JPH1188427A - Routing table retrieving device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a reouting table retrieving device by which a routing table retrieval processing is executed at high speed and also a retrieval error due to that a sub-net mask is provided with variable length is prevented. SOLUTION: This device is provided with a routing table storing memory 3 for storing a routing table 2 consisting of plural sorted entries, a matching comparing part 13 for comparing the sizes of a retrieval IP address and a register IP address without executing a sub-net mask in the retrieval IP address, a matching comparing part 12 for respectively executing the AND operations of one register IP address and the variable length sub-net mask and the other register IP address and the variable length sub-net mask in two entries which directly pinch the retrieval IP address by bit, and a control part 14 for retrieving the register IP address matched with the retrieval IP address.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for receiving an IP (Internet Protocol) packet and transferring data or the like in an inter-computer communication. Is applied to a router that outputs an IP packet to a route connected to a destination.
The present invention relates to a routing table search device used in destination information search using a P address as a key.

[0002]

2. Description of the Related Art Conventionally, in computer-to-computer communication, a device called a router has been used to transfer an IP packet to a destination registered in advance. This router searches a routing table in which a plurality of registered IP addresses to be registered are registered by a microprocessor and software executed by the microprocessor based on an IP address described in an input IP packet. Then, the IP packet is output to a route connected to the destination based on the transfer destination information obtained therefrom.

The search algorithm performed by the above software has the best agreement, such as a tree algorithm or a binary search, as introduced in, for example, the document "Interconnections-Bridges and Routers" (by Radis Perlman, translated by Akira Kato). This is to find a registered IP address having the largest number of matching bits in the routing table.

[0004] As another router having a different configuration, an input IP packet (hereinafter referred to as an input IP packet) is used.
Is searched for a routing table having a plurality of pre-registered IP addresses (hereinafter, referred to as registered IP addresses) using the IP address obtained from the IP address as a search key, and a destination is connected based on the transfer destination information obtained from the routing table. There is a device that outputs an IP packet to a route.

Here, the input and registered IP addresses are, as is well known, a network address indicating a network address, a subnet address indicating an address of a subnet existing in the network, and a host connected to the subnet. The address is represented by 32 bits, which is composed of a host address indicating the address.

Further, a router uses a routing table search device having a function of searching a routing table stored in a storage unit in advance at a high speed. Here, the routing table includes a registered IP address (32 bits), a subnet mask (32 bits),
When the destination information of the IP packet is put together into one entry, it is composed of a plurality of entries.

In the routing table,
Each entry is sorted in ascending order (or descending order) by regarding a plurality of registered IP addresses as numerical values. Registration I above
In the P address, the portion of the host address is all “0”. Here, in the subnet mask, the bit corresponding to the network address and the subnet mask in the IP address obtained from the input IP packet described above is set to “1”, and the bit corresponding to the host address is set to “0”. ing.
In the subnet mask, the length of the “1” bit portion is fixed.

In the above configuration, when an input IP packet is input to the routing table search device, the routing table search device converts the IP address and subnet mask obtained from the input IP packet into bits for each bit.
Perform ND operation. The operation result (hereinafter, referred to as an input IP address) is an address other than the host address in the IP address obtained from the input IP packet, that is, a network address and a subnet address.

[0009] Next, the routing table search device,
In the routing table, an address of an entry (hereinafter, referred to as a first center entry) located at the center (hereinafter, referred to as a first center) between the head entry and the tail entry is obtained. Then, the routing table search apparatus registers the input IP address and the registration I of the first central entry.
The P address is collated and various processes are performed based on the collation result.

That is, if the input IP address and the registered IP address of the first central entry match, the routing table search device determines that the matching result matches, and then determines the transmission corresponding to the registered IP address. The destination information is output to the corresponding route.

On the other hand, if the input IP address is smaller than the registered IP address, the routing table search apparatus sets the first center as a new tail entry and sets the center between the tail entry and the head entry (hereinafter, the second center). ) (Hereinafter referred to as a second central entry). Thereafter, the routing table search device performs the input I
The P address is collated with the registered IP address of the second central entry, and various processes are performed based on the collation result. Then, the input IP address and the registered IP of the second central entry
If the address does not match, the third central entry is again sought. That is, in the matching process, the search range in the routing table is sequentially narrowed down to half.

On the other hand, if the input IP address is larger than the registered IP address, the routing table search device sets the first center as a new head entry and sets the center between the head entry and the tail entry (hereinafter referred to as a fourth center). ) (Hereinafter referred to as a fourth central entry). Thereafter, the routing table search device performs the input I
The P address is collated with the registered IP address of the fourth central entry, and various processes are performed based on the collation result. Then, the input IP address and the registered IP of the fourth central entry
If the address does not match, the fifth center entry is determined again. That is, in the matching process, the search range in the routing table is sequentially narrowed down to half.

[0013]

By the way, in the above-mentioned conventional software router, a process of searching for transfer destination information of an IP packet using an IP address as a key is executed by software on a microprocessor. Naturally, the processing capacity was limited. Therefore, the above-mentioned conventional router has a problem that it is difficult to receive a large amount of input IP packets at high speed and to output the IP packets to a route to a destination instantly.

On the other hand, since the conventional routing table search apparatus uses a fixed-length subnet mask, if the subnet mask is of variable length,
There is a disadvantage that a search error occurs for the following reasons. Now, the routing table has at least the following 2
It is assumed that there are two entries A and B corresponding to one subnet NA and NB, and each entry is composed of a registered IP address and a variable-length subnet mask.・ Entry A (subnet NA) → registered IP address AA
(32 bits), subnet mask SA of first length (20 bits) Entry B (subnet NB) → registered IP address AB
(32 bits), subnet mask SB of second length (24 bits)

Here, the relationship between the sizes of the respective network spaces of the subnets NA and NB is as follows.
A> Subnet NB. Also, subnet NB
Is included in the subnet NA. Therefore, the 20 bits of the registered IP address AB correspond to the registered I
It is the same as the 20 bits of the P address AA.

Under the above conditions, if the registered IP address AB in the routing table is to be searched,
It is assumed that an input IP packet has been input to the routing table search device. Thus, the routing table search device selects one entry to be searched from the routing table by the above-described method. In this case, it is assumed that the entry B is selected as the entry to be searched.

