WO2015032214A1

WO2015032214A1 - High-speed routing lookup method and device simultaneously supporting ipv4 and ipv6

Info

Publication number: WO2015032214A1
Application number: PCT/CN2014/077912
Authority: WO
Inventors: 程晨; 李彧; 张炜; 白栎旸; 徐宝魁; 王晓明
Original assignee: 中兴通讯股份有限公司
Priority date: 2013-09-03
Filing date: 2014-05-20
Publication date: 2015-03-12
Also published as: CN104426774A

Abstract

Disclosed are a high-speed routing lookup method and device simultaneously supporting IPv4 and IPv6, which relate to the technical field of network switching. The device comprises three parts, i.e. a routing lookup algorithm software module, a routing update interface module and a routing lookup hardware module, wherein the routing lookup algorithm software module completes software calculations and update entry instruction issuance for a routing entry; after receiving a software update entry instruction, the routing update interface module controls a data stream of the routing lookup hardware module according to the actual operating state of the routing lookup hardware module, and writes an update entry into a memory of the routing lookup hardware module; and the routing lookup hardware module is in a two-way parallel pipelined architecture, respectively responds to IPv4 and IPv6 routing lookup requests of a hardware system, and returns a lookup result of the longest prefix match.

Description

TECHNICAL FIELD The present invention relates to the field of network switching technologies, and in particular, to a method and apparatus for high-speed route searching based on B-Tree and supporting both IPv4 and IPv6. BACKGROUND OF THE INVENTION With the rapid development of the Internet on the Internet, the core routers used for backbone network interconnection have an interface rate of up to 100 Gbps, which requires the core router to achieve a route lookup rate of several million times per second while supporting a large-capacity routing table.

IP search needs to get the longest matching prefix. IPv6 is much larger than IPv4 address space. It requires higher storage and search speed. In recent years, researchers have proposed a variety of hardware search methods to improve the search rate, which is most popular with Trie tree structure and TCAM. .

The Trie tree structure is the most widely used tree structure, and it is easy to implement pipeline operations on hardware, which is beneficial to improve the throughput of route lookup. However, the Trie tree structure has certain limitations. Generally, the Trie tree structure has a large number of pipeline stages, which causes an excessive search delay. The route lookup design using a multi-bit Trie tree structure can greatly reduce the number of pipeline stages, but it will bring a lot of extra memory consumption, and this extra memory consumption is related to the prefix distribution of the routing table, so that the multi-bit Trie tree is based. There is a large ups and downs in the routing table capacity of the route lookup hardware design. In particular, because the key value width difference between IPv4 and IPv6 is 4 times, it is difficult for the same hardware architecture to support IPv4 and IPv6 route lookup at the same time, which increases the complexity and cost of hardware design.

TCAM can implement IPv4 and IPv6 route lookup in the same chip, but to meet the lOOGbps search rate and the core router's routing table capacity requirements, multiple TCAM cascaded lookups are required, and power consumption and heat generation will become serious problems.

B-tree is a software algorithm widely used in database file management. It is characterized by a tree node with M-1 keywords and M child nodes. The depth of B-Tree is determined by the order M and the number of keywords. Decision, regardless of the distribution of keywords. These features enable B-Tree to overcome the shortcomings of the above Trie tree structure, and implement route lookup design with few pipeline stages and insensitive to route prefix distribution. It should be noted that when applying B-Tree for route lookup, you need to ensure that the result returned is the longest prefix match result. Since the lookup backtracking cannot be implemented in the hardware pipeline operation, the route check of the B-Tree is required. Find an algorithm for special handling. There are a number of publicly available methods of doing this, such as copying a parent prefix into multiple copies onto all of its sub-prefixes, or a method of merging the parent prefix with the child nodes at the top of the tree structure. At present, we have proposed a high-speed route search method and device based on B-Tree, which can realize high-performance IPv4 or IPv6 route lookup by hardware pipeline structure, but it can not meet the requirements of supporting both IPv4 and IPv6 route lookup. In view of this, the present invention proposes a method and apparatus capable of simultaneously performing IPv4 and IPv6 route lookup. SUMMARY OF THE INVENTION An object of the embodiments of the present invention is to provide a high-speed route search method and apparatus for supporting both IPv4 and IPv6. In order to solve the problem of IPv4 or IPv6 route lookup in the prior art, it cannot satisfy the same system structure. Support for the need for IPv4 and IPv6 route lookups. According to an aspect of the present invention, a high-speed route search method for supporting both IPv4 and IPv6 is provided, including the following steps: Responding to a route entry update command sent by a route forwarding system in real time, and saving the IPv4 and IPv6 shared tree nodes The routing entry information is updated. During the route lookup, it is determined whether an IPv4 route lookup or an IPv6 route lookup is required according to the route lookup request. If it is determined that an IPv4 route lookup is required, the IPv4 and IPv6 shared tree structure is used to find an IPv4 route lookup. An IPv4 root node, and obtains IPv4 routing information from respective nodes of the shared tree structure according to a pipeline search manner from an IPv4 root node; if it is determined that an IPv6 route lookup is required, the IPv4 and IPv6 shared tree structure is used to find an IPv6 route. The IPv6 root node is searched, and IPv6 routing information is obtained from each corresponding node of the shared tree structure according to a pipeline search manner from the IPv6 root node. Preferably, the IPv4 and IPv6 routing tables are stored in the B-Tree node of the same bit width, and the IPv4 routing table and the IPv6 routing table dynamically share the multi-layer nodes, and each layer node stores multiple routing information. Preferably, the memory area stores: an IPv4 root node address, an IPv6 root node address, the multi-layer node, and a result table. Preferably, obtaining the IPv4 routing information includes:

