KR20020081569A - The Management Method of tables that have entries of same size on CAM in Network device - Google Patents

Abstract

PURPOSE: A method for managing plural of tables having identical entry size in a CAM(Content Addressable Memory) of a network device using a ternary CAM is provided to locate a table separator capable of separating tables having identical entry size in a table entry, so as to search the tables even for the identical entry size. CONSTITUTION: Entry size of a table to make is defined(401). The size of a CAM entry is defined to be larger than the entry size of the table(402). A difference between data tuple size of a table having the biggest data tuple among different tables having identical size of CAM entry and the size of the CAM entry is regarded as the size of a table separator(403). Whether the size of the table separator is larger than 0(zero) is decided(404). If the size of the table separator is larger than 0, an actual separator value is set within the size of the table separator(405). If the size of the table separator is smaller than 0, the step 402 is returned. Accordingly, information on the table separator is added on a CAM table(406).

Description

{The Management Method of tables that have entries of same size on CAM in Network device} in a network device using Ternary Content Addressable Memory (CAM)

The present invention relates to a network device using a CAM (Content Addressable Memory) as a table for retrieval, and more particularly to a method for more efficiently using tables managed in a CAM.

Conventional techniques of the present invention include a CAM method and a ternary CAM method.

The CAM method has the advantage of finding a particular entry in the CAM within a hardware-specified time, i.e., a fixed number of clock cycles.

Data in the table, such as MAC address 130, IP address 140, {sending IP address, sending port number, receiving IP address, receiving port number} 160, as in a table structure managed within a network environment and network devices. After storing a tuple of fields in the CAM, the CAM is searched and found, and the result is used for packet transmission and reception.

The ternary CAM method enables not only the specific fields in the entries that do not need to be compared to the don't care by making it possible to have "don't care" values as well as the 0 and 1 values of the existing binary CAM. This makes it much easier to apply CAM to a variety of applications. CAM has emerged as a secondary memory that is indispensable for next-generation network devices that demand high speed.

Conventional patents related to the present invention are the presently filed "Dynamic Content Addressable Memory (CAM) Cells (Application No. 10-2001-712599, Publication No. 2002-0012168, Publication Date February 15, 2002)." This patent is not about the application of CAM but rather a patent on CAM's own functions.

The disadvantage of the conventional CAM is that the entry size is fixed, such as 36bit, 72bit, 144bit, 288bit, etc. due to the characteristics of the CAM, and there is no way of searching separately between tables having the same size. Because a table is logically divided from the CAM user's point of view, and not divided in hardware, the CAM itself does not search only the table to be searched. Instead, it searches all entries of the same size in the CAM according to the size of the search entry. Regardless of the type of search, the present invention produced the effect.

In the embodiment of the tables in the CAM, if the user issues a search command to search only in the A table 211 in the CAM memory 210, there is a B table 212 having the same entry size in the CAM. Will be searched at the same time and the first searched in the two tables will result in matching entry values. Thus, if a match occurs first in the A table, a desired result value may be obtained. However, if a match occurs first in the B table 212, a situation occurs in which a desired result is not obtained.

It is an object of the present invention to invent a method for adding a CAM entry that can overcome the prior art, which has the disadvantage of not being able to search for a network device constituting various tables in the CAM having the same CAM entry size. According to the present invention, a table having an entry of the same size can be searched by overcoming a disadvantage that the CAM method cannot search.

Figure 1 Table structure managed within network environment and network device

Figure 2 An embodiment of the tables falling within the CAM

Figure 3 Packet forwarding method in a network device

Fig. 4 How to add an entry including table separator information in CAM table

Figure 5. How to Set Table Separator Values

Fig. 6 Entry structure added by including table separator information in CAM table

7 Example of Setting a Table Separator Value

Fig. 8 Entry Search Method in CAM Table

Referring to the configuration of the present invention by the accompanying drawings as follows.

