JP2001111616A - Storage exchange switch and storage exchange method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a storage exchange switch and a storage exchange method that a plurality of paths between nodes can simultaneously be used and occurrence of a loop in frames can be prevented. SOLUTION: When the one storage exchange switch is started, the switch receives information denoting a reachable terminal address and denoting a required relay number up to each terminal from the other switch connected to itself, and registers paths to a path management table by each terminal address on the basis of the received information. The switch increments the relay numbers by one in the case of the registration. In the case of transferring frames, the switch retrieves the path management table by using a destination address of a frame by a key, selects an idle port corresponding to the least relay numbers and transfers the frames through the selected port. Looping of the frames is not sunstrantially caused according to this method and the switch can simultaneously use a plurality of paths.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to data communication, and more particularly to a storage and switching switch mainly used for a network such as Ethernet (registered trademark).

[0002]

2. Description of the Related Art In an Ethernet (registered trademark) or the like, a plurality of communication terminals connected to a network share a communication line and a CSMA / CD (Carrie).
r Sense Multiple Access with Collision Detection)
For example, there is a method of controlling contention avoidance of communication lines by a method such as the above, and a method of allocating a communication line to each terminal or each group of terminals and exchanging frames by a switch device.

[0003] Compared with the former method, the latter method has a higher network construction cost, but can secure a wider communication band per terminal, and has less effective retransmission due to frame collisions. Has the characteristic that it can be increased.

In such a network configuration using switches, in order to prevent infinite transfer of frames, a network topology that does not have a closed loop may be adopted.

FIG. 4 is a network configuration diagram showing an example of a tree-type network. In FIG. 4, three 4-port switches S1, S2, and S3 are connected so as not to form a closed circuit, and the switch S1 has a terminal N
1 and N2 are connected to the terminals N3, N5, N
The terminal N4, N7, and N8 are connected to the switch S3. In such a network, an infinite loop of frame transfer usually does not occur.

As a method of avoiding an infinite loop of frame transfer in a network having a closed circuit, there is a spanning tree protocol. FIG.
FIG. 2 is a network configuration diagram illustrating an example of a network configured in a mesh type. In FIG. 3, for example, switch S1-switch S2-switch S3-switch S5
There are closed circuits such as switch S1 and switch S4-switch S5-switch S6-switch S2-switch S4.

When the spanning tree protocol is used in such a network, a specific path of the above-described closed circuit is used as a main path, and another path is assumed to be an alternative path. For example, when the switch S1-switch S2-switch S3-switch S5-switch S1 is closed, switch S1-switch S2-switch S1
When the main path 3 is used as a main path, the switches S1-switch S5-switch S3, which are alternative paths, are not used unless this main path becomes a failure. In the spanning tree protocol, an infinite loop of frame transfer is avoided in this way.

Also, in the inventions disclosed in JP-A-5-244299, JP-A-5-344120, JP-A-7-131459, and JP-A-8-288982, the number of times a frame is relayed by a relay device is counted. When the number of relays exceeds a predetermined value, it is determined that an infinite loop has occurred, and the transfer of the frame is stopped.

In the invention of Japanese Patent Application Laid-Open No. Hei 8-204879, data to be relayed and identification information for uniquely identifying the data are exchanged between the relay devices, and the relay has already been performed using this identification information. When data is received again by the same relay device, it is determined that the data is in an infinite loop.

[0010]

Each of the above-mentioned prior arts has its own problems. First, in the case of a network configuration that does not create a closed circuit as shown in FIG. 4, for example, when communication from the terminal N1 to the terminal N7 and communication from the terminal N2 to the terminal N8 occur simultaneously, the switches S1 and S3 Since there is only one path between the two, there is a problem in that either communication is kept waiting.

In the case of a network having no closed circuit, there is only one path connecting one node to another node, and there is no alternative path when a failure occurs in a communication path on the path. Therefore, there is a problem that a disabled child cannot communicate between those nodes.