Therefore, the routing table search device firstly obtains the IP address obtained from the IP packet,
An AND operation is performed on the subnet mask SB (24 bits) corresponding to the entry B for each bit. As a result, a network address and a subnet address (input IP address) each consisting of 24 bits are extracted from the IP address.

Next, the routing table search device:
The registered IP address corresponding to the entry B is compared with the input IP address. In this case, since the registered IP address matches the input IP address, the search was correctly performed.

On the other hand, as another example, when the registered IP address AB in the routing table is a search target, it is assumed that the entry A is selected as the search target entry before the entry B described above. I do.

As a result, the routing table search device firstly obtains the IP address obtained from the IP packet and the subnet mask SA (20) corresponding to the entry A.
Bit) and AND operation for each bit. This allows
From the IP address, a network address and a subnet address (input IP address) each consisting of 20 bits are extracted.

Next, the routing table search device:
The registered IP address corresponding to the entry A is compared with the input IP address. Here, although the search target is the registered IP address corresponding to the entry B, the registered IP address corresponding to the entry A and the input IP address apparently match as described above. For this reason, the routing table search apparatus determines that the matching result is a match, and in this search, a search error occurs that the search result is regarded as entry A regardless of whether the search result is originally entry B. I have. Such a search error occurs firstly because the subnet mask has a variable length, and secondly, when the entry A is selected before the entry B.

The present invention has been made under such a background, and is capable of performing a routing table search process at a high speed, receiving a large amount of input IP packets at a high speed, and immediately connecting to a destination. It is an object of the present invention to provide a routing table search device capable of outputting an IP packet to a path and preventing a search error caused by a variable length subnet mask.

[0023]

According to the first aspect of the present invention, all bits of each host address portion of a plurality of registered IP addresses to be registered are set to "0", and the registration I
After setting the bit of each subnet mask portion corresponding to the P address to “1”, the one registered IP address, the subnet mask, and the transmission destination information are combined into one.
After making the entry data, there is further provided a routing table generating means for generating a routing table composed of a plurality of entries by sorting the registered IP addresses as numerical values, and further comprising a search IP address read from the received IP packet; In a routing table search device that performs a search by narrowing the search range in a routing table by half by comparing the size of the search IP address with the registered IP address, the search IP address and the registration are applied without applying a subnet mask to the search IP address. The comparison with the IP address is performed, and the search I
Comparing means for directly searching for two entries in the routing table, directly interposing the P address;
In one entry, the result of AND operation of one registered IP address and the variable length subnet mask for each bit, and the result of AND operation of the other registered IP address and the variable length subnet mask for each bit And searching means for searching the routing table for a registered IP address that matches the search IP address based on a result of the comparison. Further, according to a second aspect of the present invention, in the routing table search device according to the first aspect, the routing table generating means replaces all bits of a host address portion of each registered IP address in the routing table with the "". "1" instead of "0", and the search means replaces one of the registered IP address and the inverted bit of the variable length subnet mask ("0" and "1") in the two entries. ) For each bit, and the other registered IP address and an inverted bit (“0”) of the variable-length subnet mask
Based on the result of comparison between the result of the OR operation for each bit and the result obtained by replacing “1” with “1”, the registration I that matches the search IP address in the routing table is determined.
It is characterized by searching for a P address. According to a third aspect of the present invention, all bits of each host address portion of a plurality of registered IP addresses to be registered are set to “0”,
After setting the bit of each subnet mask portion corresponding to the registered IP address to “1”, one of the registered IP addresses
A routing table generating unit that collects the address, the subnet mask, and the transmission destination information into one entry data, and sorts the registered IP address as a numerical value to generate a routing table including a plurality of entries. A routing table search device for performing a search by narrowing the search range in the routing table by half by comparing the search IP address read from the received IP packet with the registered IP address of the entry, Without applying a subnet mask to the
A comparison unit that performs the magnitude comparison between the registered IP address and the registered IP address and directly interposes the search IP address to search for two entries in the routing table; and a variable length of one of the two entries in the two entries. The length of “1” of the subnet mask and the variable length “1” of the subnet mask of the other entry
Registered IP address matching the search IP address in the routing table based on the comparison result
A search unit for searching for an address.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION

<First Embodiment> An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the routing table search device according to the first embodiment of the present invention. In this figure, reference numeral 1 denotes a routing table generator for generating the same routing table 2 as described above in accordance with the flowcharts shown in FIGS.

The routing table 2 includes a plurality of entries when the IP address 7 to be registered, the subnet mask 8 and the destination information 6 of the IP packet are grouped into one entry. Each entry is sorted in ascending order (or descending order) by regarding 7 as a numerical value. Here, in the routing table 2, the length of the subnet mask for each entry is variable.

However, the routing table 2 generated by the routing table generation unit 1 includes
Entry corresponding to address 7 (original entry)
In addition, a dummy entry is included. This dummy entry is composed of a dummy IP address, a dummy subnet mask, dummy destination information, and a dummy identifier 17 which is a flag indicating a dummy entry. The process of generating the routing table 2 by the routing table generator 1 and the details of the routing table 2 and the dummy entry will be described later.

Reference numeral 3 denotes a routing table storage memory that stores (stores) the routing table 2. 1
Reference numeral 1 denotes an address generation unit that generates a memory address of the routing table storage memory 3 corresponding to an address generation instruction 16 input from the control unit 14 described below. The memory address corresponds to an entry in the routing table.

Reference numeral 5 denotes a search IP address holding unit that holds an IP address obtained from an input IP packet (hereinafter, referred to as a search IP address). 9 is search I
This is a size comparison unit that compares the magnitude relationship between the search IP address input from the P address holding unit 5 and the IP address 7 input from the routing table storage memory 3. A dummy check unit 10 checks whether an entry in the routing table 2 is an original entry or a dummy entry. If the entry includes the dummy identifier 17, the entry is determined to be a dummy entry.

Numeral 12 denotes a match comparison unit, which is a bit-wise AND operation of the search IP address input from the search IP address holding unit 5 and the subnet mask 8 (hereinafter referred to as an input IP address), and a routing. A comparison is made with the IP address 7 input from the table storage memory 3 to determine whether or not they match. 13
Is a match comparison unit that compares the search IP address input from the search IP address holding unit 5 with the IP address 7 input from the routing table storage memory 3 and determines whether they match. judge.