The table structure managed in the network environment and the network device has a MAC address 130 and an IP address 140 table. These tables exist in the switch router 120 and perform data transmission via the Internet 110 or data transmission with the computers 121, 122, and 123 connected to the switch router 120. Information for transmitting data from the source to the destination consists of a sending IP address 150, a sending port number 160, a receiving IP address 170, a receiving port number 180, and the like. A tuple of data fields in a table with this information is stored in the CAM, and then the CAM is searched for and used to properly transmit and receive packets.

FIG. 2 shows an example of tables in the CAM. When a user issues a search command to search only in Table A 211 having a 36-bit entry in the CAM memory 210, Table A having the same entry size in the CAM. Since there are 211 and Table B 212, Table A 211 and Table B 212 will be searched at the same time and the first searched in both tables will result in matching entry values. Therefore, if the desired entry is in the A table 211, if a match occurs first in Table A 211, a desired result value may be obtained. However, if a match occurs first in Table B 212, the desired result may not be obtained. Problem occurs.

In the case of Table C (213) having a 72-bit entry, there is no problem because only one is always selected because there is only one same size.

In the case of Table D (214), Table E (215), and Table F (216) with 144bit-size entry, there are three tables of the same size in CAM, so Table D, Table E, and Table F are searched simultaneously and three The first searched entry in the table will result in the matched entry value. Thus, if the desired entry is in the F table 216, if the match occurs first in Table F 216, the desired result can be obtained, but if the first match occurs in Table D (214) or Table E (215). The problem arises that the desired result is not achieved.

FIG. 3 illustrates a method of packet forwarding within a network device. In the table initialization 310, a CAM entry is added 311, a data packet is received 320, a table search tuple is created 330, and a table is searched ( 340), the table search result analysis 350. In the table search result analysis 350, the CAM entry is added 351, and the packet is transmitted or discarded to the corresponding port. When the series of procedures is completed as described above, the process moves to the data packet reception 320.

In other words, we first perform a basic initialization routine for a table to be in the CAM. You can also create the tables you need at this time. It then receives the data packet, which determines whether the received packet is valid and to which port to forward it. This is the table in the CAM. When initializing a table (310) and analyzing a search result of a table (350), there is an item called 'Add CAM entry' (311, 351). When adding an entry, a plurality of tables having the same CAM entry size can be distinguished. Adding a table delimiter as a field in a table is the subject of this invention.

FIG. 4 illustrates a method for adding an entry including table separator information to a CAM table in FIG. 3. A method of adding CAM entries 311 and 351, which is not available in the prior art, is described in FIG. First, an entry size value of a table to be created is defined (401). A CAM entry size value larger than the entry size value of the table is determined (402). The difference between the data tuple size value and the CAM entry size value of the table having the largest data tuple among other tables with the same size CAM entry becomes the size of the table separator (403), and the table separator size is compared (404). If the table delimiter size is greater than zero, set the actual delimiter value within the table delimiter size (405), compare the table delimiter size (404), and if the table delimiter size is less than zero, select CAM entry size selection (402) again. Run

5 illustrates a table separator value setting method 405 in detail. The table delimiter value can range from 0 to (2 ^{table delimiter size} -1) (501), which also means the number of possible tables with entries of the same size. Set the first table value in the table (502). After initializing the delimiter value of the first table to 0 (503), if the delimiter value is compared (504) and smaller than 2 ^{(table delimiter size)} , the next table having the same size entry in the table is set (505). If the existing table is present (506), or terminates (508), and if the existing table, the delimiter value of the table is increased by one to set the delimiter value (507). If the delimiter value is compared (504) to be greater than or equal to 2 ^{(table delimiter size)} , the method of obtaining the table delimiter size in FIG. 4 is performed again (401) so that a larger delimiter value can be allocated. . The table delimiter value set in this way is created together when creating an entry in the table, and becomes one entry (406).

