JPH0730587A - Frame relay control method and its device - Google Patents

Abstract

PURPOSE:To decide a frame relay destination at a high speed by receiving a frame to decide a relay destination network and rearranging the connection of plural relay table entries depending on the frequency of use. CONSTITUTION:Every time a repeater 1 receives a frame and an entry retrieval device 3 retrieves a relay control table entry, an entry revision device 42 rearranges a control table entry of a relay control table entry group 4 or a grouped relay control table entry. That is, when a frame of same destination address is continuously received, the rearrangement of the relay control table entry is executed by the reception of one frame. The relay control table entry used to retrieval of the destination network of this frame is rearranged to be retrieved with priority from next time. Thus, the relay destination network in the reception frame is decided at a high speed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a frame relay control method in a relay device and its device.

[0002]

2. Description of the Related Art Conventional frame relay control methods (reference: I
NTERNETWORKING WITH TCP / IP VOLUMEII, Author: DOUGL
ASE E.COMER and DAVID L.STEVENS, Publisher: Prentice-
HallInternational, Inc, ISBN: 0-13-465378-5 Publication year: 1991), the sequence of relay control table entries is fixed at the time of registration, and the relay control table is always in sequence at the time of registration of relay control table entries. I was doing a search for an entry.

[0003]

The above-mentioned conventional method does not consider preferentially searching the relay control table entry that is frequently used, and even when a frame having a high relay frequency is received, It took a relay control table entry search time equivalent to a frame with a low relay frequency. Therefore, there is a problem that the relay of a large number of data frames having the same destination address cannot be speeded up.

An object of the present invention is to provide a frame relay control method and apparatus capable of preferentially searching a relay control table entry that is frequently used and accelerating the relay of a large number of data frames having the same destination address. To provide.

Another object of the present invention is to comply with a specified ratio even when a frame addressed to a certain network is distributed and relayed to a plurality of networks at a specified ratio. A frame relay control method and a device therefor capable of easily realizing load distribution.

[0006]

In order to solve the above object, the present invention uses a plurality of relay control table entries and an entry search mechanism for determining a relay destination network of a frame. In the frame relay control method, a function of executing a rearrangement of the relay control table entries according to the frequency of use is added to preferentially search a relay control table entry having a high frequency of use. Make up.

In order to achieve the above-mentioned other object, according to the present invention, rearrangement control information indicating a frame relay ratio is added to a relay control table entry and a relay destination corresponding to a frame destination address is added. From the relay control table entries with different networks, by adding the function of rearranging each relay control table entry according to the ratio specified by this rearrangement control information, the specified ratio And a frame relay control method capable of frame relaying.

Further, according to the present invention, the relay control table entries are grouped according to the frame destination, and the function of rearranging the relay control table entries in group units is added. By constructing a frame relay control method capable of determining a relay destination network of a received frame at high speed by preferentially searching a group of frequently used relay control table entries. .

Further, in order to achieve the above-mentioned object of the present invention, a network connected to a plurality of networks and connected.
In a relay device for relaying frames between networks, a frame receiving mechanism for receiving the frame from the network and a relay control mechanism for determining the relay destination network according to the destination address of the received frame. And a frame transmission mechanism for transmitting the frame whose relay destination network has decided the relay destination network to the relay destination network. This relay control mechanism determines the relay destination of the frame. Multiple relay control table entries in order to search for multiple relay control table entries in response to frame reception, and multiple relay control table entries for searching frequently used relay control table entries first. A frame relay control device having an entry rearrangement mechanism for rearranging relay control table entries is configured.

Furthermore, in the present invention, the relay control table
The bull entry has rearrangement control information indicating the transmission ratio, and the entry rearrangement mechanism of the relay control mechanism relays the relay destination network different from the destination address of the frame based on the ratio indicated by the rearrangement control information. The control table entries are rearranged.

Further, in the present invention, the relay control table entries are grouped by the destination of the frame, and the entry rearrangement mechanism is arranged to execute the rearrangement of the relay control table entries in units of this group.

