 | 2011-04-27 | 北京中能普瑞技术有限公司 | Routing control method of radio sensor network for mine |
| CN102377654B (en) * | 2010-08-17 | 2014-06-18 | 国基电子（上海）有限公司 | Router and method for realizing IPv6 (Internet Protocol Version 6) message traversing on IPv4 (Internet Protocol Version 4) router |
| CN101969698B (en) * | 2010-10-25 | 2014-02-12 | 中山大学 | Method for establishing and using mobile IP table of mobile IP application-level gateway |
| CN102025644B (en) * | 2010-12-31 | 2012-10-17 | 华为技术有限公司 | Method and device for load sharing |
| CN102760114B (en) * | 2011-04-29 | 2015-07-08 | 无锡江南计算技术研究所 | Communication emulation method, engine and system for multi-processor system |
| CN103179031B (en) * | 2011-12-23 | 2016-05-11 | 上海博达数据通信有限公司 | Based on multi-service forwarding and the processing method of stream mode |
| CN102546405B (en) * | 2011-12-27 | 2015-05-13 | 华为技术有限公司 | Business processing method and device of protocol stack |
| CN102420772B (en) * | 2011-12-31 | 2014-05-14 | 杭州华三通信技术有限公司 | Tunnel message transmission and receiving methods and devices |
| CN102497385B (en) * | 2011-12-31 | 2015-09-16 | 曙光信息产业股份有限公司 | A kind of network traffics auditing method and auditing system |
| CN102594680B (en) * | 2012-02-15 | 2015-06-17 | 迈普通信技术股份有限公司 | Message fragment processing method |
| CN102904813B (en) * | 2012-11-05 | 2016-03-02 | 华为技术有限公司 | A kind of method of message repeating and relevant device |
| EP2938026B1 (en) | 2013-01-14 | 2018-05-02 | Huawei Technologies Co., Ltd. | Cluster and forwarding method |
| CN103414594A (en) * | 2013-08-23 | 2013-11-27 | 烽火通信科技股份有限公司 | IP stream information statistical method for charging and monitoring |
| WO2015081551A1 (en) * | 2013-12-06 | 2015-06-11 | 华为技术有限公司 | Method, device and system for implementing packet routing in network |
| CN103716253A (en) * | 2013-12-27 | 2014-04-09 | 广州华多网络科技有限公司 | Method and device for requesting data |
| CN104954288B (en) * | 2014-03-28 | 2019-06-14 | 华为技术有限公司 | Method for sending information, device and communication system |
| WO2016023148A1 (en) | 2014-08-11 | 2016-02-18 | 华为技术有限公司 | Packet control method, switch and controller |
| CN104735073B (en) * | 2015-03-30 | 2018-07-06 | 广州杰赛科技股份有限公司 | IPv4-IPv6 transition protocols dispatching method and device |
| CN105072038B (en) * | 2015-08-28 | 2018-12-21 | 华讯方舟科技有限公司 | A kind of data message forwarding method and device |
| CN105391704A (en) * | 2015-10-29 | 2016-03-09 | 国网智能电网研究院 | Configuration port isolation exchange equipment based on service type and application method |
| CN106657436B (en) * | 2016-11-29 | 2019-07-09 | 杭州迪普科技股份有限公司 | Message processing method and device |
| CN107888521B (en) * | 2017-10-20 | 2021-01-01 | 深圳市楠菲微电子有限公司 | Method and device for sharing table resource pool by multiple protocols |
| CN109361782B (en) * | 2018-11-02 | 2021-10-08 | 迈普通信技术股份有限公司 | Message forwarding method and network equipment |
| CN111385212B (en) * | 2018-12-29 | 2021-08-31 | 华为技术有限公司 | Data transmission technology and neural network system |
| CN111147519A (en) * | 2019-12-31 | 2020-05-12 | 奇安信科技集团股份有限公司 | Data detection method, device, electronic equipment and medium |
| CN111711679A (en) * | 2020-06-09 | 2020-09-25 | 宏图智能物流股份有限公司 | Warehouse network unified management platform method |
| CN112003792B (en) * | 2020-07-23 | 2022-04-15 | 烽火通信科技股份有限公司 | Software and hardware cooperative message acceleration method and device |
| CN112422695A (en) * | 2020-12-07 | 2021-02-26 | 重庆忽米网络科技有限公司 | Industrial equipment data forwarding method supporting multiple protocols and multiple rules |
| CN112559246B (en) * | 2020-12-10 | 2024-02-27 | 苏州盛科通信股份有限公司 | Data comparison method and device for hot standby smoothing stage of switch |
| CN115086233B (en) * | 2022-08-17 | 2022-11-11 | 北京左江科技股份有限公司 | FPGA-based network message key information extraction and forwarding method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0840482A1 (en) * | 1996-11-01 | 1998-05-06 | Hitachi, Ltd. | Communicating method between IPv4 terminal and IPv6 terminal and IPv4-IPv6 converting apparatus |
