CN1787477A - Method for searching IPv6 routing table - Google Patents
Method for searching IPv6 routing table Download PDFInfo
- Publication number
- CN1787477A CN1787477A CNA2004100099954A CN200410009995A CN1787477A CN 1787477 A CN1787477 A CN 1787477A CN A2004100099954 A CNA2004100099954 A CN A2004100099954A CN 200410009995 A CN200410009995 A CN 200410009995A CN 1787477 A CN1787477 A CN 1787477A
- Authority
- CN
- China
- Prior art keywords
- prefix
- route
- subclass
- searching
- ipv6
- 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.)
- Pending
Links
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
This invention relates to a looking up method for IPv6 route list including the following steps: a, dividing the IPv6 route list into two prefix sub-sets, b, applying an index list to index the prefix sub-set with more or longer prefix sub-sets and utilizing Trie tree of multiple branches to look up the sets with less number and shorter length, and c, utilizing at least two microcomputer modules in the whole sphere to look up the sub-sets in parallel and looking up them by route pre-computation.
Description
Technical field
The present invention relates to interference networks agreement of future generation, more specifically, relate to a kind of IPv6 routing table finding method, can fast and effeciently search the IPv6 routing table of core network.
Background technology
Internet user's continuous growth and multimedia application universal day by day makes that the data traffic exponentially level in the Internet network increases.Physical link and connected node are two fundamentals that constitute the Internet network.The breakthrough of optical transport technology research makes the transmission rate of physical link can satisfy the growth of Internet network traffic data.And,, be difficult to catch up with the speed that the Internet network traffics increase owing to innerly need carry out complicated data processing as the router of Internet network connected node.How to design and realize that high-performance router becomes the research focus in Internet field.
Router is responsible for routing forwarding is carried out in grouping in network.When transmitting grouping, router need carry out route querying based on the purpose IP address of each packet header in routing table, finds correct next of grouping to jump out mouth.In high-performance router, interface rate is more and more higher, so corresponding inner forwarding processing also must be more and more faster, just can match.And the fast routing lookups technology is to improve the key of router forwarding performance.
In IP route table, routing address is called as prefix, and P/L represents with symbol, and wherein P represents the route prefix or the network address, and L represents the length of this network address.The process of route querying is process of long matched and searched, and promptly address prefix is searched the bit that not only needs with prefix and carried out matched and searched, but also needs to consider the length of address prefix, selects in all occurrences that of the longest coupling.The longest matched and searched makes route querying become very complicated, becomes a key factor that improves router performance.Especially constantly soaring along with the core router interface rate, and the continuous increase of the forwarding-table item number of supporting, the routing table lookup technology has seriously restricted the further raising of router performance.
The appearance of Next Generation Internet network agreement (IPv6) makes fast routing lookups become difficult more.With the IPv4 ratio that adopts 32 bit address, IPv6 adopts 128 bit address, searches an IPv6 address prefix and need spend the longer time.Though the IPv6 agreement is in order just to consider this problem at the beginning of design, proposed to adopt the address allocation policy of hierarchical structure, dwindle means such as routing table by route assemblage, but following factor causes the IPv6 route querying can not utilize this level distribution mechanism to optimize fully: require address assignment and topology space to be consistent during (1) route assemblage, and the IPv6 address assignment is based on provider, can not guarantee this consistency, so the effect that converges is difficult to guarantee.(2) source routing and broadcast new business such as (AnyCast) arbitrarily and need the main frame route has for example increased the complexity of IPv6 routing table.(3) increase of the new business that may occur future can cause the very big increase of IPv6 address need, and it is complicated that the IPv6 routing table also can become.
There are many IP looks up table technique at present, the overwhelming majority proposes at the IPv4 look up table technique, do not make full use of the characteristics that the IPv6 prefix distributes, this makes them can not expand to 128 IPv6 space, location longways, so just can't be applied directly in the IPv6 route querying.Mainly contain following problem: (1) can not realize the high-speed searching of IPv6 routing table; (2) adopt complex data structures and calculating means, cause the time of routing table update very long, can not realize incremental update; (3) need bigger memory space.Though the TCAM technology is appropriate to the IPv6 route querying, the cost height of TCAM, power consumption is big, and integrated level is low, and that is that all right is ripe for its technology.And the IPv6 seek rate of TCAM also is subject to system bus bandwidth, and 0 (1) seek rate can not really be provided, and this is considerably beyond the highway width of existing system because of IPv6 routing table purpose width.