[0012]

In the above-described structure of the present invention, every time the relay device receives a frame and the entry retrieval mechanism retrieves a relay control table entry, the entry rearrangement mechanism of the relay control mechanism causes the relay control table to operate. Bullet entry group relay control table
Bullet entries or grouped relay control table entries are rearranged, the relay control table entries that are frequently used are searched preferentially, and the relay network of the receiving frame is transmitted at high speed. To decide.

That is, when frames with the same destination address are continuously received, the relay control table entries are rearranged by the first frame reception, and the destination network of this frame is received. The relay control table entries used for searching the network are rearranged so as to be preferentially searched from the next time. Therefore, from the second frame, the search is started from the relay control table entry having the destination network of this frame, and the relay destination network of the received frame is determined at high speed.

Further, according to the rearrangement control information of the relay control table entry, the rearrangement is performed only when the specified transmission rate is exceeded, and the frame is relayed according to the specified transmission rate. By doing so, the traffic of each relay destination network is distributed.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Specific embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows an embodiment of the internal schematic configuration of a relay device 1 for implementing the present invention and a configuration example of a network to which the relay device 1 is connected. The relay device 1 is directly connected to the network A (7a), the network B (7b), and the network C (7c). The relay device 8b includes a network B (7b), networks D1, D2, D3,
Connect directly to D4 and E. The relay device 8c is directly connected to the network C (7c) and the network D1. The relay devices 8b and 8c have the same configuration as the relay device 1, but the illustration thereof is omitted.

By the way, the relay device 1 uses the network A (7
a), frame receiving mechanisms 5a, 5b, 5c for receiving frames from B (7b), C (7c) and a next hop address for transmitting the frame received by the frame receiving mechanism to the destination address of the frame. It comprises a relay control mechanism 2 and frame transmission mechanisms 6a, 6b, 6c for transmitting a frame to the network having the next hop address determined by the relay control mechanism 2. The frame transmission mechanisms 6a, 6b and 6c are respectively connected to the network A (7
a), B (7b), and C (7c).

The relay control mechanism 2 searches the relay control table entry group 4 for the next hop address from the destination address of the received frame, the relay control table entry group 4, and the entries described in detail later. And an update mechanism 42. The relay control table entry group 4 is, as illustrated later, a relay control table entry having information of a pair of destination subnetwork address and next hop address, and a hash table for searching the relay control table entry by hash calculation. Composed of. For a general description of relay control table entries, see Douglas E. Comer et al., "Internet working with TCP / IP VOLUME 2" (above),
See pages 81-84 for details.

Now, FIG. 2 shows an embodiment of the structure of the relay control table entry 20. The relay control table entry 20 has a configuration common to the relay control table entries 21 to 28 in FIGS.
Also, in the following description, the "mask" is a bit unit value indicating which part of the destination address is the subnet work address. The "mask length" is the number of bits in which the bits of the mask are set. The "root entry" indicates an entry chained by the pointer of the entry having a small mask, and the subnet mask has the same mask.
This is the entry from which an entry with a different address is searched. "Address" consists of network address and individual address. "Network address"
Is the network address part of the address. "Subnet work address" is an address that combines a network address and a subnet work identifier,
The subnet network identifier is a part of the individual address part and is used to construct a plurality of networks with one network address. For example, if the destination address is 200.0.1.25 and the mask is 255.255.255.0,
Subnet work address is 200.0.1, mask length is
24, Subnet work identifier is 1, Destination network
The address is 200.0 and the individual address is 1.25 (numerical value 200.0.1.25 is a 4-byte hexadecimal data 0xC8000119 separated by a period "." In each byte and expressed in decimal. Is).
The "next hop address" is a subnet work address to be relayed next to the received frame. For a detailed description of these, see, for example, "Internet Standard S
ubnetting Procedure, RFC 950, 1985 ".