That is, while the match comparison unit 12 compares the search IP address with the subnet mask, the match comparison unit 13 sets the search IP address without the subnet mask as the comparison target. This is different from the function of the match comparison unit 12 in that

The control unit 14 outputs the search result of the routing table 2 as search result data 15 in accordance with the procedure of the flowchart shown in FIG. 4 and the search result output table shown in FIG. Details of the operation of the control unit 14 will be described later. Reference numeral 18 denotes a work memory that stores data of various calculations performed by the control unit 14.

Next, the operation of the routing table search device according to the first embodiment will be described with reference to FIGS. <Routing Table Generation Operation> First, the routing table generation operation performed by the routing table generation unit 1 will be described with reference to FIGS. In step SA1 shown in FIG. 2, the routing table generating unit 1 regards each IP address (destination network address) of a plurality of entries as a numerical value, sorts the entries in ascending order (or descending order), and then proceeds to step SA2. Hereinafter, the case of sorting in ascending order will be described.

Here, each entry is composed of an IP address (destination network address), a subnet mask, and transmission destination information. Also,
The IP (destination network address) includes a network address, a subnet address, and a host address, and the host address is all “0”. Further, in the subnet mask, bits corresponding to the network address and the subnet mask are all “1”, and bits corresponding to the host address are all “0”.

In step SA 2, if there are a plurality of entries having the same destination network address in the routing table sorted in ascending order, the routing table generating unit 1 regards each subnet mask as a numerical value, and Are further sorted in ascending order, and the process proceeds to step SA3.

At step SA3, the routing table generator 1 stores the destination network address (IP address), subnet mask and destination information in the first entry of the routing table in the memory d1 (not shown), and then returns to step SA4.
Proceed to. Hereinafter, the destination network address and the subnet mask stored in the memory d1 are abbreviated as “d1.IP” and “d1.subnetmask”, respectively.

At step SA4, the routing table generator 1 sets the destination network address (IP) in the entry next to the first entry of the routing table (hereinafter referred to as the second entry).
After storing the address), the subnet mask, and the destination information in the memory d2 (not shown), the process proceeds to step SA5. Hereinafter, the destination network address and the subnet mask stored in the memory d2 are abbreviated as “d2.IP” and “d2.subnetmask”, respectively.

At step SA5, the routing table generator 1 determines whether or not there is an entry not yet stored in the memory d1 or the memory d2 among a plurality of entries of the routing table. In this case,
It is assumed that there is an entry that is not stored in the memory d1 or the memory d2, and the routing table generation unit 1 determines “NO” in step SA5 and proceeds to step SA8.

In step SA8, the routing table generator 1 firstly sets the subnet mask (d1.subne.sub.n) in the first entry stored in the memory d1.
tmask) and the destination network address (d2.IP) in the second entry stored in the memory d2, for each bit. Next, the routing table generation unit 1 compares the destination network address (d1.IP) in the first entry stored in the memory d1 with the result of the AND operation (hereinafter, d2.IP ^* d1.subnetm).
ask) is compared as a numerical value.

That is, the destination network address (d1.IP)> the result of the AND operation (d
2. IP ^* d1. If the relationship (subnetmask) is established, the routing table generation unit 1 proceeds to step SA9, and proceeds to step SA1 and step SA2.
It is determined that sorting of a plurality of entries in was not performed correctly, and the process returns to step SA1 to execute sorting from the beginning. The above relationship does not hold when sorting is performed correctly.

On the other hand, in step SA8, the destination network address (d1.IP) <A
Result of ND operation (d2.IP ^* d1.subnetm)
When the relationship “ask” is established, the routing table generator 1 sets the contents of the second entry (destination network address (d1.IP), subnet mask (d1.subn) stored in the memory d2.
etmask) and destination information) are copied (stored) in the memory d1, and the process returns to step SA4.

In step SA4, the routing table generator 1 stores the contents of the entry next to the second entry in the memory d2, and then proceeds to step SA5 to repeat the above-described process.

On the other hand, in step SA8, the destination network address (d1.IP) = A
Result of ND operation (d2.IP ^* d1.subnetm)
ask), the memories d1 and d
2 as arguments d1 and d2, respectively, and proceed to step SB1 shown in FIG.
Call.

In step SB2, the routing table generator 1 copies (stores) the contents stored in the memory d1 to the memory D (not shown), and then copies (stores) the contents stored in the memory d2 to the memory d1. ), The process proceeds to Step SB3. Here, the destination network address stored in the memory D is “D.IP”, and the subnet mask is “D.subne.
tmask ".

In step SB3, the routing table generator 1 sets the entry next to the second entry (hereinafter referred to as the second entry).
After storing the destination network address, subnet mask, and destination information of the third entry (called the third entry) in the memory d2, the process proceeds to step SB4. That is, in this case, the following data are stored in the memory d1, the memory d2, and the memory D, respectively.・ Memory d1 → contents of second entry ・ Memory d2 → contents of third entry ・ Memory D → contents of first entry

In step SB4, the routing table generator 1 determines whether a plurality of entries in the routing table are empty. In this case, assuming that the entry of the routing table is not empty, the routing table generator 1 sets the determination result of step SB4 to "NO" and proceeds to step SB5. On the other hand, when the entry of the routing table is empty, the routing table generation unit 1 sets the determination result of step SB4 to “YES” and proceeds to step SB8.

In step SB5, the routing table generator 1 firstly sets the subnet mask (D.subnetm) in the first entry stored in the memory D.
ask) and the destination network address (d2.IP) in the third entry stored in the memory d2 is ANDed for each bit. next,
The routing table generation unit 1 regards the destination network address (D.IP) in the first entry stored in the memory D and the result of the AND operation (hereinafter abbreviated as d2.IP ^* D.subnetmask) as a numerical value. And compare the size.

That is, in step SB5, the destination network address (D.IP)>
AND operation result (d2.IP ^* D.subnetm)
If the relationship (ask) is established, the routing table generation unit 1 proceeds to step SB11 and proceeds to step SA11.
It is determined that sorting of the plurality of entries in step 1 and step SA2 was not correctly performed, and
Return to 1 and execute sorting from the beginning. The above relationship does not hold when sorting is performed correctly.