The IPv4 lookup logic obtains an IPv4 key value from the route lookup request, and obtains an IPv4 root node address from the memory area;

The IPv4 lookup logic uses the IPv4 root node address to find the IPv4 root node from the IPv4 and IPv6 shared tree structure, reads and parses the node data of the node, finds the routing information of the node, and searches the routing information of the next-level node according to the pipeline search mode. Until the routing information of the last-level node is found; wherein the IPv4 search logic obtains the next-level node pointer by comparing the IPv4 key value with the route prefix in the node data of each node being parsed, and Corresponding IPv4 routing information is obtained in the matched routing entry data. Preferably, obtaining the corresponding IPv4 routing information from each matched routing entry data comprises: saving a result table entry pointer in each matching routing entry in a cache of the searching module; a result table entry of the routing entry matching each level of the pipeline The pointer directly covers the contents of the cache; after ending the route lookup for all nodes, the result table is searched according to the final result table entry pointer in the cache to obtain the corresponding IPv4 routing information. Preferably, obtaining the IPv6 routing information includes:

The IPv6 lookup logic obtains an IPv6 key value from the route lookup request, and obtains an IPv6 root node address from the memory area;

The IPv6 lookup logic uses the IPv6 root node address to find the IPv6 root node from the IPv4 and IPv6 shared tree structure. By reading and parsing the node data of the node, the routing information of the node is found, and the routing information of the next-level node is searched according to the pipeline searching manner. Until the routing information of the last-level node is found; wherein the IPv6 search logic compares the IPv6 key value with the route prefix in the node data of each node that is parsed, respectively, to obtain the next-level node pointer, and Corresponding IPv6 routing information is obtained in the matched routing entry data. Preferably, obtaining the corresponding IPv6 routing information from each matched routing entry data includes: saving a result table entry pointer in each matched routing entry in a cache of the searching module; a result table entry of the routing entry matching each level of the pipeline The pointer directly covers the contents of the cache; After the route lookup is completed for all nodes, the result table is searched according to the final result table entry pointer in the cache, and the corresponding IPv6 routing information is obtained. According to another aspect of the present invention, a high-speed route finder device for supporting both IPv4 and IPv6 is provided, including: a route entry information update module, configured to respond to a route entry update command sent by a route forwarding system in real time, to IPv4 The route entry information saved by the IPv6 shared tree node is updated; the route lookup module is determined to be configured to determine whether an IPv4 route lookup or an IPv6 route lookup is required according to the route lookup request during the route lookup request;

The IPv4 route lookup module is configured to: when it is determined that an IPv4 route lookup is required, find an IPv4 root node for IPv4 route lookup from the IPv4 and IPv6 shared tree structure, and start from the IPv4 root node according to the pipeline search manner from the shared tree structure. Each corresponding node obtains IPv4 routing information;

The IPv4 result obtaining module is configured to look up the IPv4 routing information search result table obtained by the IPv4 route lookup module to obtain an actual IPv4 route search result.

The IPv6 route lookup module is configured to: when it is determined that an IPv6 route lookup is required, find an IPv6 root node for IPv6 route lookup from the IPv4 and IPv6 shared tree structure, and start from the IPv6 root node according to the pipeline search manner from the shared tree structure. Each corresponding node obtains IPv6 routing information.

The IPv6 result obtaining module is configured to look up the IPv6 routing information search result obtained by the IPv6 route lookup module to obtain the actual IPv6 route search result. Preferably, the IPv4 route lookup module includes: an IPv4 acquisition unit, configured to obtain an IPv4 key value from the route lookup request by using an IPv4 lookup logic, and obtain an IPv4 root node address from the memory area;

The IPv4 search unit is set to use the IPv4 root node address to find the IPv4 root node from the IPv4 and IPv6 shared tree structure by using the IPv4 root node address, and by reading and parsing the node data of the node, searching for the routing information of the node, and searching according to the pipeline search manner. Routing information of the first-level node until the routing information of the last-level node is searched; wherein the IPv4 search logic compares the IPv4 key value with the route prefix in the node data of each node that is parsed to obtain the next level. The node pointer, and obtain corresponding IPv4 routing information from each matched routing entry data. Preferably, the IPv6 route lookup module includes:

An IPv6 obtaining unit, configured to obtain an IPv6 key value from the route lookup request by using an IPv6 lookup logic, and obtain an IPv6 root node address from the memory area;