The most important thing is the route forwarding table structure Design in the IP route querying, the structures shape of transmitting route querying mode and routing update mode.Table 1 has shown the logical construction of route forwarding table, can be divided into two classes to method for searching route by the Storage Format of routing table: a kind of lookup method that is based on general-purpose storage, a kind of lookup method that is based on private memory.
The logical construction of table 1 route forwarding table
| Destination address prefix/length | Next hop address | |
| 10*/2 | H1 | P1 |
| 111*/3 | H2 | P2 |
(1) based on the lookup method of general-purpose storage
In Route Lookup Scheme based on common memory device, generally be that routing table is expressed as a kind of well-designed data structure, then this data structure that is used for searching for be put into SRAM (Static Random Access Memory) or SDRAM (Synchronous DynamicRandom Access Memory).The route search process realizes by the microprocessor executive software, and perhaps the IP that is realized by the hardware control logic special circuit of tabling look-up is realized.Software realization mode is more flexible, and hardware implementations can improve the throughput of route querying by modes such as streamlines.The route search structure has various ways, and wherein, a kind of multiple-limb Trie (Multi-bit trie) structure is introduced it below because of the characteristics of its simple and flexible are most widely used.
Trie is a kind of tree data structure.Fig. 1 (a) shows the address prefix table of representing with binary system trie structure (each inner node comprises two child nodes at most in the tree), and Fig. 1 (b) has shown 4 Trie of the branch tree of its correspondence.In multiple-limb Trie, some bit hytes are started anew to be divided in a route prefix address, and the value of these bit hytes is used to decide which branch that continues search tree successively.With the bit number that needs in searching each time to check be called search the step wide (stride).In fact two trie of branch tree is exactly that searching the step wide is 1 trie tree.Searching the step is wider than 1 trie tree and is called as multiple-limb trie tree (multibit trie), and for searching wide multiple-limb trie tree for k of step, the maximum branch number of each node is 2
kSearching wide can fixing of step, also can be variable.As the IP of 32 bits table look-up can adopt successively 10-10-8-4bit to search the step wide.
What Fig. 1 showed is the logical construction of Trie, and the physical structure of multiple-limb trie can be represented with Fig. 2, whole trie structure is made up of by pointer chain Trie node together some, and each trie node is made of one group of Trie entry array (201), each clauses and subclauses wherein comprise two pointers: the pointer (nt that points to next Trie node, 203) and point to the pointer (nh, 202) of next redirect photos and sending messages entry.Nt and nh can simultaneously effectively also may have only one effectively.When the two simultaneously effectively the time, represent that this entry correspondence a prefix, also have longer sub-prefix simultaneously under this prefix.When nt effectively and nh when invalid represents that this entry is an inner transition node; Otherwise then represent a longest-prefix on from the root of Trie to current clauses and subclauses path.
The IP process of tabling look-up is as described below: the top node by trie sets out, wide according to the step of setting, extract bit hyte index in the array of expression trie node of purpose IP address successively, determine the trie node that next step need be searched for, search procedure lasts till that always the node of least significant end or all bits of destination address use until exhausted.In search procedure, write down next jumping output port pointer (pn) of the route prefix node of last process, after search finished, this value was the result that IP tables look-up.
Adopt this structure, if the longest prefix length is W, search the wide k of being of step, worst case is searched next time needs the W/k memory access.Each step of multiple-limb trie tree search procedure needs to check a plurality of bits, so it can not support the address prefix of random length.For prefix length just is the IP prefix of the integral multiple of k, need not to carry out any change, but for prefix length is not the IP prefix of the integral multiple of k, then need this prefix is expanded, making the length of expansion route prefix afterwards is the integral multiple of k, this moment, a route prefix will copy in a plurality of storage item, and this can cause the memory space utilance to descend.Go on foot wide selection and played conclusive effect for the overall performance of multiple-limb trie tree, the bigger step is wide can improve searching speed, but also corresponding increase of storage demand of trie tree; Though the wide use that can reduce memory space of less step, seek rate has corresponding decline.A kind of more suitable scheme is exactly to select the suitable step wide according to the characteristic distributions of actual address prefix, thereby reaches the purpose of optimized Algorithm memory space.How some other just trade off on yardsticks such as the memory space of speed of route lookup, algorithm consumption and routing update complexity based on the lookup method of multiple-limb trie.