In FIG. 2, a pointer 201 to an entry having a different network address is a pointer indicating an entry having the same hash key but a different network address. When updating the relay control table entry, the pointer 201 stores the relay control table entry in the same hash key when the hash calculation result obtained by calculating the network address of the destination sub-network address entry of the relay control table entry is duplicated. Used to chain. Also, it is used to retrieve the relay control table entry having the same hash key at the time of the search. A pointer 202 to a relay control table entry having the same mask but different subnetwork addresses is a pointer indicating a relay control table entry having the same mask but different subnetwork addresses. This pointer 202 is used to chain relay control table entries having the same mask length but different destination subnetwork addresses when updating the relay control table entries. Also, it is used to retrieve relay control table entries having the same mask length but different destination subnetwork addresses at the time of the search.

Pointer 20 to entry with small mask
Reference numeral 3 is a pointer indicating a relay control table entry having a smaller mask length. When updating the relay control table entry, this pointer 203 is used for chaining the relay control table entries in which a part of the destination sub-network address matches in order of increasing mask length. For example,
Mask 25 with destination subnet work address 200.0.32.0
Mask 255.255.255.0 with relay control table entry of 5.255.240.0 and destination subnet work address 200.0.33.0
For the relay control table entry of, the subnet work address matches up to 200.0.32 and the mask is 255.255.25.
Since 5.0 is longer, the latter is chained in order of the former.
Also, at the time of the search, it is used to search from the relay control table entry having a longer mask length.

The pointer 204 to the same cost entry is
2 routes to the destination network address of the receiving frame
If there is more than one, it is a pointer to a relay control table entry having a next hop address that differs depending on the destination subnetwork address. This pointer 204 is used to chain the relay control table entries of the same cost when updating the relay control table entries. It is also used to retrieve the relay control table entry of the same cost when searching.

The destination subnetwork address 205 is
This is the value of the destination subnetwork address that this relay control table entry has. Next hop address 206
Indicates the next hop address to be transmitted next to the destination subnetwork address. The rearrangement control information 207 indicates how many times this relay control table entry should be used for rearrangement to use the next relay control table entry indicated by the pointer to the same cost entry. Each element 2 of this relay control table entry 20
In 01 to 207, when 0 is written, it indicates that the element is an invalid or final pointer. The mask 208 is a mask for the destination subnet work address and the frame destination address held by the relay control table entry.

FIG. 3 shows an example of the initial configuration of the relay control table entry group 4 in the network configuration of FIG. The relay control table entry group 4 has a hash table 10 for searching a relay control table entry by a hash operation, and the structure shown in FIG.
Multiple relay control table entries 2 chained to 0
1 to 28. Relay control table entry 21
Items 28 to 28 are chained in the hash table 10 that performs a hash operation using the destination network address of the received frame as a hash key, and is searched by the entry search mechanism 3. If there are two or more routes to the destination network address of the reception frame, the relay control table entry 2
Destination subnetwork address 205, such as 3, 25
But the same next hop address 206 has different relay control table entries such as B and C. In this example, the destination subnetwork address is network D1.
In this case, the rearrangement control information 207 is designated to be transmitted to the networks B and C at a ratio of 2 to 5.

The relay control table entry 21 is the next hop address 2 because the destination subnetwork address is the network A directly connected to the relay device 1.
There is no 06 (indicating 0). In this example, a case is shown in which the hash calculation results for the relay control table entry 22 and the relay control table entry 23 match. The relay control table entry 22 does not have the next hop address 206 (indicates 0) because the destination sub-network address 205 is the network B directly connected to the relay device 1, and the hash calculation result overlaps with the relay control table entry 23. Therefore, a pointer to the relay control table entry 23 (a pointer 2 to an entry whose network address is different but the hash calculation result is the same)
01). The relay control table entry 24 has the network B as the next hop address 206 because the destination subnetwork address 205 is the network E that is not directly connected to the relay device 1.