The IPv6 search unit is set to the IPv6 lookup logic to find the IPv6 root node from the IPv4 and IPv6 shared tree structure by using the IPv6 root node address, and by reading and parsing the node data of the node, searching for the routing information of the node, and searching according to the pipeline search manner. Routing information of the first-level node until the routing information of the last-level node is found; wherein the IPv6 search logic compares the IPv6 key value with the route prefix in the node data of each node that is parsed to obtain the next level. The node pointer, and obtain corresponding IPv6 routing information from each matched routing entry data. Compared with the prior art, the embodiments of the present invention have the following advantages: The embodiment of the present invention can support large-scale IPv4 and IPv6 routing tables, and the entry capacity does not fluctuate with the distribution of routing prefixes, and the convergence of the IPv4 and IPv6 tree structures is greatly reduced. The memory requirement improves the flexibility of the routing system and satisfies the high-speed route lookup requirements of IPv4 and IPv6 in the core router, which brings convenience to the user and improves the user experience. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a flowchart of a high-speed route searching method for supporting both IPv4 and IPv6 according to an embodiment of the present invention; FIG. 2 is a schematic diagram of a high-speed route searching device for supporting both IPv4 and IPv6 according to an embodiment of the present invention; 3 is a block diagram of a device for performing IPv4 and IPv6 route lookup according to an embodiment of the present invention; FIG. 4 is a schematic diagram of a structure of IPv4 and IPv6 tree structure according to an embodiment of the present invention; FIG. FIG. 6 is a block diagram of a function module of a route lookup hardware module according to an embodiment of the present invention; FIG. 7 is a flowchart of a route lookup hardware module search according to an embodiment of the present invention; FIG. 8 is a prefix provided by an embodiment of the present invention; Schematic. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. FIG. 1 is a flowchart of a high-speed route search method for supporting both IPv4 and IPv6 according to an embodiment of the present invention. As shown in FIG. 1, the method includes the following steps: Step S101: Respond to a route entry update command sent by a route forwarding system in real time. And updating the routing entry information saved by the IPv4 and the IPv6 shared tree node; Step S102: determining, during the route searching, whether the IPv4 route search or the IPv6 route search is performed according to the route lookup request; Step S103: If it is determined that an IPv4 route lookup is required And searching for an IPv4 root node for IPv4 route lookup from the IPv4 and IPv6 shared tree structure, and obtaining IPv4 routing information from each corresponding node of the shared tree structure according to a pipeline search manner from the IPv4 root node; Step S104: An IPv6 route lookup is required, and an IPv6 root node for IPv6 route lookup is searched from the IPv4 and IPv6 shared tree structure, and IPv6 routing information is obtained from each corresponding node of the shared tree structure according to a pipeline search manner from the IPv6 root node. In the embodiment of the present invention, the IPv4 and IPv6 routing tables are stored in the B-Tree node of the same bit width. The IPv4 routing table and the IPv6 routing table dynamically share the multi-layer nodes, and each layer node stores multiple routing information. The memory area stores: an IPv4 root node address, an IPv6 root node address, the multi-layer node, and a result table. Obtaining the IPv4 routing information in the embodiment of the present invention includes: the IPv4 search logic acquires an IPv4 key value from the route lookup request, and obtains an IPv4 root node address from the memory area; the IPv4 lookup logic uses the IPv4 root node address to share from the IPv4 and the IPv6. The tree structure finds the IPv4 root node, finds and analyzes the node data of the node, finds the routing information of the node, and searches the routing information of the next-level node according to the pipeline searching manner, until the routing information of the last-level node is found; The IPv4 lookup logic obtains the next-level node pointer by comparing the IPv4 key value with the route prefix in the node data of each node that is parsed, and obtains corresponding IPv4 routing information from each matched route entry data. The IPv4 route prefix in the node includes a VPN, an IPv4 prefix, and an IPv4 prefix vector (Vector). Each bit of the IPv4 prefix vector indicates a prefix length of the prefix, if one of them If the bit is ', the prefix with the corresponding length of the bit exists in the prefix. In this method, multiple IPv4 route prefixes with different parent lengths can be stored in the same IPv4 prefix. Preferably, obtaining the corresponding IPv4 routing information from each matched routing entry data comprises: storing a result table entry pointer in each matching routing entry in a cache of the searching module; a result table entry of the routing entry matching each level of the pipeline The pointer directly covers the contents of the cache; after ending the route lookup for all nodes, the result table is searched according to the final result table entry pointer in the cache to obtain the corresponding IPv4 routing information. Obtaining the IPv6 routing information in the embodiment of the present invention includes: the IPv6 search logic acquires an IPv6 key value from the route lookup request, and obtains an IPv6 root node address from the memory area; the IPv6 lookup logic uses the IPv6 root node address to share from IPv4 and IPv6. The tree structure finds the IPv6 root node, finds and analyzes the node data of the node, finds the routing information of the node, and searches the routing information of the next-level node according to the pipeline searching manner, until the routing information of the last-level node is found; The IPv6 lookup logic obtains the next-level node pointer by comparing the IPv6 key value with the route prefix in the node data of each node that is parsed, and obtains corresponding IPv6 routing information from each matched route entry data. The IPv6 route prefix in the node includes a VPN, an IPv6 prefix, and an IPv6 prefix vector (Vector). Each bit of the IPv6 prefix vector indicates a prefix length of the prefix. If one of the bits is '1' The prefix of the corresponding length of the bit exists in the prefix. In this method, multiple IPv6 route prefixes with different parent lengths can be stored in the same IPv6 prefix. Preferably, obtaining the corresponding IPv6 routing information from each matched routing entry data includes: saving a result table entry pointer in each matched routing entry in a cache of the searching module; a result table entry of the routing entry matching each level of the pipeline The pointer directly covers the contents of the cache; after ending the route lookup for all nodes, the result table is searched according to the final result table entry pointer in the cache to obtain the corresponding IPv6 routing information. The IPv4 and IPv6 shared tree structure in the embodiment of the present invention has a multi-layer node stored in a memory area, and the IPv4 and IPv6 nodes have the same data bit width, and each layer node can store multiple pieces of IPv4 routing information and multiple IPv6 routes. Information, dynamically assigned to IPv4 or IPv6 usage. FIG. 2 is a schematic diagram of a high-speed route finder device for supporting both IPv4 and IPv6 according to an embodiment of the present invention. As shown in FIG. 2, the method includes: a route entry information update module 201, configured to respond to a route delivered by a route forwarding system in real time. An entry update command, updating routing entry information saved by the IPv4 and IPv6 shared tree nodes; determining a route lookup module 202, configured to determine whether an IPv4 route lookup or an IPv6 route lookup is required according to the route lookup request during the route lookup; IPv4 route The searching module 203 is configured to: when determining that an IPv4 route lookup is required, find an IPv4 root node for IPv4 route lookup from the IPv4 and IPv6 shared tree structure, and start from the IPv4 root node according to the pipeline search manner from the shared tree structure. Each corresponding node obtains The IPv6 routing information module is configured to: when it is determined that an IPv6 route lookup is required, find an IPv6 root node for IPv6 route lookup from the IPv4 and IPv6 shared tree structure, and start from the IPv6 root node according to the pipeline search manner. Each corresponding node of the shared tree structure obtains IPv6 routing information. The embodiment of the present invention further includes: an IPv4 result obtaining module 205, configured to perform an IPv4 routing information search result table obtained by the IPv4 route search module, and obtain an actual IPv4 route search result; the IPv6 result obtaining module 206 is set to In the IPv6 route lookup, the IPv6 route information lookup result table obtained by the IPv6 route lookup module is obtained, and the actual IPv6 route search result is obtained. The IPv4 route lookup module 203 includes: an IPv4 acquisition unit, configured to obtain an IPv4 key value from the route lookup request, and obtain an IPv4 root node address from the memory area; and an IPv4 search unit, set to an IPv4 lookup The logic uses the IPv4 root node address to find the IPv4 root node from the IPv4 and IPv6 shared tree structure, reads and parses the node data of the node, finds the routing information of the node, and searches the routing information of the next-level node according to the pipeline search manner until Finding routing information of the last-level node; wherein the IPv4 search logic obtains the next-level node pointer by comparing the IPv4 key value with the route prefix in the node data of each node that is parsed, and The corresponding IPv4 routing information is obtained in the routing entry data. The IPv6 route lookup module 204 includes: an IPv6 obtaining unit, configured to obtain an IPv6 key value from the route lookup request by using an IPv6 lookup logic, and obtain an IPv6 root node address from the memory area;