(2) based on the Route Lookup Scheme of private memory
Route Lookup Scheme based on private memory is that the primary routing tabulation is directly deposited, utilize hardware control logic to walk abreast then and search whole routing table, to reach the purpose of high-speed searching, the representative of this respect is TCAM (Ternary Content Addressable Memory).TCAM is development and come on the basis of CAM.Different with CAM, TCAM makes it can preserve the keyword list item of random length by the mode of preserving the keyword mask.This improvement has overcome CAM only can accurately mate the shortcoming of tabling look-up, and makes TCAM be fit to very much carry out searching of longest-prefix route.
In the IP look-up method based on SDRAM or SRAM structure, each IP tables look-up and needs to carry out repeatedly memory read operation, thereby has limited the table look-up raising of performance of IP, and relatively poor for the autgmentability of IPv6.And TCAM uses concurrent technique, has 0 (1) and searches complexity.So TCAM is the best solution that ultrahigh speed IP tables look-up, but it that is that all right is ripe, have following weak point: the first, compare with general random access memory ram, the TCAM of unit bit is more expensive, and the capacity of storage chip is smaller relatively; The second, because TCAM uses is the PARALLEL MATCHING manner of comparison, so the power consumption of TCAM chip is bigger.All keyword list items of search procedure all compare, but the actual keyword that can mate is several, and therefore most compare operation all has been wasted; Three, TCAM need guarantee that the long route of prefix is kept at before the short route of prefix, and this ordinal relation makes the routing update work of the TCAM relative complex that becomes.The TCAM technology also is in developing stage at present.
Summary of the invention
Therefore, the present invention proposes a kind of routing table of IPv6 fast and effectively address lookup method that can be used for core network, may further comprise the steps: a) entire I Pv6 routing table is divided at least two prefix subclass; B) the long prefix subclass of more or prefix length adopts concordance list to retrieve to number of prefixes, and prefix subclass that prefix length is short less to number of prefixes utilizes multiple-limb Trie tree to search; And c) in global scope, utilizes at least two microprocessor modules to come each part prefix subclass walked abreast and search, in subrange, utilize the route precomputation to search.
In addition, when each part is adopted multiple-limb Trie tree representation, adopt and search wide searching of step accordingly, when guaranteeing seek rate, to improve the memory space service efficiency of routing table.
The inventive method has taken into full account the distributed model characteristics of IPv6 address prefix, and has adopted quick indexing, routing table based on Hash table to cut apart and the technology such as search that walks abreast reaches raising speed and realizes the purpose of routing table incremental update.This method can very be supported the IPv6 route querying effectively, does not rely on TCAM, can realize on general hardware platform.
Description of drawings
Fig. 1 is the Trie structural representation of prior art.
Fig. 2 is the data structure of multiple-limb Trie.
Fig. 3 is the IPv6 prefix splitting scheme schematic diagram according to the inventive method.
Fig. 4 is the IPv6 route querying engine structure schematic diagram according to the inventive method.
Fig. 5 is that router data is transmitted structure.
Fig. 6 is a network processing unit basic structure.
Fig. 7 is the realization of IPv6 route querying engine in network processing unit.
Embodiment
Next, will be with reference to the accompanying drawings, embodiments of the invention are described.
(1) basic thought and data structure
The method of mentioning among the present invention adopts multiple-limb Trie tree as the basic structure of searching.But consider the characteristics that the IPv6 address is long and the IPv6 prefix distributes, earlier all prefixes are divided into several sections by length, for example, the part that prefix length can be higher than predetermined threshold is defined as the longer part of length, number of prefixes wherein number be can also be defined as than many parts more than the part of predetermined threshold, and the short part of length or number respectively remainder be defined as than small part.Implement different search strategies then to each part multiple-limb Trie tree representation, and in conjunction with its prefix characteristic distributions.Also the prefix to major part adopts Hash table to come quick indexing and parallel searching, to improve seek rate.
On basis, can obtain some characteristics that the IPv6 prefix distributes to the analysis of existing Ipv4 routing table and Ipv6 address allocation policy:
(1.1) the IPv6 prefix is mainly concentrated respectively on some length, on other length, then seldom has the IPv6 prefix;
(1.2) in long IPv6 prefix, middle very long one section bit is that 0 or 1 prefix accounts for significant proportion continuously.