The networks D1, D2, D3 and D4 are
Since it is a subset of the network D, the relay control table entries 23, 25, 26, 27, 28 form a relay control table entry D group 30. The relay control table entry 23 has the network B as the next hop address 206 because the destination sub-network address 205 is the network D1 that is not directly connected to the relay device 1, and the rearrangement is the transmission ratio designated as “2”. The control information 207, the pointer 204 to the relay control table entry 25 that is the same cost entry that specifies the transmission ratio, the pointer 202 to the relay control table entry 26 that has the same mask and a different destination subnetwork address, and the relay with a small mask It has a pointer 203 to the control table entry 28. In addition, the relay control table entry 25 includes the network D1 whose destination sub-network address is not directly connected to the relay device 1.
Therefore, the network C is assigned to the next hop address.
And the rearrangement control information 207, which is the transmission ratio designated as “5”.

In the relay control table entry 26, since the destination network address 205 is the network D2 that is not directly connected to the relay device 1, the next hop address 206 has the same mask as the network B, but the destination subnetwork address is different. It has a pointer 202 to the table entry 27. The relay control table entry 27 has a destination subnetwork address 205.
Is the network D3 that is not directly connected to the relay device 1, and therefore has the network B at the next hop address 206. The relay control table entry 28 has the network B as the next hop address 206 because the destination subnetwork address 205 is the network D4 that is not directly connected to the relay device 1.

Next, this relay control table entry group 4
A specific example of the rearrangement control method of the present invention will be described. In addition, the network A shown in FIG.
When indicating the network address of B, C, D1, D2, D3, D4, E, for example, the network address is referred to as network A, and when indicating the subnet work address, for example, The network address is called network A and is used properly according to the needs.

For example, when the initial state of the relay control table entry group 4 is as shown in FIG. 3 and a frame is received from the network A (7a) and the network address of the destination address of the frame indicates the network D1,
The entry retrieval mechanism 3 retrieves the relay control table entry group from the network address of the destination address of the frame passed from the frame reception mechanism 5a, and the destination subnet work address 20 of the relay control table entry 22.
After checking the network address of No. 5, the relay control table entry 23 in which the network address of the destination subnetwork address indicates the network D1 is searched.

In this embodiment, the relay control table entry 23 and the relay control table entry 22 are exchanged in order to check the relay control table entry 23 first from the next time by using the entry updating mechanism 42 which will be described in detail later. With such a configuration of the relay control table entry as described above, when the frames having the network address of the destination address of the reception frame indicating the network D1 are continuously received, the relay control table is first searched when two or more frames are searched. Since the entry 23 can be checked, the search can be speeded up. In this embodiment, the relay control table entry 23 is the relay control table entry D group 30.
The relaying unit belongs to the relay control table entry D group 30 as a whole.

In the initial state of the relay control table entry group 4 shown in FIG. 4, when a frame is received from the network A (7a) and the network address of the destination address of the frame indicates the network D2, The entry retrieval mechanism 3 retrieves the relay control table entry group from the network address of the destination address of the frame passed from the frame reception mechanism 5a, and the destination subnet work address 205 of the relay control table entry 23.
Of the relay control table entry 23, the network address of the relay control table entry 23 and the destination address of the received frame are set to the mask 208 of the relay control table entry 23. The multiplied subnetwork address is checked. Since they do not match, the destination sub-network address 205 searches the relay control table entry 26 indicating the network D2.

Then, in order to check the relay control table entry 26 first from the next time, the entry control mechanism 42 is used to replace the relay control table entry 26 with the relay control table entry 23, and the relay control table entry as shown in FIG. With the above configuration, when the frames whose network address of the destination address of the reception frame indicates the network D2 are continuously received, the relay control table entry 26 can be checked first when searching for two or more frames. Search can be speeded up. In addition,
Since the relay control table entry 23 is chained with the relay control table entry 25 that is the same cost entry, the relay control table entry 25 is replaced while being chained.