On the other hand, in step SB5, the destination network address (D.IP) <AN
D operation result (d2.IP ^* D.subnetmas)
If the relationship k) is established, the routing table generator 1 proceeds to step SB8.

In step SB8, first, the routing table generator 1 stores the subnet mask (d1.subnet) of the second entry stored in the memory d1.
mask) by one bit to the left and the memory d
The exclusive-OR is calculated for each bit with the subnet mask (d1.subnetmask) of the second entry stored in 1. Next, the routing table generator 1
Calculates the sum of the result of the exclusive OR operation and the destination network address of the second entry stored in the memory d1. Hereinafter, the result of the sum operation is abbreviated as “d1.IP ′”. Next, the routing table generator 1 calculates the result of the sum operation (d
1. IP ′) and the subnet mask (D.subnetmask) stored in the memory D for each bit A
Perform ND operation. Hereinafter, the result of the AND operation is referred to as “d
1. IP ' ^* D. abbreviated as “subnetmask”.

Then, the routing table generator 1
Is the result of the sum operation (d1.IP ') ≠ the case in the memory d2.
Destiny in the third entry contained
Connection network address (d2.IP)
(Hereinafter referred to as a first relationship) is determined.
Next, the routing table generator 1 performs an AND operation.
Result (d1.IP ' ^*D. submask) <
The disk at the first entry stored in the memory D
Is the network address (D.IP)
Whether a relationship (hereinafter, referred to as a second relationship) is established
to decide.

Then, in step SB8, the routing table generator 1 determines whether or not the above-described first or second relationship is established.
If "S", the process proceeds to Step SB10. Step SB1
0, the routing table generator 1 stores the memory d2
Is used as the return value of the function f (· d2), the process returns to step SA10 shown in FIG. 2, and then proceeds to step SA11.

On the other hand, if both the first and second relationships described above are established, the routing table generator 1
Proceeds to step SB9 after making the determination result of step SB8 shown in FIG. 3 "NO". In step SB9,
The routing table generation unit 1 inserts a dummy entry between the second entry stored in the memory d1 and the third entry stored in the memory d2 in the routing table, and then proceeds to step SB10. .

Here, the dummy entry is a result of the above-described sum operation (d1.IP ') as a destination network address, and a subnet mask (D1) of the first entry stored in the memory D as a subnet mask. .Subnetmask), destination information (hereinafter, D.next_hop), and a dummy identifier 17 (see FIG. 1).

On the other hand, in step SB5, the destination network address (D.IP) = AN
D operation result (d2.IP ^* D.subnetmas)
If the relationship k) is established, the routing table generation unit 1 proceeds to step SB6.

At step SB6, the routing table generator 1 firstly stores the second data stored in the memory d1.
Subnet mask (d1.subn) in the entry of
etmask) and the destination network address (d2.IP) in the third entry stored in the memory d2 are bit-wise ANDed.
Next, the routing table generation unit 1 determines the destination network address (d1.IP) in the first entry stored in the memory d1 and the A
The result of the ND operation (hereinafter, d2.IP ^* d1.subn)
and the magnitude relation is compared with each other as a numerical value.

That is, in step SB6, the destination network address (d1.IP)
> AND operation result (d2.IP ^* d1.subne)
When the relationship of (tmask) is established, the routing table generation unit 1 proceeds to step SB7 and proceeds to step SB7.
It is determined that sorting of the plurality of entries in A1 and step SA2 was not correctly performed, and
Returning to A1, sorting is executed from the beginning. The above relationship is
If sorting is performed correctly, it will not hold.

On the other hand, in step SB6, the destination network address (d1.IP) <A
Result of ND operation (d2.IP ^* d1.subnetm)
If the relationship (a) is established, the routing table generator 1 proceeds to step SB13.

In step SB13, first, the routing table generator 1 performs the above-described sum operation (d1.I) according to the procedure described in step SB8.
P ′). Next, the routing table generator 1
Is determined as to whether or not the above-described first relationship holds, that is, the result of the sum operation (d1.IP ′) ≠ the destination network address (d2.IP) in the third entry stored in the memory d2. I do.

Next, the routing table generation unit 1
The result of the AND operation (d1.IP ' ^*D. subnet
mask) <to the first entry stored in the memory D
Destination network address
(D.IP) whether or not the above-mentioned second relationship holds
Judge.

Then, the routing table generator 1
Determines whether the above-described first or second relationship is established, and if the determination result is “YES”, the process proceeds to step SB.
Proceeding to 15, the content stored in the memory d2 is copied (stored) in the memory d1, and the process returns to step SB3.

On the other hand, if both the first and second relationships described above are established, the routing table generator 1
Is “NO” in step SB13,
Proceed to step SB14. In step SB14, the routing table generation unit 1 inserts a dummy entry between the second entry stored in the memory d1 and the third entry stored in the memory d2 in the routing table, Proceed to step SB15.

Here, as described above, the dummy entry is a result of the above-described sum operation (d1.IP ′) as a destination network address, and a leading entry stored in the memory D as a subnet mask. Subnet mask (D.subnetma)
sk), destination information (hereinafter, D.next_hop), and a dummy identifier 17 (see FIG. 1).

In step SB6, the destination network address (d1.IP) = A
Result of ND operation (d2.IP ^* d1.subnetm)
If the relationship (ask) is established, the routing table generator 1 proceeds to step SB12. Step SB
In 12, the routing table generator 1 sets the function f
After (.) Is recursively read, the process returns to step SB5.

Now, assuming that there are no more entries in step SA5 shown in FIG. 2, the routing table generator 1 sets the determination result to "YES" and proceeds to step SA6. In step SA6, when there are a plurality of entries having the same network address in the routing table, the routing table generation unit 1 deletes the entries other than the entry having the longest subnet mask from the plurality of entries. Proceed to step SA7.

In step SA7, the routing table generation unit 1 stores the routing table having the dummy entries generated by the above-described processing as the routing table 2 shown in FIG. Is completed.

<Routing table search operation>
Regarding the routing table search operation in the routing table search device according to the first embodiment described above,
This will be described with reference to FIGS. FIG. 4 is a flowchart illustrating a routing table search operation in the routing table search device according to the first embodiment. FIG. 5 is a flowchart illustrating search results used when the control unit 14 illustrated in FIG. 1 outputs search result data 15. It is an output table.