In prefix division and search procedure, make full use of these characteristics that the IPv6 prefix distributes, entire I Pv6 prefix table has been divided into the individual prefix subclass of n (be greater than one positive integer), as shown in Figure 3, the characteristics of these prefix subclass can the divide into several classes, is described as follows: 306: length is 1 to n
1The prefix of position.The quantity of this a part of prefix can be few, even do not have.If exist, it be stored in inner buffer (Cache) lining of searching engine.307: length is n
1+ 1 to n
2The prefix of position.With a multiple-limb Trie tree representation (below be referred to as Trie1,301).This a part of prefix distributes more sparse, and small step is wide searches so adopt under the situation that seek rate allows as far as possible, to improve operating factor of memory space.
308: length is n
2+ 1 to n
3The prefix of position.With a quick indexing table (below be referred to as a Hash1,302) and a plurality of multiple-limb Trie tree representation (below be referred to as Trie2,303).This part is the major part of IPv6 route prefix, and prefix distributes can be than comparatively dense, so adopt bigger slightly searching to go on foot the wide seek rate that improves, it is many that the memory space utilance is descended.The quick indexing table can be a Hash table, and its each entry comprises the preceding n of route prefix
2Root address and the preceding n of the pairing Trie2 in position
2The longest prefix match information of bit address in P1 (below be referred to as bmp), bmp is used for shortening and searches the time.
309: length is n
3+ 1 to n
4The prefix of position.This a part of prefix distributes can be more sparse, and concentrate on the bottom of Trie tree, but consider the needs of seek rate, so it is wide to select bigger searching to go on foot.With a quick indexing table (below be referred to as a Hash2,304) and a plurality of multiple-limb Trie tree representation (below be referred to as Trie3,305).The quick indexing table is a Hash table, and each entry comprises the preceding n of route prefix
3The root address of the pairing Trie3 in position.
Preserve a longest-prefix precomputation information field in the quick indexing table that upgrades easily, its content is write routing table entry when route prefix upgrades, and this territory can dwindle seek scope rapidly.But in the route prefix association concordance list how this territory is not set, in order to avoid cause the routing table update difficulty, entry increases access cycle.
(2) search engine structure and route querying process
Fig. 4 shows basic structure of the present invention.The whole engine of searching is by m microprocessor module (401), and m route forwarding table (402) formed.Microprocessor internal provides buffer memory, Hash resource (403), memory unit access interface, and and the connecing of processor controls (405) etc.Also have SYN register (406) and lookup result selector (407) between the microprocessor in addition.M parallel the searching of processor, each microprocessor module is responsible for searching of several prefix subclass.What link to each other with each microprocessor module transmits the route forwarding table that memory is deposited its prefix subclass correspondence of being responsible for searching.Processor controls operation routing table management software is responsible for the renewal of route forwarding table.
The basic process of route querying is described below:
(2.1) purpose IP address is sent to all microprocessor modules;
(2.2) all microprocessor module executed in parallel are searched, and lookup result is delivered to selector.Each microprocessor module is carried out the quick indexing operation earlier, finds the root pointer of Trie tree, searches in the Trie of its sensing tree then.In the process of searching, the route prefix subset range that the microprocessor module that has utilizes the route precomputation to dwindle rapidly need to search.SYN register is responsible for the simultaneous operation between each processor module.
(2.3) selector is selected longest matching prefix from a plurality of lookup results as a result.
(2.4) according to the result of selector output as a result, visit next redirect photos and sending messages table
(408), obtain next final redirect and deliver information.
(3) routing table update
In the Internet network, topology is dynamic change, and route also is dynamically to change.Router is correct for the forwarding that guarantees to divide into groups, and must change route fast to make reflection, timely new route is joined routing table, and expired route is deleted from routing table.The processing speed of routing update becomes the important evaluation index of Route Lookup Scheme.Lookup method based on Trie all has " similar " these characteristics of routing update process and route querying process.And compare with basic multiple-limb Trie lookup method, the process of Hash table visit that this method is just many still can realize the incremental update of routing table.The present invention also reduces the complexity of routing update by following peculiar means, to guarantee to realize the purpose of quick routing update in high speed route lookup.