The IPv6 search unit is set to the IPv6 lookup logic to find the IPv6 root node from the IPv4 and IPv6 shared tree structure by using the IPv6 root node address, and by reading and parsing the node data of the node, searching for the routing information of the node, and searching according to the pipeline search manner. Routing information of the first-level node until the routing information of the last-level node is found; wherein the IPv6 search logic compares the IPv6 key value with the route prefix in the node data of each node that is parsed to obtain the next level. The node pointer, and obtain corresponding IPv6 routing information from each matched routing entry data. FIG. 3 is a structural block diagram of a device for simultaneously performing IPv4 and IPv6 route searching according to an embodiment of the present invention. As shown in FIG. 3, the method includes a route lookup algorithm software module 301, a route update interface module 302, and a route lookup hardware module 303. The route search algorithm software module 301 completes the software calculation and update entry instruction delivery of the route entry. After receiving the software update entry instruction, the route update interface module 302 searches the actual working state of the hardware module 303 according to the route, and controls the route lookup. The data stream of the hardware module 303 is written into the memory of the route lookup hardware module 303; the route lookup hardware module 303 is a two-way parallel pipelined architecture that responds to the hardware system's IPv4 and IPv6 route lookup requests and returns The longest prefix matches the search result. Preferably, the route lookup algorithm software module 301 runs on the main control CPU, and may be a B-Tree algorithm software program written in a high-level algorithm language (such as C, C++ language) to calculate and update the hardware routing table. The route update interface module 302 is a hardware module, which includes a buffer area and a logical processing area. The buffer area is used to store the update node information written by the route lookup algorithm software module 301. The logic processing area is configured to search the related state of the hardware module 303 according to the route, and write the update node data into the route lookup hardware module 303 according to a certain order. In the node memory. The route lookup hardware module 303 is the main module of hardware route lookup, and is a pipeline design mainly composed of an IPv4 lookup logic area, an IPv6 lookup logic area, and a memory area. The IPv4 and IPv6 lookup logic respectively processes the corresponding lookup requests and returns the search results. The memory area is also a hierarchical structure for storing node data of IPv4 and IPv6. The specific implementation method of the present invention is mainly divided into three parts: tree structure design, software module design and hardware module design. FIG. 4 is a schematic diagram showing the fusion of the IPv4 and IPv6 tree structures provided by the embodiment of the present invention. As shown in FIG. 4, the tree structure design of the present invention is to determine the node width of the IPv4 and IPv6 nodes, and the tree structure orders M1 and M2. The tree structure heights N1 and N2, the number of nodes per layer, and the number of result table entries, thereby determining the corresponding number of pipeline stages of the hardware, the occupied space of each layer node, and the space occupied by the result table. The structure design needs to comprehensively consider the basic data bit width of IPv4 and IPv6, the maximum requirements of the routing table, the cost of reading and writing hardware memory, hardware layout and other factors.