In FIG. 1, when an IP packet is input to the search IP address holding unit 5, the search IP address holding unit 5 extracts a search IP address from the IP packet and holds it. Next, the search IP address holding unit 5 outputs a signal instructing the search of the routing table 2 to the control unit 14. Thereby, the control unit 14
Proceeds to step SC1 shown in FIG.

At step SC 1, the control unit 14 sets a result obtained by dividing the total number of entries (table size) X in the routing table 2 stored in the routing table storage memory 3 by “2” as an operation result A.
This calculation result A is
It corresponds to the entry located at the center between the first entry and the last entry, and the address where the entry is stored.

Then, the control unit 14 substitutes "1" for an internal variable i indicating the number of searches. Next, the control unit 14 outputs an address generation instruction 16 for instructing generation of the address corresponding to the operation result A to the address generation unit 11. As a result, the address generation unit 11 generates the memory address 4 corresponding to the operation result A, and outputs this to the routing table storage memory 3. As a result, from the routing table storage memory 3, the destination network address IP (hereinafter, referred to as IPA) of the A-th entry stored at the address corresponding to the operation result A is compared with the magnitude comparison unit 9, Part 13
And output to the coincidence comparison unit 12.

As a result, in step SC2, the match comparison unit 13 sets the destination network address IPA to which no subnet mask has been applied and the search I
It is determined whether or not the search IP address held in the P address holding unit 5 matches. Here, the destination network address IPA and the search IP
If the addresses match, the match comparison unit 13 outputs a signal indicating that both match, and the destination network address IPA to the control unit 14. Accordingly, the control unit 14 determines in step SC3
Then, the destination network address IPA of the A-th entry and the destination information 6 are output as search result data 15.

On the other hand, in step SC2, when there is a relationship of destination network address IPA ≠ search IP address, the match comparison unit 13
A signal indicating that the two do not match is output to the control unit 14. Accordingly, the control unit 14 proceeds to step SC4, and determines whether or not the relation X / 2 ⁱ <1 is satisfied. In other words, the control unit 14 determines whether or not the destination network addresses IP1 and IP2 having the relationship of destination network address IP1 <search IP address <destination network address IP2 have been narrowed down in the search of the routing table. .

In this case, assuming that the relationship of X / 2 ⁱ <1 is not established, the control unit 14 sets the determination result to “N
"O" and the process proceeds to Step SC5. In step SC5, the control unit 14 increments the internal variable i by 1, and then proceeds to step SC6. In step SC6, the magnitude comparing unit 9 determines whether the destination network address IPA in the A-th entry is smaller than the search IP address. Here, if the destination network address IPA is smaller than the search IP address, the magnitude comparison unit 9 proceeds to step SC1.
The judgment result of No. 6 is set to "YES", and the judgment result is outputted to the control unit 14.

As a result, in step SC7, the control unit 14 performs an operation of A = A + (X / 2 ⁱ ), and generates an address corresponding to the operation result A (hereinafter, referred to as A1) on the left side of the equation. The instruction 16 is output to the address generator 11. Thereby, the address generation unit 11 determines that the operation result A
After the memory address 4 corresponding to 1 is generated, it is output to the routing table storage memory 3. This allows
The routing table storage memory 3 outputs the destination network address IPA1 of the A1 entry corresponding to the operation result A1.

Then, in step SC2, the match comparison unit 13 determines again whether the destination network address IPA1 matches the search IP address.

On the other hand, in step SC6, if the destination network address IPA in the A-th entry is not smaller than the search IP address, the magnitude comparison unit 9 sets the determination result to “NO”, and then determines the above determination result to the control unit. 14 is output. Accordingly, in step SC8, the control unit 14 performs an operation of A = A− (X / 2 ⁱ ), and generates an address generation instruction 16 corresponding to the operation result A (hereinafter, referred to as A2) on the left side of the same expression. Is output to the address generation unit 11. Thereby, the address generation unit 11
Generates a memory address 4 corresponding to the operation result A2, and outputs it to the routing table storage memory 3. Thereby, the routing table storage memory 3
Output the destination network address IPA2 of the A2 entry corresponding to the operation result A2.

On the other hand, in step SC4, X / 2 ⁱ
When the relation of <1 is established, the control unit 14 sets the determination result to “YES”. Thereby, step SC9
Then, the match comparison unit 12 sets the subnet mask 8 of the A-th entry held in the search IP address holding unit 5.
(Abbreviated as maskA) is read from the routing table storage memory 3.

Next, the match comparison unit 12 searches the search IP address (IP address) held in the search IP address storage unit 5.
) And the above subnet mask 8 (maskA) for each bit. Hereinafter, the result of the AND operation is abbreviated as “IP ^* maskA”.

Further, the match comparison unit 12 stores the IP address 7 (abbreviated as IPA) of the A-th entry and the above-described subnet mask 8 (maskA) in A bits
Perform ND operation. Hereinafter, the result of the AND operation is referred to as “IP
A ^* maskA ".

Next, the coincidence comparing section 12 performs the above-described AND operation.
It is determined whether or not the calculated result (IP ^* maskA) matches the ANDed result (IPA ^* maskA). Here, “IP ^* maskA” and “IPA ^* maskA”
If “A” matches, the match comparison unit 12 sets the determination result of step SC9 to “YES”.

Thus, in step SC10, the dummy check unit 10 checks whether the A-th entry in the routing table 2 is a dummy check. Specifically, when the dummy identifier 17 is included in the A-th entry, the dummy check unit 10
While the first dummy entry flag is set to “1”, if the dummy identifier 17 is not included, the first dummy entry flag is set to “0”. That is, when the first dummy entry flag is “1”, it indicates that the A-th entry is a dummy entry.
When the first dummy entry flag is “0”,
This indicates that the A-th entry is the original entry. Next, in step SC11, the control unit 14 stores the contents of the A-th entry in the working memory 18 as search result data OUTtmp. Here, if the A-th entry is later than the last entry, the destination network address IP of the A-th entry
(Hereinafter abbreviated as IPtmp) is a “null” character.