In the initial state of the relay control table entry group 4 shown in FIG. 4, when a frame is received from the network A (7a) and the network address of the destination address of the frame indicates the network D1, The entry retrieval mechanism 3 retrieves the relay control table entry group from the network address of the destination address of the frame passed from the frame reception mechanism 5a, and the destination subnet work address 205 of the relay control table entry 23.
Network address and the number of times of use of this entry is checked by the rearrangement control information 207. If the number of times of use exceeds the rearrangement control information 207, it is determined that the relay control table entry 25 is the same cost entry. The relay control table entry 23 is replaced, and FIG.
By configuring the relay control table entry as described above, the number of times the same cost entry is used can be controlled.

FIG. 7 is a schematic system configuration diagram of the relay control mechanism 2 shown in FIG. 2, in other words, an operation relational diagram of the entry retrieval mechanism 3, the entry update function 42, and the relay control table entry group 4. In the present embodiment, the entry search mechanism 3 and the entry update mechanism 42 are programs stored in the memory 41 and executed by the processor 40. As the processor 40, for example, the router processor, the bruta processor, etc. described in FIGS. 1 and 2 of Japanese Patent Laid-Open No. 05-22293 mentioned above are used.
Further, a processor or the like in a routing accelerator (relay device) described in Japanese Patent Application No. 04-196730 "Internetwork Device and Communication Network System" is used. An example of a specific system configuration will be described later with reference to the drawings.

When the entry search mechanism 3 receives a frame and searches the relay control table entry group 4, the following processing is performed. For example, the frame receiving mechanism 5a
When the frame is received, the frame receiving mechanism 5a receives the bus control L
The control line 43 is used to notify the entry retrieval mechanism 3 via the SI 105, and the data line 50 is used to notify the frame buffer 46.
The frame is stored in. Next, the entry search mechanism 3 reads the frame from the frame buffer 46 by the data line 45, reads the relay control table entry group 4 with the network address of the destination address of the frame, and performs a search. . When the next hop address is determined to be the network C by the search, when the transmission instruction is given to the frame transmitting mechanism 6c through the control line 44, the frame transmitting mechanism 6c
The frame is read out from the data line 51 and transmitted.

When the frame is received and the relay control table entry group 4 is updated by the entry updating mechanism 42, the following processing is performed. The frame updated by the entry updating mechanism 42 is a special frame for updating the entry, and its frame structure is determined by the routing protocol used. For example, when the frame receiving mechanism 5a receives this frame, the entry updating mechanism 42 notifies the entry retrieving mechanism 3 via the control line 43, and the data line 5
At 0, the frame is stored in the frame buffer 46. The entry update mechanism 42 reads the frame from the frame buffer 46 by the data line 45, and reads the relay control table entry group 4 according to the contents of the information section of the frame to update the frame.

FIG. 8 shows the processing flow of the entry search mechanism 3.
It is a figure which shows one Example. A hash operation is performed with the network address of the destination address of the reception frame and the hash table 10 is drawn (process 60). Next, the relay control table
The network address of the destination address of the received frame matches the network address of the destination subnet network address of the relay control table entry by grabbing the pointer 201 to an entry having a different network entry network address. The entry to be searched is searched (process 61). If there is an entry, the process 62 is carried out. If there is no entry, there is no entry corresponding to the network address of the destination address of the received frame, so the process is terminated.

In the process 62, the pointer 202 to the entry having the same mask but a different subnet work address, including the root entry which is the entry chained by the pointer 203 to the entry having a small mask, is laid out. The destination subnet work address 205 of the relay control table entry and the destination address of the received frame are multiplied by the mask 208 of the relay control table entry to find an entry having a matching subnet work address. If there is no entry, the process 63 is performed. If there is an entry, the search is terminated, the next hop address is obtained from the matching entry (process 64), this entry is rearranged with the root entry, and the relay control to which this entry belongs The table entry group is sorted into the first entry chained by the pointer 201 to the entry having a different network address in the hash table (process 65).
The process 65 will be described later in detail. Process 63 is a rule
20 Pointer to entry with small mask
Move to entry with smaller mask than indicated by 3.
If there is a next entry to be moved, the process returns to step 62 to continue the process, and if there is no next entry to be moved, this entry is the last entry and the process is terminated.