Both IPv4 and IPv6 route lookups are implemented by two pipeline lookup logics, but their tree node storage space is shared, and IPv4 and IPv6 nodes of the same pipeline level are stored in the same memory address space. Under the condition that the tree node data bit width of IPv4 and IPv6 are guaranteed to be the same, the tree structure generated by IPv4 and IPv6 can be as shown in Figures 401 and 402, since each key-value width of IPv4 is 32-bit, and each of IPv6 The key-value width is 128-bit, so under the same bit-width node data structure, the order M1 of the B-Tree of IPv4 is greater than the order M2 of IPv6 (the tree node of IPv4 can index more child nodes). Therefore, from the image of the tree, the tree shape of IPv4 is short and fat, and the tree shape of IPv6 is high and thin. If you simply add two routing tables, the memory requirement is a simple sum of the two tree areas, and their hardware design will be independent of each other. The design of the embodiment of the present invention is as shown in FIG. 403. The two trees are merged, and the bottoms of the trees are at the same level. The tree nodes of each level are determined by their respective maximum demands, and the maximum number of levels is determined by the tree height of IPv6. Decide. The advantages of this processing are: Reduce the memory space requirement, and meet the requirements of large-capacity routing tables of IPv4 and IPv6; IPv4 and IPv6 node space sharing, can automatically adjust the actual capacity of the two routing tables automatically, and improve the actual application of the search design. Adaptability; Simplify hardware design complexity and reduce hardware costs. FIG. 5 is a block diagram showing the function modules of the route lookup algorithm software module provided by the embodiment of the present invention. As shown in FIG. 5, the entire routing algorithm software module is implemented based on the B-Tree algorithm, and the overall framework is divided into four parts: the IPv4 algorithm module 501 The IPv6 algorithm module 502, the memory management module 503, and the update hardware operation module 504. The IPv4 algorithm module architecture 501 includes two parts: an insert operation module and a delete operation module. The main operation of the insert operation module is to respond to the route insert instruction of the route forwarding system, and insert the delivered route entry into the B-Tree tree structure. The main operation of the delete operation module is to respond to the route deletion instruction of the route forwarding system, and delete the delivered route entry from the B-Tree tree structure. The IPv4 algorithm module 501 and the IPv6 algorithm module 502 use the same algorithm, which are both B-Tree-based route lookup algorithms, except that the bit width of the key value and the order of the B-Tree are different. The function of the memory management module 503 is to manage the node data memory and the result table data memory in the route lookup algorithm, wherein the space of the IPv4 and IPv6 nodes and the result table are uniformly allocated by the management function, that is, the nodes of the IPv4 and the IPv6. The result tables are allocated in the same resource pool, enabling them to share resources with each other and improve the adaptability of the search device. The primary function of the update hardware operation module 504 is to update tree nodes and result table entries in hardware memory. First, the update hardware operation module 504 records all the information that needs to be updated in the cache; then, after inserting the operation module or deleting the operation module, converting the tree node from the software data format to the agreed hardware data format; finally, the converted hardware The data and the corresponding hardware mapped address can be continuously written into the route update interface module through the software and hardware interaction interface, but not limited to the LocalBus or PCIe interface. The significance of the update hardware operation module 504 is to reduce the interface interaction of the hardware and software, save the update time, and reduce the impact of the item update on the actual hardware lookup flow. FIG. 6 is a block diagram showing the function module of the route lookup hardware module provided by the embodiment of the present invention. As shown in FIG. 6, the route update interface module 601 and the route lookup hardware module 604 are included. The route update interface module 601 includes a buffer area 602 and a logical processing area 603. The main function of the buffer area 602 is to receive the node data written by the storage software through the software and hardware interaction interface, the result table entry data and their corresponding hardware addresses, which have been converted into hardware data formats by software. The main function of the logical processing area 603 is to update the tree node and the result table entry data in the hardware memory under the condition that the correctness of the normal search is not affected. The main function of the route lookup hardware module 604 is to implement IPv4 and IPv6 routing hardware lookup functions, including IPv4 lookup logic 605, IPv6 lookup logic 606, memory area 607, instruction dispatch 609, and result aggregation 610. The main function of the IPv4 lookup logic 605 is to complete the IPv4 route lookup and return the IPv4 route lookup result. The IPv4 lookup pipeline has a total of N1+1 levels, where LV_1 to LV_N1 are tree node lookup logic layers, and LV_N1+1 is a result table lookup logic layer. Each search logic layer corresponds to a separate memory space in the memory area, ensuring that the accesses of the layers to the memory area do not conflict with each other. The main function of the IPv6 lookup logic 606 is to complete the IPv6 route lookup and return the IPv6 route lookup result. The