FIG. 6 shows an entry structure added to the CAM table including the table separator information by 72 bits and 144 bits. In the case of having a 48-bit search data tuple in a 72-bit CAM entry, there is a table separator 601 having a maximum size of 24 bits. If the size of the data tuple to be searched is 112 bits and the CAM can support 144 bits, the reserved identifier field 603 is added as follows when the table separator value of another table having the same size of 144 bits is 16 bits. The table delimiter value is set to 16 bits (602).

7 is an embodiment of a method of setting a table separator value. Tables A and B are tables having 36-bit CAM entries, which use 0 value 701 and 1 value 702 as table delimiter values, respectively. The table separator is 6 bits in size. Tables D, E, and F are tables with 144-bit CAM entries, which use 0 value 704, 1 value 705, and 2 value 706 as table delimiter values, respectively. The table separator is 8 bits in size.

By creating a table with table separators within the table entries, the search results can be obtained only from the target table even if two or more tables with the same size CAM entries exist in the CAM.

In the searching method, as shown in FIG. 8, a table search tuple is first created (801) with a table separator, a table separator for searching is added (802), and a CAM search is performed (803). It determines whether there is a matched entry (804), and if there is no entry to search for in the table (806), and if it finds in the table (805) in the table separator. The result thus obtained determines whether the packet is discarded at the end of FIG. 3 or transmitted to the corresponding destination (360).

As described above, in the present invention, a table delimiter capable of distinguishing a table having the same entry size may be placed in a table entry, thereby overcoming the shortcomings that the CAM method cannot search in the case of having an entry of the same size.

A disadvantage of the conventional technology, CAM, is that there is no way to search separately among tables having entries of the same size, such that the entry size is fixed, such as 36 bits, 72 bits, 144 bits, 288 bits. In order to solve this drawback, the effect of the invention that only the table to be searched when searching a plurality of tables has occurred.

Claims (3)

It has a table structure such as MAC address 130 and IP address 140 managed in the network environment and network devices, these tables exist in the switch router 120, and transmits data through the Internet 110, Data for transmitting data with the computers 121, 122, and 123 connected to the switch router 120, and information for transmitting data from the source to the destination may include a source IP address 150, a source port number 160, a destination IP address ( 170), the receiving port number 180, and the like, In a network device using Ternary CAM, if there are tables having CAM entries of the same size, a table delimiter for distinguishing the tables may be added as a field in the table. The packet forwarding method in the network device includes adding (311) a CAM entry in a table initialization (310), receiving a data packet (32), and writing a tuple for a table search (330), and then searching (340) the table. In the search result analysis 350, the CAM entry is added 351 in the table search result analysis 350, the packet is transmitted or discarded to the corresponding port, and the data packet is received 320 when the series of procedures is completed. In moving to, When initializing the table in CAM (310) and analyzing the search results of the table (350), there is an item called 'Add CAM Entry' (311, 351). To Add Table Separator as a Field in a Table The method of claim 1, Define an entry size value of the table to be created (401), define a CAM entry size value larger than the entry size value of the table (402), and have the table with the largest data tuple among other tables with the same size CAM entry. If the difference between the data tuple size value of the CAM entry size value is the table separator size (403) and the table separator size is compared (404), and the table separator size is greater than zero, the actual separator value within the table separator size is selected. Setting (405) and comparing the table separator size (404), if the table separator size is less than zero, execute the CAM entry size selection (402) again to add an entry including the table separator information to the CAM table. The method of claim 2, Table delimiter values range from 0 to (2 ^{table delimiter sizes} -1), After setting the first table value to the table (502), initializing the separator value of the first table to zero (503), and comparing the separator value (504) to less than 2 ^{(table separator size)} , The next table having the same size entry is set (505) and the existing table is executed (506), or terminated (508). If the existing table, the delimiter value is set by increasing the delimiter value of the corresponding table by one. If the delimiter value is greater than or equal to 2 ^{(table delimiter size)} , the method for obtaining the table delimiter size is performed again (401) so that a larger delimiter value can be allocated. To set the table delimiter value so that the set table delimiter value is created together when creating an entry in the table so that it becomes one entry 406.