FIG. 9 shows an embodiment of the processing flow of the entry updating mechanism 42 (new addition of relay control table entry).
FIG. The entry update does not require the mask length to be the same as the routing protocol (eg, OSPF;
In order to support Open Shortest Path First), relay control table entries are registered so that searches can be performed in order of increasing mask length. First, a hash operation is performed with the network address of the relay control table entry to be registered, and the hash table 10 is drawn.
(Process 70). Next, the pointer 201 to an entry having a different network address of the relay control table entry is draped, and the network address of the destination subnet work address of the relay control table entry to be registered and the hash management table are registered. -Search for an entry in which the network address of the destination subnet work address of the relay control table entry registered in the table matches (step 71). If there is no matching entry in the process 71, the relay control table entry to be registered at the end of the relay control table entry chained by the pointer 201 to the entry having a different network address of the relay control table entry. Bull entry is added and the process ends (process 75). If there is a matching entry in the process 71, the pointer 203 to the entry with the smaller mask of the relay control table entry is waged, and the route entry is searched for a relay control table entry with a matching mask length (process 72). . If there is no matching entry in the process 72, the relay control table entry to be registered is changed to the entry with the smaller mask so that the root entries are chained by the pointer 203 to the entry with the smaller mask in the order of the longer mask length. Chaining is performed by the pointer 203 (process 76). In step 73, if there is a matching entry, the pointers to the entries having the same mask but different subnet work addresses are laid, and the destination subnet work address of the relay control table entry to be registered and the route are registered. An entry that includes the entry and the destination subnetwork address of the relay control table entry registered in the hash management table is searched for (step 73).

At step 73, if there is a matching entry,
The relay control table entry to be registered is added to the end of the relay control table entry chained by the pointer 204 to the same cost entry of the relay control table entry, and the process ends (step 77). If there is no matching entry in the process 73, it should be registered at the end of the relay control table entry chained by the pointer 202 to the entry having the same relay control table entry mask but different subnet work addresses. The relay control table entry is added and the process ends (process 74).

FIG. 10 is a diagram showing an embodiment of the overall system configuration of an apparatus equipped with the relay apparatus 1 of FIG. Such a system configuration is described, for example, in Japanese Patent Application No. 4-196730, "Internetwork device and communication network system" described above. Figure 10
Relay device (routing accelerator) 1
Connects the relay protocol control device (router management unit) 100 to another relay device 1 through the high-speed bus 106.
The frame relay is performed by the line control LSI 10 in the relay device 1.
Not only 7 but another relay device 1 connected to the high-speed bus
Also done to. The relay protocol control device 100 is an OS
Performs relay protocol processing such as PF and
101, processor 40, internal bus 104, external storage device 102, external communication device 103, frame buffer 4
6. Consists of a bus control LSI 105.

The relay device 1 is a relay protocol control device 1
00 relay protocol processing, relay processing is performed based on the information of the relay control table entry, and the memory 41
(General term for main storage 101 and frame buffer 46), processor 40, internal bus 104, bus control LS
I 105 and line control LSI 107. Figure 1
The frame transmitting / receiving mechanisms 5a and 5b are the line control LSI 10
It is configured within 7. The external storage device 102 is a storage medium such as an IC memory card. The external communication device 103 is a communication interface with the outside, such as RS-232C or Ethernet, and is used for maintenance. The main storage 101 stores a relay protocol control program. Frame buffer 46
Is a bus control LSI 105 or a line control LSI 107
Stores the send / receive frame for.

The bus control LSI 105 uses the internal bus 10
4. Control of the high speed bus 106 and the frame buffer 46 bus. The processor 40 reads and writes the frame buffer 46 through the bus control LSI 105. The line control LSI 107 includes FDDI, Ethernet, leased line, I
A group of control LSIs such as SDN and a plurality of connection ports of the same type (for example, 5a, 6a, 5b, 6b, 5c, 6c)
have. The relay control table entry used in the relay device 1 is created in the relay protocol control device 100 and distributed to the relay device 1. The distributed relay control table entries are distributed to the entry updating mechanism 4 described in detail by each relay device 1.
At step 2, it is registered in the relay control table entry group 4.