IPv6 lookup pipeline has a total of N2+1 levels, where LV_1 to LV_N2 are tree node lookup logic layers, and LV_N2+1 is a result table lookup logic layer. Each search logic layer corresponds to a separate memory space in the memory area, ensuring that the accesses of the layers to the memory area do not conflict with each other. The memory area 607 is a data storage area in the route lookup hardware module, and each memory area in the memory area is independent of each other, and is divided into a root node address area, each layer node area, and a result table area. The root node address 608 includes an IPv4 root node address and an IPv6 root node address for respectively storing the addresses of the root nodes of the B_Tree tree structure of IPv4 and IPv6. Each layer of the tree node area is independent of each other and does not overlap with other layer tree node areas, so as to avoid conflicts in the simultaneous access of the pipeline. The resulting table area is also a separate area accessed by the last level of the search flow. In the IPv4 lookup logic 605 and the IPv6 lookup logic 606, the lookup pipeline bottom layer is at the same level, but the upper layer may have different IPv6 pipeline stages than the IPv6 pipeline level due to the characteristics of the IPv4 tree structure and the IPv6 tree structure. The N1+1-level pipeline and the N2+1-level pipeline are respectively configured. Therefore, the sequence between the IPv4 and IPv6 route lookup requests is not guaranteed in the design of the embodiment of the present invention, and only the order between the IPv4 request and the IPv4 request is guaranteed. And the ordering between IPv6 requests and IPv6 requests. At the same time, since the number of levels of the lookup pipeline is different, there are cases where the IPv4 lookup logic 605 and the IPv6 lookup logic 606 simultaneously access the same layer node in the memory area. This situation can be solved by a variety of methods, such as dual-port RAM with two sets of read-write ports in the memory RAM selection; logical scheduling can also be used to schedule read requests for lookup logic. The main function of the instruction distribution module 609 is to distinguish whether it is an IPv4 instruction or an IPv6 instruction according to the characteristics of the external input instruction, and then distribute it to the corresponding search logic for processing. The main function of the result aggregation module 610 is to schedule the IPv4 and IPv6 search results and return them separately. FIG. 7 is a flowchart of searching for a route lookup hardware module according to an embodiment of the present invention. As shown in FIG. 7, the method includes the following steps: Step S701: The route lookup hardware module receives the input search request. Step S702: The input search request search key value is distributed, the IPv4 request is sent to the IPv4 search logic, and the IPv6 request is sent to the IPv6 check logic. Step S703: The search key value information and the root node address information in the memory area are simultaneously sent to the first level search stream; Step S704: In the logic of the pipeline, first determine whether the root node address is the current level node, if it is the current level node , initiating a read node request to the memory, waiting for the memory to return the node data; if not the current level node, not requesting to the memory, leaving the current node information unchanged; step S705: parsing the obtained node information, and the key value If the node is not accessed, the analysis comparison is not required, and the waiting is performed; Step S706: If there is a hit, the result of the comparison is recorded, and the result corresponding to the hit route entry is recorded, and the result of the previous hit is overwritten; If the node of the level is not accessed, no recording is required, and the beating waits; Step S70 7: If a comparison is made, the node address that should be accessed by the next level is recorded in the node address from the comparison result; if the node of the current level is not accessed, the current node information is retained unchanged; Step S708: Searching for the water carrying key value information The node address and the hit result enter the next stage flow processing logic; Step S709: If the next stage flow is not the last stage, the process is the same as the above S704-S708 process, and if the next stage flow is the last stage, the access is The processing logic S710 of the result table; Step S710: In the result table water level, the logic will access the result table of the memory area according to the hit result table entry address, and obtain the hit actual result; if it is missed, the result table is not accessed, and the output misses Information: Step S711: The search result is sent to the result aggregation logic and the result of another search flow is scheduled; Step S712: return the actual result of the hit through the hardware interface, and if not, return the miss information. FIG. 8 is a schematic diagram of a prefix structure provided by an embodiment of the present invention. As shown in FIG. 8, an IPv4 prefix structure 801, an IPv6 prefix structure 802, an IPv4 node pointer 20-bit 803, and an IPv6 node pointer 20-bit 804 are included. Each IPv4 prefix structure 801 includes: a 32-bit prefix, a 16-bit VPN, and a 24-bit result supports more than 16M routing tables. The 26-bit Vector does not have an IPv4 prefix with a prefix length of 1_7 in the actual routing table. . So the bit width required for the IPv4 prefix structure is 98-bit. Each IPv6 prefix structure 802 includes a 128-bit prefix, a 16-bit VPN, and a 24-bit result supports a routing table of more than 16M. The 114-bit Vector does not have an IPv6 prefix with a prefix length of 1_15 in the actual routing table. . So the bit width of the required IPv6 prefix structure is 282-bit.