(3.1) prefix is divided into a plurality of parts, and long prefix utilizes the quick indexing table to come index, has shortened the time of upgrading in the Trie tree;
(3.2) take particular prefix length to be used as cut-point,, simplified the memory allocation management to guarantee the consistency of various storage unit space sizes;
(3.3) only in subrange, longest-prefix is carried out precomputation, can not have a significant impact update time route;
(3.4) only leave in the processor cache for considerably less ultrashort route prefix, do not influence the renewal of outside route forwarding table.
As shown in Figure 5, the IPv6 method for searching route table that provides among the present invention can be used in the forwarding engine design of High Speed IP v6 router.Because the flexibility of network processing unit, novel router uses network processing unit as forwarding engine usually.Describe below and in network processing unit, how to realize this method.
Network processing unit is the product that modern ISDN(Integrated Service Digital Network) network constantly develops renewal, be a kind of SOC (system on a chip) of new generation (System on Chip) chip, be specifically designed to the exploitation of network communication equipment based on programmable asic (Application-Specific IntegratedCircuit) structure.During the method that in network processing unit, realizes providing among the present invention, only require that network processing unit has following basic characteristics:
(1) the integrated high performance parallel processing architecture of having realized based on a plurality of micro engines and processor controls nuclear on single chip.
(2) support the multichannel memory interface, comprise SRAM and SDRAM.
As shown in Figure 6, existing diverse network processor substantially all possesses top characteristics, all can satisfy above-mentioned requirements as the IXP2400 and the IXP2800 of Intel Company.Can adopt following route to divide and search procedure:
At first, entire I Pv6 prefix table is divided into four parts (S0-S3), wherein: S0: length is 1 prefix to the F position, and F is the parameter that can regulate between 16 to 24.The quantity of this a part of prefix can be few, even do not have.If exist, it be stored in inner buffer (Cache) lining of searching engine.
S1: length is F+1 to 31 prefix.Begin to search than rising abruptly from the 1st of prefix, it searches the wide 6-6-6-6-7bit of being followed successively by of step.
S2: length is 32 to 64 prefix.With a quick indexing table (below be referred to as a Hash32) and a plurality of multiple-limb Trie tree representation (below be referred to as Trie64).Comprise longest-prefix precomputation information among the Hash32, searching of Trie64 goes on foot the wide 8-8-8-9bit of being followed successively by.
S3: length is 65 to 128 prefix.With a quick indexing table (below be referred to as a Hash64) and a plurality of multiple-limb Trie tree representation (below be referred to as Trie128).Searching of Trie128 goes on foot the wide 9-9-9-9-9-9-10bit of being followed successively by.
When P1 and P2 are divided, adopt the reason of 31 rather than 32 boundary to be: to allow the wide allocation strategy of step of Trie32 and Trie64 cooperate corrective measure to guarantee that the internal node of trie structure and the memory space that leaf node takies equate, help the distribution and the management of memory space to the trie structure.The time that worst case (Worst Case) can also be upgraded the hash32 table down shortens half.
A longest-prefix precomputation territory is arranged among the Hash32, and do not have among the Hash64.This is to consider the data-bus width problem of routing table update and processor and the choice done.It is to write routing table entry that longest-prefix precomputation territory is upgraded at route prefix, can dwindle seek scope rapidly, as long as find this territory, just needn't search in S1.If but in Hash64, increase this territory, and can cause Hash64 table entry too wide, processor needs the extra one-period that increases to visit a hash entry.Because the modification of any one the route prefix information between 32 to 64 all may cause the modification of the longest-prefix precomputation information in a plurality of Hash64 entries, this makes that the renewal of routing table is also very difficult simultaneously.And among the Hash32 because the entry width is smaller, and the prefix that will be shorter in length than F when upgrading has been put into the buffer memory of microprocessor, so just there are not the problems referred to above.
As shown in Figure 7, the present invention can utilize two micro engines of network processing unit to realize.Route forwarding table leaves in respectively among the outer SRAM (703,704) of two-way sheet, and micro engine inside provides buffer memory, also has mechanism such as SYN register and hash units in the network processing unit and is also utilized by this method just.Search algorithm by programming realizes to micro engine.Search for preceding 64 of the responsible IPv6 address of microprocessor Ma (701), Mb (702) is responsible for searching at 65 to 128, two parallel searching of processors.The route forwarding table that links to each other with Ma is stored in this whole route prefix and is shorter in length than 64 prefix; The route forwarding table that links to each other with Mb is stored the prefix of length above 64 in this whole route prefix.Processor controls operation routing table management software in the network processing unit is responsible for the renewal of route forwarding table.