On the other hand, in step SC9, if the result of the AND operation (IP ^* maskA) does not match the result of the AND operation (IPA ^* maskA), the match comparison unit 12 determines that the result of the determination is “NO”. ". As a result, in step SC12, the magnitude comparison unit 9 determines that A
It is determined whether the IP address 7 (IPA) of the second entry is smaller than the search address IP, and if the IP address 7 (IPA) is smaller than the search address IP,
The judgment result is “YES”.

As a result, in step SC13,
The control unit 14 performs an operation of A = A + 1, and generates an address generation instruction 1 for generating an address of the routing table storage memory 3 corresponding to the A-th entry on the left side of the same expression.
6 is output to the address generation unit 11. The control unit 14
Sets “1” in the magnitude comparison result flag j as the comparison result in step SC12. As a result, the address generation unit 11 outputs the memory address 4 corresponding to the A (= A + 1) -th entry to the routing table storage memory 3.

On the other hand, in step SC12, the I
If the P address 7 (IPA) is not smaller than the search address IP, the magnitude comparison unit 9 sets the determination result to “NO”.
As a result, in step SC14, the address generation unit 11 performs an operation of A = A−1, and the A
An address generation instruction 16 for generating an address of the routing table storage memory 3 corresponding to the third entry is output to the address generation unit 11. In addition, the control unit 14
As a result of the comparison in step SC12, the magnitude comparison result flag j is set to “0”. As a result, the address generation unit 11
, The memory address 4 corresponding to the A (= A−1) th entry is output to the routing table storage memory 3.

In step SC15, the match comparison unit 12
Performs the next comparison based on the A-th entry stored in the memory address 4 generated in step SC13 or step SC14. That is, the match comparison unit 12 first ANDs the search IP address (IP) held in the search IP address holding unit 5 and the subnet mask 8 (maskA) of the A-th entry for each bit. Hereinafter, the result of the AND operation is abbreviated as “IP ^* maskA”.

Further, the match comparison unit 12 sets the IP address 7 (abbreviated as IPA) of the A-th entry and the above-described subnet mask 8 (maskA) on a bit-by-bit basis.
Perform ND operation. Hereinafter, the result of the AND operation is referred to as “IP
A ^* maskA ".

Next, the coincidence comparing section 12 performs the above-described AND operation.
It is determined whether or not the calculated result (IP ^* maskA) matches the ANDed result (IPA ^* maskA). Here, “IP ^* maskA” and “IPA ^* maskA”
If “A” does not match, the match comparison unit 12 sets the determination result of step SC15 to “NO”. As a result, in step SC18, the control unit 14 executes step SC11.
At the destination network address IPtm of the A-th entry stored in the working memory 18
It is determined whether or not p is “null”. Here, the destination network address IP is "nu
ll ”, the control unit 14 determines in step S
The result of the determination at C18 is set to "YES", and step SC1
Then, the process proceeds to step S9, where search result data 15 representing a mishit as a search result is output.

On the other hand, in step SC18, the destination network address IPtmp is changed to “nu”.
If not “11”, the control unit 14 proceeds to step SC20. In step SC20, the control unit 14 outputs the search result data OUTtmp stored in the working memory 18 in step SC11 as the search result data 15.

On the other hand, in step SC15, when the result of the AND operation (IP ^* maskA) and the result of the AND operation (IPA ^* maskA) match, the match comparison unit 12 sets the determination result to “YES”. ".
Thus, in step SC16, the dummy check unit 10 checks whether the A-th entry in the routing table 2 is a dummy check. Here, the A-th means the (A + 1) -th or (A-1) -th calculated in step SC13 or SC14.

Specifically, when the dummy identifier 17 is included in the A-th entry, the dummy check unit 10 sets the second dummy entry flag to “1”, while the dummy identifier 17 is included. If not, the second dummy entry flag is set to "0". That is, when the second dummy entry flag is “1”,
The Ath ((A + 1) th or (A-1) th) entry is a dummy entry. On the other hand, if the second dummy entry flag is “0”, the Ath ((A + 1)) Or (A-1) -th entry) is the original entry.

Then, in step SC17, the search result data 15 contains the combination of the result of the dummy check 1 in step SC10, the result of the dummy check 2 in step SC16 and the magnitude comparison result flag j, and the search output result table shown in FIG. , A search result to be output as search result data 15 is obtained. Specifically, as shown in FIG. 5, when the results of the dummy checks 1 and 2 are both dummy entries and the magnitude comparison result flag j is “0”, the search that is the content of the A-th entry is performed. The result OUTA is output as the search result data 15.

If both the dummy checks 1 and 2 are dummy entries and the magnitude comparison result flag j is "1", the A-th entry stored in the working memory 18 in step SC11. Is output as search result data 15. If both the dummy checks 1 and 2 are original entries and the magnitude comparison result flag j is “0”, the above-described search result OUTtm
p is output as the search result data 15.

Further, both dummy checks 1 and 2 are original entries, and the magnitude comparison result flag j
Is “1”, the above-described search result OUTA
Is output as the search result data 15. Further, when the dummy check 1 is an original entry and the dummy check 2 is a dummy entry, the above-described search result OUTtmp is output as the search result data 15. In this case, the magnitude comparison result flag j is not considered. In addition, when the dummy check 1 is a dummy entry and the dummy check 2 is an original entry, the above-described search result OUTA is output as search result data 15. In this case, the magnitude comparison result flag j is not considered.

<Second Embodiment> FIG. 6 is a block diagram showing a configuration of a routing table search device according to a second embodiment of the present invention. In this figure, parts corresponding to the respective parts in FIG. 1 are denoted by the same reference numerals, and description thereof is omitted.
6, a subnet length comparison unit 30 is provided instead of the dummy check unit 10 shown in FIG.

The routing table search device according to the second embodiment performs a routing table search operation according to the procedure of the flowchart shown in FIG. In FIG. 9, portions corresponding to the respective portions in FIG. 4 are denoted by the same reference numerals, and description thereof will be omitted. In FIG. 9, step SC10 shown in FIG. 4 is deleted, and step SD16 and step SC16 shown in FIG.
1, SD2 and SD3 are employed.

When the determination result of step SC15 shown in FIG. 9 is “YES”, the subnet length comparing unit 30 shown in FIG. 6 determines the length of “1” of the subnet mask maskA and “1” of the subnet mask masktmp. Compare with length. Here, the subnet mask maskA
Is the subnet mask of the Ath ((A + 1) th or (A-1) th) entry calculated in step SC13 or step SC14. On the other hand, the subnet mask masktmp is the subnet mask of the A-th entry stored in the working memory 18 in step SC11.