Both IPv4 and IPv6 node pointers are 20-bit, which satisfies a large-scale tree structure. Then the data bit width of each layer node is A*M+B*M+1. Where A is the bit width of the prefix structure, M is the number of prefix structures, and B is the bit width of the node pointer. Since most of the tree nodes may be stored on the DDR, the access data bit width of the DDR is an integer multiple of 128-bit, so the data bit width of each layer node is preferably an integer multiple of 128. Then A1*M1+20*M2+1 128N, A2*M2+20*M2+1 128N are required. A1 is the prefix structure width of IPv4, A2 is the IPv6 prefix structure bit width, and Ml and M2 are the number of IPv4 and IPv6 prefixes in each layer node respectively. The values that can be selected are:

Ml = 8, M2 = 3, N = 8; At this time, the IPv4 node requirement bit width is 964_bit, the IPv6 requirement bit width is 926-bit, and the actually allocated node bit width is 1024-bit. In this configuration, the number of pipelines under the large capacity requirement is small, the DDR performance is generally required, and the wasted memory is small. It can be expected that in this configuration, the 4M capacity IPv4 routing table and the 1M capacity IPv6 routing table requirement are required, and 11 levels of pipeline water and 600 Mb of node memory space are required. You can also choose a value that is:

Ml = 5, M2 = 2, N=5; at this time, the IPv4 node requirement bit width is 610-bit, the IPv6 requirement bit width is 624-bit, and the actually allocated node bit width is 660-bit. In this configuration, the DDR performance requirements are low, and the wasted memory is the least, but the number of pipeline stages is high. Especially for IPv6 routing, more pipeline stages are needed to satisfy the larger-scale routing table. It can be expected that in this configuration, to meet the 512k capacity of IPv4 routing and 32k capacity IPv6 routing, 11 levels of pipeline and 70 Mb of node memory space are required. The data format of the embodiment of the present invention is mainly divided into a node data format and a result table data format. The node data includes: key value, tag vector, result table address, and next level pointer. The key value is the route prefix in the routing table; the tag vector is a tag vector in the B-Tree algorithm that marks different prefix length entries in the same prefix; The table address is the address information of the result list corresponding to the routing entry in the result table; the next level pointer is the node address that should be accessed by the next level after the level search. The IPv4 B-Tree node and the IPv6 node have the same node bit width, and the tree nodes of the same pipeline level are stored in the same memory address space, and are distinguished by different node resolution logics of IPv4 and IPv6. The hardware lookup structure designed in the embodiment of the present invention is a bottom-up B-Tree tree structure. First, define the order M1&M2 and tree height N1&N2 of B-Tree according to the IPv4 and IPv6 routing table capacity requirements, and set the corresponding pipeline level according to their respective tree heights. Then define the maximum tree structure of each layer according to the algorithm filling situation. The number of nodes needs to be set according to the maximum number of nodes; finally, the memory space of the result table is set according to the actual capacity requirement of the routing table. The embodiment of the invention designs the software algorithm module, the hardware data structure, the update process, the search pipeline structure and the memory structure of the route lookup system as a whole, and can simultaneously satisfy the route lookup of IPv4 and IPv6, and satisfy the high performance search of the large-capacity routing table. demand. In summary, the embodiments of the present invention have the following technical effects: The embodiments of the present invention can support large-scale IPv4 and IPv6 routing tables, and can simultaneously satisfy the high-speed route searching requirements of IPv4 and IPv6 in the core router, and bring convenience to the user. And improve the user experience. Although the invention has been described in detail above, the invention is not limited thereto, and various modifications may be made by those skilled in the art in accordance with the principles of the invention. Therefore, modifications made in accordance with the principles of the invention are understood to fall within the scope of the invention. Industrial Applicability The technical solution provided by the embodiments of the present invention can be applied to the field of network switching technologies, and can support large-scale IPv4 and IPv6 routing tables. The capacity of the entry does not fluctuate with the distribution of the routing prefix. The fusion of the IPv4 and IPv6 tree structures greatly reduces the memory. The demand increases the flexibility of the routing system and satisfies the high-speed route lookup requirements of IPv4 and IPv6 in the core router, which brings convenience to the user and improves the user experience.