Next, when avoiding an infinite loop by a spanning tree protocol in a network having a closed circuit as shown in FIG. 3, a part of a path forming a loop is an alternative path and is not used at normal times. There is a problem that the communication path is not used effectively. For example, in FIG. 3, when the communication from the terminal N1 to the terminal N5 and the communication from the terminal N2 to the terminal N6 occur at the same time, the path of the switch S1-switch S2-switch S3 and the path of switch S1-switch S5-switch S3 Because only one of these is used, either communication is kept waiting. That is, the main path becomes a bottleneck, and the end-to-end transmission efficiency deteriorates.

Also, a method for detecting a loop by counting the number of relays, as in the inventions of JP-A-5-244299, JP-A-5-344120, JP-A-7-131459, and JP-A-8-288982. Not only requires a counter area in the frame, but also detects it only when a loop actually occurs, so that unnecessary frame transfer cannot be prevented beforehand and network resources are wasted. Problem.

Further, in the method of detecting a loop by uniquely identifying data, as in the invention of Japanese Patent Application Laid-Open No. 8-204879, an overhead of transmitting identification data with identification information added to all data is generated. At the same time, since a loop is actually detected only when it occurs, there is a problem that network resources are wasted similarly to the above method.

The present invention has been made in view of the above circumstances, and when there are a plurality of paths between nodes, these paths can be used at the same time. It is an object of the present invention to provide a storage and exchange switch and a storage and exchange method for relaying a frame so that such a loop is not generated in the first place, instead of detecting the loop.

[0016]

In order to solve the above-mentioned problem, the invention according to claim 1 is a storage and switching switch for relaying a frame in a data communication network, wherein a plurality of switches for transmitting and receiving the frame are provided. A port table that holds the port, the terminal address, the port through which a frame is transmitted to the terminal address, and the number of frame relays until the terminal address is reached after transmitting the frame via the port. Referring to the routing table,
And a transfer unit for sending the frame to a port corresponding to the terminal address of the destination of the frame.

The invention according to claim 2 is characterized in that the transfer means refers to the routing table and sends the frame to a port having the least number of frame relays with respect to the terminal address of the destination of the frame. And

According to a third aspect of the present invention, the transfer unit refers to the routing table when a port corresponding to the terminal address of the destination of the frame is in use, and When another port corresponding to the terminal address is available, the frame is transmitted to the other port.

According to a fourth aspect of the present invention, there is provided a routing table initializing means. The routing table initializing means receives and receives the information of the routing table from another connected storage and exchange switch when the apparatus is started. It is characterized in that information on a reachable terminal address is constructed based on the information, and is written in its own route table.

According to a fifth aspect of the present invention, the store-and-forward switch stores a terminal address, a port for transmitting a frame addressed to the terminal address, and the number of frame relays required to reach the terminal address via the port. A path table for storing, for the terminal directly connected to the switch, the switch stores the terminal address, the port, and the number of relays in the path table based on information held by the switch, For a terminal connected via another store-and-forward switch, the route table information is received from the other store-and-forward switch, and based on the received information, a terminal address, a port, and a relay count are stored in the own route table. When forwarding a frame, refer to the routing table using the terminal address of the destination of the frame as a key. By Rukoto, and gist store-method characterized by sending the frame to the selected port.

According to a sixth aspect of the present invention, when transferring a frame, the route table is referred to by using a terminal address of a frame destination as a key, and a port having the least number of relays for the terminal address is selected. Features.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. When the power supply is turned on, the switch (storage and exchange switch) according to the present embodiment first
By passing information to and from other connected switches, a process of writing the route information to its own route table is performed.

FIG. 1 is a flowchart showing a procedure for initializing the route table when the power is turned on. After the power is turned on, first, in step 1000, the number of switch stages of the network configured in the mesh structure is acquired and stored.