In step SD1, the subnet length comparing unit 30 determines whether the subnet mask "maskA"
The length of “1” is “1” of the subnet mask masktmp.
Judge whether it is shorter than the length. And step S
When the determination result of D1 is “YES”, the control unit 14 proceeds to step SD2, and outputs the contents of the A-th entry stored in the working memory 18 in step SC11 as search result data 15 as a search result. . on the other hand,
If the decision result in the step SD1 is "NO", the control unit 1
4 proceeds to step SD3, and outputs the contents of the A-th entry calculated in step SC3 or SC14 as search result data 15 as a search result.

FIG. 7 is a functional block diagram showing functions of the coincidence comparison unit 12, the control unit 14, and the subnet length comparison unit 30 shown in FIG. In this figure, the functions of the comparison unit 401 and the comparison unit 402 are the same as those of the coincidence comparison unit 12 shown in FIG.
The function of the comparison unit 403 corresponds to the function of the subnet length comparison unit 30 shown in FIG. Further, the function of the determination circuit 41 corresponds to the function of the control unit 14 shown in FIG.

Therefore, in the following description, the comparison unit 12, the subnet length comparison unit 30, and the control unit 14 shown in FIG.
A description will be given with reference to FIG. here,
In the description with reference to FIG. 7, the following items are redefined in order to achieve consistency with the routing table search device shown in FIG. The “k-th term” shown in FIG.
13 (see FIG. 9) indicates the entry number in the routing table 2 shown in FIG.
This means the “A-th” in step SC11. The “(k + 1) -th item” shown in FIG. 7 indicates that the routing table 2 shown in FIG.
, The "A (= A + 1) -th" in step SC13.

On the other hand, the “k-th item” shown in FIG. 7 indicates the number of the entry in the routing table 2 shown in FIG. 6 when the process of step SC14 (see FIG. 9) is performed. (= A-1) th "
Say. The “(k + 1) -th item” shown in FIG. 7 indicates the “A-th” in step SC11, which indicates the number of the entry in the routing table 2 shown in FIG. 6, when the processing in step SC14 is performed.

In the above definition, the comparison unit 401 (coincidence comparison unit 12) shown in FIG. 7 adds the subnet mask of the k-th entry in the routing table 2 to the search IP address held in the search IP address holding unit 5. Is compared with a bit string obtained by multiplying the IP address of the k-th entry by the subnet mask of the k-th term. That is, the comparing unit 401 determines whether or not the two bit strings match. Here, the “apply a subnet mask” refers to the search I as described above.
This refers to performing an AND operation on the P address and the subnet mask for each bit.

The comparing unit 402 (matching comparing unit 12) calculates a bit string obtained by multiplying the search IP address held in the search IP address holding unit 5 by the subnet mask of the (k + 1) -th entry in the routing table 2. A comparison is made with a bit string obtained by multiplying the (k + 1) term IP address by the subnet mask of the (k + 1) term entry.
That is, the comparing section 402 determines whether or not the two bit strings match.

Comparison section 403 (subnet length comparison section 30)
Compares the length of “1” of the subnet mask of the k-th entry in the routing table 2 with the length of “1” of the subnet mask of the (k + 1) -th item. That is, when the two match or the two do not match, the comparing unit 403 determines the magnitude relationship between the two.

The determination circuit 41 converts the search result of the routing table 2 into the search result data 31 shown in FIG. 6 based on the comparison results of the comparison units 401, 402 and 403 and the comparison result output table shown in FIG. Output. Here, in the comparison result output table shown in FIG. 8, the first term in the upper column corresponds to the comparison result of the comparison unit 403, and the second term in the upper column corresponds to the comparison result of the comparison unit 401. ,
The third term in the upper column corresponds to the comparison result of the comparison unit 402. Details of the operation of the determination circuit 41 will be described later.

Next, the operation of the routing table search device according to the second embodiment will be described with reference to FIG. In the following description, the description of the same parts as those described with reference to FIG. 4 will be omitted. Further, the following description is a description of a portion corresponding to steps SD1 to SD3 shown in FIG. In FIG. 7, it is assumed that the comparison result of the comparison unit 403 is “k-th item> (k + 1) -th” shown in FIG. 8 and the comparison result of the comparison unit 401 is “=” (coincidence). . At this time, if the comparison result of the comparison unit 402 is “=” (match) or “≠” (mismatch), the determination circuit 41 determines that the search IP address is the IP address of the entry of the k-th term in the routing table 2. Is determined, and the content of the k-th entry is output as search result data 31.

On the other hand, the comparison result of the comparison unit 403 is now shown in FIG.
It is assumed that “the k-th item> the (k + 1) -th item” shown in FIG. At this time, when the comparison result of the comparing unit 402 is “=” (match), the determination circuit 41 determines that the search IP address matches the IP address of the (k + 1) -th entry in the routing table 2. Then, the content of the entry of the (k + 1) -th item is output as the search result data 31. If the comparison result of the comparing unit 402 is “≠” (mismatch), the judgment circuit 41
Produces a search result indicating that there is no entry corresponding to the search IP address in the routing table 2.

On the other hand, the comparison result of the comparison unit 403 is now shown in FIG.
, And the comparison result of the comparing unit 401 is “=” (match).
At this time, if the comparison result of the comparison unit 402 is “=” (match), the determination circuit 41 issues a search result indicating that there is no entry corresponding to the search IP address in the routing table 2. If the comparison result of the comparing unit 402 is “比較” (mismatch), the determination circuit 41 determines that the search IP address matches the IP address of the k-th entry in the routing table 2. Then, the content of the entry of the k-th item is output as the search result data 31.

On the other hand, the comparison result of the comparison unit 403 is now shown in FIG.
It is assumed that the “k-th term = (k + 1) -th term” shown in (1) and the comparison result of the comparing unit 401 is “≠” (mismatch). At this time, when the comparison result of the comparing unit 402 is “=” (match), the determination circuit 41 determines that the search IP address matches the IP address of the (k + 1) -th entry in the routing table 2. Then, the content of the entry of the (k + 1) -th item is output as the search result data 31. If the comparison result of the comparing unit 402 is “≠” (mismatch), the judgment circuit 41
Produces a search result indicating that there is no entry corresponding to the search IP address in the routing table 2.