The each several part route prefix to search performance as shown in table 2:
Table 2
| Prefix types | The maximum external memory access times | The memory location | Prefix quantity |
| P0 | 0 | Inner buffer | Seldom, almost do not have |
| | 1 Hash visit, the read-write of 5 external memories | The Ma external memory storage | |
| P2 | |||
| 1 Hash visit, 4 external memories | The Ma outside | At most |
| Read- | Memory | ||
| P3 | |||
| 1 Hash visit, the read-write of 7 external memories | The Mb external memory storage | Few |
SRAM can select the QDRAM of two 16 Mbytes for use, just can support up to a million routes.Suppose that its read and write access time is that (this is a conventional speeds to 5ns, also have faster), the width of data/address bus is 32, one time Hash lookup needs three SRAM visits, so in the method, a route querying at most only needs 10 SRAM visits, be to handle 20,000,000 messages each second to search, and only require at 10G linear speed IPv6 message and can handle 12.5 million messages (size of supposing message is 100 bytes) each second that the inventive method can satisfy the 10G surface speed forwarding requirement of message.In this programme, 1 routing update needs about No. 1000 holders visit at most, so if 1000 routing updates take place per second, only need 5ms to handle, the route querying performance also only descends 0.5%.
The present invention has provided a kind of method for searching route table of IPv6 fast and effectively, and its advantage is:
(1) this method has taken into full account the distributed model characteristics of IPv6 address prefix, can effectively support big capacity IPv6 routing table lookup.
(2) this method is supported the routing table incremental update, does not need extra hardware resource just can realize the quick renewal of routing table, has saved system cost.
(3) this method does not rely on the specific store device and realizes, has strengthened the flexibility that realizes.
(4) under the existing hardware condition, utilize this method, just can cost-effective realization IPv6 message in other surface speed forwarding of OC-192 level.
(5) this method is suitable for realizing in the diverse network processor, can be used for constructing effectively flexibly the IPv6 core router.
Claims (7)
1. IPv6 method for searching route table may further comprise the steps:
A) entire I Pv6 routing table is divided at least two prefix subclass;
B) the long prefix subclass of more or prefix length adopts concordance list to retrieve to number of prefixes, and prefix subclass that prefix length is short less to number of prefixes utilizes multiple-limb Trie tree to search; And
C) in global scope, utilize at least two microprocessor modules to come each part prefix subclass walked abreast and search, in subrange, utilize the route precomputation to search.
2. method according to claim 1 is characterized in that in described step a, and described route prefix subclass is left in respectively in the different memories.
3. method according to claim 1 is characterized in that in described step b, when each part is adopted multiple-limb Trie tree representation, adopts and searches wide searching of step accordingly.
4. method according to claim 1 is characterized in that among the described step c, only comprises route precomputation information in the short concordance list of prefix subclass length, realizes the routing table incremental update when searching.
5. method according to claim 1 is characterized in that in described step c, when densely distributed prefix subclass is searched, utilizes route precomputation information that seek scope is narrowed down to the regional area of this subclass, to reduce the parallel processing element number.
6. according to claim 1 or 4 described methods, it is characterized in that described concordance list is a Hash table.