Next, in step 1001, it is determined whether the device of the adjacent node connected through each port is a terminal or a switch. If the terminal is a terminal, the address of the terminal is stored as a reachable address. Also, since the frame can be sent to the terminal without relaying by the switch, the fact that the number of times of passing through the switch related to the address is “0” is stored.
Based on the address information collected in this manner, in step 1002, a route table is created for each port, and information on reachable addresses and the number of times of passing through the switch is written in the table.

If it is determined in step 1003 that the adjacent node is a switch, the processing in steps 1004 to 1007 is performed. In step 1004, the information of the route table of the adjacent switch is acquired for each port, and the information is added to the route table of each port based on the acquired information. Information on the reachable address and the number of times the address has passed through the switch is received from the adjacent switch. When registering this information in the routing table, 1 is added to the number of times of passing through the switch because the number of relays to the adjacent switch required for frame transfer is added. Steps 1006 to 100
By the loop control process of No. 7, the processes of Steps 1004 to 1005 are repeated by the number of switch stages stored in Step 1000.

FIG. 2 is a flowchart showing a procedure for determining a frame transfer destination port while referring to the routing table created in the above procedure. When a frame is received, the destination address written in the frame is read, and in step 2001, it is determined whether the destination address exists on the routing table. If not, step 2002 discards the frame as an error.

Next, in step 2003, it is determined whether or not a plurality of destination addresses of the frame currently being transferred are entered in the routing table, that is, whether or not there are a plurality of frame delivery routes to the destination address. If there is only one such route, in step 2004, the frame is transferred to the port whose destination address is registered as a reachable address. When there are a plurality of routes, in step 2005, a port having the least number of times of passing through the switch is selected from the entries related to the destination address, and the frame is transferred to the selected port.

At the selected port, when the communication line is in use, or when the input / output related resources are in use and cannot be used, if another port having the same number of times of passing through the switch as the port exists, Attempt to transfer to this different port as an alternate route. In this way, when there are a plurality of routes with the same number of times of passing through the switch, even if one route is in use, the frame is transmitted using another route without waiting for the route to be free. be able to.

Next, a specific example of the initialization of the routing table described with reference to FIG. 1 will be described. In this example, it is assumed that the network is configured as shown in FIG. The switches S1 to S6 are turned on almost at the same time.
The switches S1 to S6 perform initialization processing while synchronizing with each other, and acquire a routing table from an adjacent switch.

Each of the switches S1 to S6 is shown in FIG.
, The value of the variable H is set to “3” which is the number of switch stages in this case. Then, in step 1002, a table of the reachable address and the number of times of passing through the switch is created for each port.

FIG. 5 is a diagram showing the values held in the path table of each switch at this time. FIG.
(D) shows the information contents held by the switches S1, S3, S4, and S6, respectively. For example, FIG.
In the path table of the switch S1 shown in (a), two reachable addresses are entered, and the terminal N1 can be reached via the port P1 with the number of relays “0”, and via the port P2. This indicates that the terminal N2 can be reached with the relay count “0”. The same applies to FIGS. 5B to 5D, in which addresses of terminals directly connected to the respective switches are entered.

Since there is no terminal directly connected to the switches S2 and S5, there is no information entered in the routing table at this time.

Next, in step 1004 of FIG.
The switches S1 to S6 obtain the routing table information from each adjacent switch, and in step 1005,
The information is reflected in its own route table. FIG.
(A) and (b) show switches S2 and S2, respectively.
The routing table No. 5 shows the information contents held at this point.

The switch S2 transmits information on the addresses of the terminals N1 and N2 via the port P1 to the switch S1.
, And 1 is added to the number of relays, and registered as “1” in the table. Similarly, information of the terminals N3 and N4 is transmitted from the switch S4 via the port P2, and information of the terminals N5 and N6 is transmitted from the switch S3 via the port P3.
The information of the terminals N7 and N8 is obtained from the switch S6 via the port P4 and registered in the table.