On the other hand, the comparison result of the comparison unit 403 is now shown in FIG.
It is assumed that “the k-th item <the (k + 1) -th item” and the comparison result of the comparing unit 401 is “=” (match).
At this time, when the comparison result of the comparing unit 402 is “=” (match), the determination circuit 41 determines that the search IP address matches the IP address of the (k + 1) -th entry in the routing table 2. And the contents of the entry of the (k + 1) -th item are searched for in the search result data 31.
Output as When the comparison result of the comparing unit 402 is “≠” (mismatch), the determination circuit 41 determines that the search IP address is the k-th IP address in the routing table 2.
It is determined that the entry matches the IP address of the entry of the item, and the content of the entry of the k-th item is searched for in the search result data 3
Output as 1.

On the other hand, the comparison result of the comparison unit 403 is now shown in FIG.
It is assumed that “the k-th item <the (k + 1) -th item” and the comparison result of the comparing unit 401 is “≠” (mismatch). At this time, when the comparison result of the comparing unit 402 is “=” (match), the determination circuit 41 determines that the search IP address matches the IP address of the (k + 1) -th entry in the routing table 2. Then, the content of the entry of the (k + 1) -th item is output as the search result data 31. If the comparison result of the comparing unit 402 is “≠” (mismatch), the judgment circuit 41
Produces a search result indicating that there is no entry corresponding to the search IP address in the routing table 2.

Although the embodiment of the present invention has been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and a design change or the like may be made without departing from the gist of the present invention. Even if present, it is included in the present invention. For example, in the routing table search devices according to the first and second embodiments described above, the I
Change all bits of the host address part of the P address to "
The example in which the AND operation is performed in the match comparison unit 12 by setting the value to "0" has been described. However, the present invention is not limited to this. That is, in the routing table search apparatuses according to the first and second embodiments described above, After setting all of the host address portion of the IP address to “1” instead of “0”, the inversion bit (“”) of the subnet mask in the selected entry.
The OR operation may be performed on the search IP address and the search IP address instead of the AND operation described above.

[0111]

According to the present invention, the search IP address is compared with the registered IP address without applying a subnet mask to the search IP address.
Even if the subnet mask has a variable length, the routing table search process can be performed at a high speed, and a large amount of high-speed input IP packets are received and the IP packets are output to a route that immediately leads to the destination. And the search error caused by the variable length of the subnet mask can be prevented.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a routing table search device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing a procedure for generating a routing table by a routing table generating unit 1 shown in FIG. 1;

FIG. 3 is a flowchart showing a procedure for generating a routing table by a routing table generator 1 shown in FIG. 1;

FIG. 4 is a flowchart illustrating an operation of the routing table search device according to the first embodiment of the present invention.

FIG. 5 is a diagram showing a search result output table used in the routing table search device according to the first embodiment.

FIG. 6 is a block diagram illustrating a configuration of a routing table search device according to the second embodiment.

FIG. 7 is a block diagram showing a configuration of a routing table search device according to the second embodiment.

FIG. 8 is a diagram showing a search result output table used in the routing table search device according to the second embodiment.

FIG. 9 is a flowchart illustrating an operation of the routing table search device according to the second embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Routing table generation part 2 Routing table 3 Routing table storage memory 4 Memory address 5 Search IP address holding part 6 Destination information 7 IP address 8 Subnet mask 9 Size comparison part 10 Dummy check part 11 Address generation part 12 Match comparison part 13 Match comparison unit 14 Control unit 15 Search result data 16 Address generation instruction 17 Dummy identifier 18 Working memory 401, 402, 403 Comparison unit 41 Judgment circuit

Claims (3)

[Claims] 1. After all bits of each host address portion of a plurality of registered IP addresses to be registered are set to "0" and bits of each subnet mask portion corresponding to the registered IP address are set to "1", After the one registered IP address, the subnet mask, and the destination information are grouped into one entry data, the registered IP address is sorted as a numerical value to generate a routing table including a plurality of entries. The search IP address read from the received IP packet and the registration I of the entry.
In a routing table search device for performing a search by narrowing a search range in a routing table by half by comparing the size with a P address, the search IP address and the registered IP address can be set without applying a subnet mask to the search IP address. A comparing unit that performs the magnitude comparison and searches for two entries in the routing table that directly sandwiches the search IP address; and compares one of the registered IP address and the variable-length subnet mask in the two entries. A for each bit
A registration that matches the search IP address from the routing table based on a comparison result between the result of the ND operation and the result of the AND operation of the other registered IP address and the variable-length subnet mask for each bit. I
A routing table searching device, comprising: searching means for searching for a P address. 2. The routing table generating means according to claim 1,
All bits of the host address portion of each registered IP address in the routing table are set to “1” instead of “0”, and the search unit determines that one of the registered IP addresses and the variable length in the two entries The result of ORing the inverted bits of the subnet mask (in which “0” and “1” are exchanged) for each bit and the other registered IP
The search is performed in the routing table based on a comparison result between an address and a result obtained by performing an OR operation on a bit-by-bit basis with a variable-length inverted bit (in which “0” and “1” are exchanged) of the subnet mask. The routing table search device according to claim 1, wherein a search is made for a registered IP address that matches the IP address. 3. After setting all bits of each host address portion of a plurality of registered IP addresses to be registered to “0” and setting bits of each subnet mask portion corresponding to the registered IP address to “1”, After the one registered IP address, the subnet mask, and the destination information are combined into one entry data, the registered IP address is sorted as a numerical value to generate a routing table including a plurality of entries. The search IP address read from the received IP packet and the registration I of the entry.
In a routing table search device for performing a search by narrowing a search range in a routing table by half by comparing the size with a P address, the search IP address and the registered IP address can be set without applying a subnet mask to the search IP address. A comparison unit that performs the magnitude comparison and directly searches for the two entries in the routing table that directly sandwiches the search IP address; and, in the two entries, one of the variable-length subnet masks “1” of one of the entries. Search means for searching the routing table for a registered IP address that matches the search IP address, based on a comparison result of the length and the length of the variable "1" of the subnet mask of the other entry; Routing characterized by comprising: Buru search device.