FIG. 11 shows the relay control table of the process 65 of FIG.
It is a detail of the rearrangement processing of the bull entries. First, the rearrangement control information 207, which is the transmission ratio, is read, and it is checked whether the transmission ratio of this entry to the next network address is greater than or equal to the transmission ratio of this entry (process 110). As a result of the check, if the transmission ratio is not higher than or equal to the processing 113, the processing 113 is performed.
Perform 11. The process 111 checks whether or not the pointer to the same cost entry is set, and if it is not set, the process 113 is performed, and if it is set, the process 1 is performed.
Twelve. In the process 112, this entry is chained to the end of the entries chained with the same cost entry, and the entries are rearranged (the entry after the rearrangement is called the main entry in the next process). An example of the processing from processing 110 to processing 112 is rearrangement of the relay control table entries 25 and 23 in FIG. The process 113 is
By rearranging this entry with the root entry, the entries are rearranged by the pointers to the entries having the same mask but different subnet work addresses (the entry after the rearrangement is called the main entry in the next process).
An example of the processing of the processing 113 is rearrangement of the relay control table entries 26 and 23 of FIG. In step 114, the relay control table entry group including this entry is rearranged so that it is directly chained by a pointer from the hash table to an entry having a different network address. An example of the processing of the processing 114 is rearrangement of the relay control table entries 22 and 30 of FIG.

In the above embodiment of the present invention, for example, when the number of networks is about 40 and the number of entries to be searched is 20 on average, the relay performance is improved by about 15%.
Further, as the number of connected networks increases, this relay performance can be improved.

[0046]

According to the present invention, the function of automatically rearranging the relay control table entries enables the relay control table entries having a high frequency of use to be searched first, so that the relay control table entries can be searched at high speed. The frame relay destination was determined and the frame could be relayed.

When there are a plurality of relay destination networks for the same frame destination address, the relay control table
By adding the rearrangement control information to the bull entry, it becomes possible to specify the ratio of transmission to the relay destination network and distribute the traffic. Therefore, for the destination address of the same frame, With respect to a plurality of relay control table entries having different relay destination networks, a relay control table entry having each relay destination network can be selected at a specified ratio.

Also, the relay control table entry is
The relay control table entries are rearranged in group units according to the destination of the system, and the relay control table entry group that is frequently used is searched first so that the frame can be read at high speed. It is now possible to determine the relay destination.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram of a relay device that realizes the present invention.

FIG. 2 is a structural diagram of a relay control table entry used in the present invention.

3 is an initial configuration diagram of a relay control table entry group (4) in the network configuration of FIG.

FIG. 4 shows that the destination address of the received frame is network D
FIG. 6 is a configuration diagram of a relay control table entry group (4) when 1 is shown.

FIG. 5: The destination address of the received frame is network D
It is a block diagram of the relay control table entry group (4) in the case where 2 is shown.

FIG. 6 shows that the destination address of the received frame is network D
FIG. 6 is a configuration diagram of a relay control table entry group (4) when 1 is shown.

FIG. 7 shows an embodiment of the system configuration of the present invention, and is a related diagram of a relay control table entry group of an entry search mechanism and an entry update mechanism.

FIG. 8 is a diagram showing an example of a processing flow when an entry is searched.

FIG. 9 is a diagram showing an example of a processing flow at the time of updating an entry (new addition).

10 is an overall configuration diagram of a device including the relay device 1 of FIG.