| WO2002045375A2 (en) * | 2000-12-01 | 2002-06-06 | Nortel Networks Limited | Auto-tunnelling in a heterogenous network |
-
2003
- 2003-10-28 CN CNB2003101017904A patent/CN100409646C/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0840482A1 (en) * | 1996-11-01 | 1998-05-06 | Hitachi, Ltd. | Communicating method between IPv4 terminal and IPv6 terminal and IPv4-IPv6 converting apparatus |
| WO2002045375A2 (en) * | 2000-12-01 | 2002-06-06 | Nortel Networks Limited | Auto-tunnelling in a heterogenous network |
Non-Patent Citations (1)
| Title |
|---|
| IPv4向IPv6过渡技术综述. 马严，赵晓宇.北京邮电大学学报，第25卷第4期. 2002 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11886325B2 (en) * | 2022-06-30 | 2024-01-30 | Browserstack Limited | Network status simulation for remote device infrastructure |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1612562A (en) | 2005-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100409646C (en) | Method and device for realizing different IP data packet repeating using strategic flow | |
| US20210051045A1 (en) | Communication switching apparatus for switching data in multiple protocol data frame formats | |
| CN103746931B (en) | A kind of method for computer network, the network equipment and service card | |
| US8718063B2 (en) | Methods and apparatus related to route selection within a network | |
| Chao et al. | High performance switches and routers | |
| JP4454499B2 (en) | Transmission system with functionality of multiple logical sub-transmission systems | |
| EP1417586B1 (en) | Dense virtual router packet switching | |
| EP2466821B1 (en) | Methods and apparatus for forwarding-state transport in a distributed control plane | |
| US7408957B2 (en) | Selective header field dispatch in a network processing system | |
| CN100448216C (en) | Data structure and storage medium for realizing multi-protocol habel exchange system engineering | |
| CN1875585B (en) | Dynamic unknown L2 flooding control with MAC limits | |
| CN102835069B (en) | Apparatus and method for synchronized networks | |
| US10728139B2 (en) | Flexible software-defined networking (SDN) protocol for service provider networks | |
| US20120155453A1 (en) | Methods and apparatus related to a switch fabric system having a multi-hop distributed control plane and a single-hop data plane | |
| US11272396B2 (en) | Frame aggregation method, network setting frame sending method, and device | |
| US20110019572A1 (en) | Method and apparatus for shared shaping | |
| CN104704779A (en) | Method and apparatus for accelerating forwarding in software-defined networks | |
| EP1943785A1 (en) | Multilink trunking for encapsulated traffic | |
| CN102468899B (en) | Channelizing STM-1 access distribution method of and system thereof | |
| CN114465946B (en) | Method, device and system for acquiring forwarding table item | |
| CN104660574B (en) | Configuration method, controlled entity and the Delivery Function of data center | |
| CN105681223A (en) | SDN data packet forwarding method and method | |
| Aweya | Switch/Router Architectures: Shared-Bus and Shared-Memory Based Systems | |
| US7970009B1 (en) | Method for performing protocol translation in a network switch | |
| CN116866249A (en) | Communication system, data processing method and related equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20170411 Address after: 430074 East Lake high tech Development Zone, Hubei Province, No. 6, No., high and new technology development zone, No. four Patentee after: Fenghuo Communication Science &. Technology Co., Ltd. Address before: 430074 Hubei, Wuhan Patentee before: Wuhan Fenghuo Network Co., Ltd. |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20080806 Termination date: 20191028 |