7. method according to claim 1 or 5 is characterized in that utilizing many engine mechanism of network processing unit to realize parallel searching.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2004100099954A CN1787477A (en) | 2004-12-09 | 2004-12-09 | Method for searching IPv6 routing table |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2004100099954A CN1787477A (en) | 2004-12-09 | 2004-12-09 | Method for searching IPv6 routing table |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1787477A true CN1787477A (en) | 2006-06-14 |
Family
ID=36784793
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2004100099954A Pending CN1787477A (en) | 2004-12-09 | 2004-12-09 | Method for searching IPv6 routing table |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1787477A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102045412A (en) * | 2010-12-28 | 2011-05-04 | 赛尔网络有限公司 | Method and equipment for carrying out compressed storage on internet protocol version (IPv)6 address prefix |
| WO2011127642A1 (en) * | 2010-04-12 | 2011-10-20 | 华为技术有限公司 | Routing table construction method and device and routing table lookup method and device |
| CN102307149A (en) * | 2011-09-23 | 2012-01-04 | 中国科学院计算技术研究所 | IP (internet protocol) lookup method and device and route updating method and device |
| CN101631086B (en) * | 2009-08-10 | 2012-09-26 | 武汉烽火网络有限责任公司 | Routing list partitioning and placing method searched by parallel IP route |
| CN103092618A (en) * | 2013-01-15 | 2013-05-08 | 东南大学 | Dalvik virtual machine just-in-time compilation (JIT) acceleration method based on software Cache |
| CN103107945A (en) * | 2013-01-10 | 2013-05-15 | 中国科学院信息工程研究所 | System and method of quick searching Internet protocol version 6 (IPV6) route |
| WO2015103868A1 (en) * | 2014-01-09 | 2015-07-16 | 中兴通讯股份有限公司 | Route searching method and router |
| CN105141525A (en) * | 2015-06-30 | 2015-12-09 | 杭州华三通信技术有限公司 | IPv6 routing lookup method and IPv6 routing lookup device |
| CN105376159A (en) * | 2014-08-25 | 2016-03-02 | 深圳市中兴微电子技术有限公司 | Packet processing and forwarding device and method |
| CN106302172A (en) * | 2015-05-18 | 2017-01-04 | 深圳市中兴微电子技术有限公司 | Support Hash lookup and the storage of route querying, lookup method and device simultaneously |
| CN106330721A (en) * | 2015-06-30 | 2017-01-11 | 杭州华三通信技术有限公司 | IP route search method and device |
| CN107154899A (en) * | 2017-03-23 | 2017-09-12 | 广东顺德中山大学卡内基梅隆大学国际联合研究院 | A kind of system that IP routes are searched with suffix index |
| CN109039911A (en) * | 2018-07-27 | 2018-12-18 | 烽火通信科技股份有限公司 | It is a kind of to search the method and system that mode shares RAM based on HASH |
| CN109327396A (en) * | 2018-12-03 | 2019-02-12 | 中国人民解放军陆军工程大学 | Hash table gradual updating method in routing table look-up circuit |
| CN113343034A (en) * | 2021-06-08 | 2021-09-03 | 湖南大学 | IP searching method, system and storage medium |
| CN115086221A (en) * | 2022-07-27 | 2022-09-20 | 新华三半导体技术有限公司 | Message processing method, device, forwarding equipment and storage medium |
-
2004
- 2004-12-09 CN CNA2004100099954A patent/CN1787477A/en active Pending
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101631086B (en) * | 2009-08-10 | 2012-09-26 | 武汉烽火网络有限责任公司 | Routing list partitioning and placing method searched by parallel IP route |
| WO2011127642A1 (en) * | 2010-04-12 | 2011-10-20 | 华为技术有限公司 | Routing table construction method and device and routing table lookup method and device |
| CN102484610A (en) * | 2010-04-12 | 2012-05-30 | 华为技术有限公司 | Routing table construction method and device and routing table lookup method and device |
| CN102045412B (en) * | 2010-12-28 | 2013-01-16 | 赛尔网络有限公司 | Method and equipment for carrying out compressed storage on internet protocol version (IPv)6 address prefix |
| CN102045412A (en) * | 2010-12-28 | 2011-05-04 | 赛尔网络有限公司 | Method and equipment for carrying out compressed storage on internet protocol version (IPv)6 address prefix |
| CN102307149A (en) * | 2011-09-23 | 2012-01-04 | 中国科学院计算技术研究所 | IP (internet protocol) lookup method and device and route updating method and device |
| CN102307149B (en) * | 2011-09-23 | 2014-05-07 | 中国科学院计算技术研究所 | IP (internet protocol) lookup method and device and route updating method and device |
| CN103107945B (en) * | 2013-01-10 | 2016-01-27 | 中国科学院信息工程研究所 | A kind of system and method for fast finding IPV6 route |
| CN103107945A (en) * | 2013-01-10 | 2013-05-15 | 中国科学院信息工程研究所 | System and method of quick searching Internet protocol version 6 (IPV6) route |