FIG. 11 is a diagram showing an example of an entry rearrangement processing flow.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Relay device, 2 ... Relay control mechanism, 3 ... Entry retrieval mechanism, 4 ... Relay control table entry group, 5a-5c ... Frame receiving mechanism, 6a-6c ... Frame transmitting mechanism, 7a ... Network -Work A, 7b ... Network B, 7c ... Network C, 8b-8c ... Relay device, 10 ... Hash table, 20-28 ... Relay control table entry, 30 ... Relay control table entry D group, 40 ... Processor, 41 ... Memory, 42 ... Entry update mechanism, 43, 44 ... Control line, 45 ... Data line, 46 ... Frame buffer, 47-51 ... Data line, 100 ... Relay protocol control device, 101 ... Main storage, 102 ... External storage device, 103 ... External communication device, 104 ... Internal bus, 105 ... Bus control LSI, 106 ... High-speed bus , 107 ... line control LSI.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Tsutomu Nakagawa Tsutomu Nakagawa 6-81, Onoue-cho, Naka-ku, Yokohama-shi, Kanagawa Hitachi Software Engineering Stock Association In-house (72) Inventor Tokuhiro Niwa 810 Shimoimaizumi, Ebina-shi, Kanagawa Stock Company Hitachi, Ltd. Office Systems Division (72) Inventor Naoya Ikeda 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Hitachi, Ltd. Microelectronics Equipment Development Laboratory

Claims (10)

[Claims] 1. A frame relay control method of a relay device for relaying a frame between connected networks, comprising: receiving a frame from one network; and a plurality of relay control table entries connected by pointers. Using the relay destination network according to the received destination address of the frame, the frame is transmitted to the determined relay destination network, and the relay control table entries of the plurality of relay control table entries are transmitted. A frame relay control method characterized in that the connections are rearranged according to the frequency of use. 2. The relay control table entry includes rearrangement control information indicating a rate of transmission to each relay destination network when there are a plurality of relay destination networks for the destination address of the frame. The relay control table entries having different relay destination networks are rearranged based on the rearrangement control information.
The frame relay control method described. 3. A plurality of relay control table entries are grouped according to a destination address of the frame, and a group is formed.
2. The frame relay control method according to claim 1, wherein the plurality of relay control table entries are rearranged in a unitized group. 4. A relay device which is connected to a plurality of networks and relays a frame between the connected networks, the frame receiving mechanism receiving a frame from the network. A relay control mechanism that determines a relay destination network according to the received destination address of the frame, and a frame transmission mechanism that transmits the frame to the determined relay destination network. However, the relay control mechanism has a plurality of relay control table entries for determining a frame relay destination, an entry search mechanism for searching the plurality of relay control tables in response to reception of a frame, and a frequency of use. Frame relay control device, which has an entry rearrangement mechanism for rearranging the plurality of relay control table entries in order to retrieve the relay control table entry having a high number. 5. The relay control table entry has rearrangement control information indicating a transmission rate, and the entry rearrangement mechanism sets a destination address of a frame based on a rate indicated by the rearrangement control information. 5. The frame relay control device according to claim 4, wherein a plurality of relay control table entries having different relay destination networks are rearranged. 6. A plurality of relay control table entries are grouped according to a frame destination, and the entry sorting mechanism sorts in units of the group. The frame relay control device described. 7. A network system in which a plurality of networks are connected via a plurality of relay devices and which relays a frame from one network to another network, wherein the relay device relays the frame. The frame receiving mechanism to receive and a plurality of relay control table entries connected in a predetermined sequence are searched, and the relay destination network of the frame is determined by the destination address of the frame received by the frame receiving mechanism. Relay which comprises entry retrieving means for deciding the number of relay control table entries, and entry reordering means for reordering the connections of the plurality of relay control table entries in order to retrieve the relay control table entry that is frequently used. And a frame transmission mechanism for transmitting the frame to the relay destination network determined by the relay control mechanism. Network system and butterflies. 8. The relay control table entry has rearrangement control information indicating a transmission rate, and the entry rearrangement means sets a destination address of a frame based on a rate indicated by the rearrangement control information. 8. The network system according to claim 7, wherein a plurality of relay control table entries having different relay destination networks are rearranged. 9. A plurality of relay control table entries are grouped according to a frame destination, and the entry sorting means sorts in units of the group. The described network system. 10. The network according to claim 7, wherein the relay control mechanism has means for selecting one of the plurality of relay control table entries by performing a hash operation with a destination address of the frame. system.