Claims

The method for claiming a high-speed route search method for supporting both IPv4 and IPv6, comprising: responding to a route entry update command sent by a route forwarding system in real time, and updating routing entry information saved by the IPv4 and IPv6 shared tree nodes;

During route lookup, it is determined whether an IPv4 route lookup or an IPv6 route lookup is required according to the route lookup request; if it is determined that an IPv4 route lookup is required, the IPv4 root node for IPv4 route lookup is searched from the IPv4 and IPv6 shared tree structure, and The IPv4 root node starts to obtain IPv4 routing information from each corresponding node of the shared tree structure according to a pipeline search manner;

If it is determined that an IPv6 route lookup is required, the IPv6 root node for IPv6 route lookup is searched from the IPv4 and IPv6 shared tree structure, and the IPv6 route is obtained from each corresponding node of the shared tree structure according to the pipeline search manner from the IPv6 root node. information. The method according to claim 1, wherein the IPv4 and IPv6 routing tables are stored in a B-Tree node of the same bit width, and the IPv4 routing table and the IPv6 routing table dynamically share the multi-layer node, and each node of the layer is saved in multiple pieces. Routing information. The method according to claim 2, wherein the memory area stores: an IPv4 root node address, an IPv6 root node address, the multi-layer node, and a result table. The method of claim 2, wherein obtaining the IPv4 routing information comprises:

The IPv4 lookup logic uses the IPv4 root node address to find the IPv4 root node from the IPv4 and IPv6 shared tree structure, reads and parses the node's node data, finds the node's routing information, and finds the routing information of the next-level node according to the pipeline search mode. Until the routing information of the last-level node is found; wherein the IPv4 search logic obtains the next-level node pointer by comparing the IPv4 key value with the route prefix in the node data of each node being parsed, and Corresponding IPv4 routing information is obtained in the matched routing entry data. The method of claim 4, wherein the obtaining the corresponding IPv4 routing information from each of the matched routing entry data comprises: The result table entry pointer in each matching routing entry is saved in the cache of the lookup module; the result table entry pointer of each level of the matching route entry directly covers the content in the cache; after ending the route lookup for all nodes, according to the cache The final result table entry pointer looks up the result table and obtains the corresponding IPv4 routing information. The method of claim 2, wherein obtaining IPv6 routing information comprises:

The IPv6 lookup logic uses the IPv6 root node address to find the IPv6 root node from the IPv4 and IPv6 shared tree structure, reads and parses the routing data of the node, finds the routing information of the node, and searches the routing information of the next-level node according to the pipeline search mode. Until the routing information of the last-level node is found; wherein the IPv6 search logic compares the IPv6 key value with the route prefix in the node data of each node that is parsed, respectively, to obtain the next-level node pointer, and Corresponding IPv6 routing information is obtained in the matched routing entry data. The method according to claim 6, wherein the obtaining the corresponding IPv6 routing information from each matched routing entry data comprises:

The result table entry pointer in each matching routing entry is saved in the cache of the lookup module; the result table entry pointer of each level of the matching route entry directly covers the content in the cache; after ending the route lookup for all nodes, according to the cache The final result table entry pointer looks up the result table and obtains the corresponding IPv6 routing information. A high-speed route finder device that supports both IPv4 and IPv6, and includes: a route entry information update module, configured to respond to a route entry update command sent by the route forwarding system in real time, and perform routing entry information saved by the IPv4 and IPv6 shared tree nodes. Updating; determining a route lookup module, configured to determine whether an IPv4 route lookup or an IPv6 route lookup is required according to the route lookup request during route lookup;

The IPv4 route lookup module is configured to: when it is determined that an IPv4 route lookup is required, find an IPv4 root node for IPv4 route lookup from the IPv4 and IPv6 shared tree structure, and start from the IPv4 root node according to the flow water lookup manner from the shared tree structure. Each corresponding node obtains IPv4 routing information;

The IPv4 result obtaining module is configured to perform an IPv4 routing search result table according to the IPv4 routing information search result table obtained by the IPv4 route lookup module, and obtain an actual IPv4 route search result. The IPv6 route lookup module is configured to: when it is determined that an IPv6 route lookup is required, find an IPv6 root node for IPv6 route lookup from the IPv4 and IPv6 shared tree structure, and start from the IPv6 root node according to the pipeline search manner from the shared tree structure. Each corresponding node obtains IPv6 routing information;

The IPv6 result obtaining module is configured to look up the IPv6 routing information search result obtained by the IPv6 route lookup module to obtain the actual IPv6 route search result. The device according to claim 8, wherein the IPv4 route lookup module comprises:

An IPv4 obtaining unit, configured to obtain an IPv4 key value from the route lookup request by using an IPv4 lookup logic, and obtain an IPv4 root node address from the memory area;

The IPv4 search unit is set to use the IPv4 root node address to find the IPv4 root node from the IPv4 and IPv6 shared tree structure by using the IPv4 root node address, and by reading and parsing the node data of the node, searching for the routing information of the node, and searching according to the pipeline search manner. Routing information of the first-level node until the routing information of the last-level node is found;

The IPv4 search logic obtains the next-level node pointer by comparing the IPv4 key value with the route prefix in the node data of each node that is parsed, and obtains the corresponding IPv4 route from each matched route entry data. information. 0. The device according to claim 8, wherein the IPv6 route lookup module comprises:

The IPv6 search unit is set to IPv6 lookup logic to find the IPv6 root node from the IPv4 and IPv6 shared tree structure by using the IPv6 root node address, and by reading and parsing the node data of the node, searching for the routing information of the node, and searching according to the pipeline search manner. Routing information of the first-level node until the routing information of the last-level node is found;

The IPv6 search logic obtains the next-level node pointer by comparing the IPv6 key value with the route prefix in the node data of each node that is parsed, and obtains the corresponding IPv6 route from each matched route entry data. information.