| CN103092618A (en) * | 2013-01-15 | 2013-05-08 | 东南大学 | Dalvik virtual machine just-in-time compilation (JIT) acceleration method based on software Cache |
| WO2015103868A1 (en) * | 2014-01-09 | 2015-07-16 | 中兴通讯股份有限公司 | Route searching method and router |
| CN105376159A (en) * | 2014-08-25 | 2016-03-02 | 深圳市中兴微电子技术有限公司 | Packet processing and forwarding device and method |
| CN106302172A (en) * | 2015-05-18 | 2017-01-04 | 深圳市中兴微电子技术有限公司 | Support Hash lookup and the storage of route querying, lookup method and device simultaneously |
| CN106330721B (en) * | 2015-06-30 | 2019-09-17 | 新华三技术有限公司 | IP method for searching route and device |
| CN106330721A (en) * | 2015-06-30 | 2017-01-11 | 杭州华三通信技术有限公司 | IP route search method and device |
| CN105141525A (en) * | 2015-06-30 | 2015-12-09 | 杭州华三通信技术有限公司 | IPv6 routing lookup method and IPv6 routing lookup device |
| CN105141525B (en) * | 2015-06-30 | 2018-08-10 | 新华三技术有限公司 | IPv6 method for searching route and device |
| US10476785B2 (en) | 2015-06-30 | 2019-11-12 | New H3C Technologies Co., Ltd | IP routing search |
| CN107154899B (en) * | 2017-03-23 | 2021-01-08 | 广东顺德中山大学卡内基梅隆大学国际联合研究院 | System for searching IP route by suffix index |
| CN107154899A (en) * | 2017-03-23 | 2017-09-12 | 广东顺德中山大学卡内基梅隆大学国际联合研究院 | A kind of system that IP routes are searched with suffix index |
| CN109039911A (en) * | 2018-07-27 | 2018-12-18 | 烽火通信科技股份有限公司 | It is a kind of to search the method and system that mode shares RAM based on HASH |
| CN109327396A (en) * | 2018-12-03 | 2019-02-12 | 中国人民解放军陆军工程大学 | Hash table gradual updating method in routing table look-up circuit |
| CN113343034A (en) * | 2021-06-08 | 2021-09-03 | 湖南大学 | IP searching method, system and storage medium |
| CN115086221A (en) * | 2022-07-27 | 2022-09-20 | 新华三半导体技术有限公司 | Message processing method, device, forwarding equipment and storage medium |
| CN115086221B (en) * | 2022-07-27 | 2022-11-22 | 新华三半导体技术有限公司 | Message processing method, device, forwarding equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1787477A (en) | Method for searching IPv6 routing table | |
| Bando et al. | Flashtrie: Hash-based prefix-compressed trie for IP route lookup beyond 100Gbps | |
| US6778532B1 (en) | Packet relaying apparatus and high speed multicast system | |
| US8625604B2 (en) | Hash-based prefix-compressed trie for IP route lookup | |
| Bando et al. | FlashTrie: beyond 100-Gb/s IP route lookup using hash-based prefix-compressed trie | |
| US8295286B2 (en) | Apparatus and method using hashing for efficiently implementing an IP lookup solution in hardware | |
| CN1859316A (en) | Packet transfer apparatus | |
| US7356033B2 (en) | Method and apparatus for performing network routing with use of power efficient TCAM-based forwarding engine architectures | |
| Waldvogel et al. | Scalable high-speed prefix matching | |
| CN1462004A (en) | Method and device for producing and using improved tree-shape bit map data structure | |
| Le et al. | Scalable tree-based architectures for IPv4/v6 lookup using prefix partitioning | |
| US7526603B1 (en) | High-speed low-power CAM-based search engine | |
| CN1270728A (en) | Method and system for fast routing lookups | |
| CN1216473C (en) | TCAM high-speed search method and system supporting multiple nexthop | |
| CN1655534A (en) | Double stack compatible router searching device supporting access control listing function on core routers | |
| Le et al. | Memory-efficient and scalable virtual routers using FPGA | |
| Jiang et al. | A memory-balanced linear pipeline architecture for trie-based IP lookup | |
| CN1561047A (en) | Distributed paralled IP route searching method based on TCAM | |
| CN106416152A (en) | Lookup device, lookup configuration method and lookup method | |
| Le et al. | Scalable high throughput and power efficient ip-lookup on fpga | |
| CN1191520C (en) | TCAM high speed updating method supporting route compress | |
| CN1279716C (en) | Parallel route searching method and system for eliminating longest prefix match search | |
| US20030236968A1 (en) | Method and apparatus for generating efficient data structures for use in pipelined forwarding engines | |
| CN104301227B (en) | High-speed low-power-consumption IP route table lookup method based on TCAM | |
| CN1571352A (en) | Method for processing five-membered stream group in network 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 | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20060614 |