The switch S5 similarly obtains information from an adjacent switch and registers the information in the routing table. Switch S
1, S3, S4, and S6 also receive information from switches S2 and S5, which are adjacent tables, respectively. However, since there is no address registered in the tables of switches S2 and S5, switches S1, There is no information to be newly added to the tables of S3, S4, and S6, and the state is as shown in FIG.

Next, at step 1006, the value of the variable H is reduced by 1 to "2", and again at step 1004.
Are repeated.

FIG. 7 shows information held in the path table of the switch S1 updated by this processing.
The entries 701 and 702 in FIG. 7 are path information for the terminals N1 and N2 originally registered in this table. Also, entries 703-71
0 is information obtained by the switch S1 from the switch S2, which is P3 as port information, and the number of relays is “2” by adding 1 to the number registered in the switch S2.
It has become. The entries 711 to 718 are information acquired and registered from the switch S5.

Switches S3, S4, and S6 also acquire information from switches S2 and S5, respectively, and update their own route tables. Also, switches S2 and S5
Also receives the information from the adjacent switch, but since the route tables of the switches S1, S3, S4, and S6 received at this time are the same as the state received at the previous time, the information additionally registered in the tables of the switches S2 and S5 is Absent.

In step 1006, the variable H
Is decreased, and since H = 1 in step 1007, the processing exits from the loop and completes the initialization of the path table of each switch. here,
Each switch is in a ready state, and starts transferring a frame using the path information created as described above.

When the switch S1 receives, for example, a frame addressed to the terminal N1, the information shown in FIG. 7 is searched, and the port P1 having the minimum number of relays is selected from the entry of the address of N1. . Also, for example, the terminal N
When a frame addressed to port P3 is received, there are two entries with the minimum number of relays being "2".
Alternatively, the frame is transmitted to the port P4. As described above, when there are a plurality of minimum relay count entries, if the port to which transmission is first attempted is busy, transmission retry by another port is performed. Therefore, it is possible to perform switching using network resources effectively by simultaneously using a plurality of paths.

As described above, by using the storage and exchange switch according to the present embodiment, it is possible to select a route that minimizes the number of relays and transfer a frame even in a network configured to include a closed circuit. Is possible,
In addition, when there are a plurality of routes that minimize the number of relays, not only at the time of failure but also all of the routes are used, and the utilization efficiency of network resources increases.

Next, suppression of a frame loop by the method of the present invention will be described. FIG. 8 is a network configuration diagram showing a connection form of a part of a network configured using the above-described switches.

The minimum number of relays to the destination terminal N11 can be either the case where there is no closed path as shown in FIG. 8A or the case where there is a closed path as shown in FIG. 8B. Either it has a difference of 1 between two adjacent switches or it is equal between two adjacent switches.

When there is a difference of 1 between adjacent switches, the frame addressed to the terminal is always transferred from the one with the largest number of minimum relays to the one with the smallest number of relays, and is not transferred in the reverse direction. Therefore, each time the frame is transferred once between the switches, the required number of relays until the frame reaches the destination terminal monotonously decreases. Thus, during the transfer of a frame,
It never goes through the same switch twice, so no frame loops occur.

There is a possibility that the minimum number of relays may be equal between adjacent switches when the switches are on a closed circuit, for example, as in switches S11 and S21 shown in FIG. 8B. . In such a case, for example, from the viewpoint of the switch S11, there is a switch S12 that is a transfer destination having a smaller number of relays than the switch S21, and a frame addressed to the terminal N11 that has reached the switch S11 is not a switch S21. The data is transferred to the switch S12. Therefore, there is no possibility that a frame loop occurs between the switches S11 and S21.

In the above embodiment, each switch creates and maintains a path table for each port. However, it is not necessary that each switch has a port table. The table structure of another form which can hold the relationship of the table may be used.

The above-described switch and switching method are applicable to various store-and-forward networks. For example,
When applied to the transfer of an Ethernet frame, the header of the Ethernet frame contains a MAC (Media Access Control).
trol) address stores a destination hardware address and a transmission source hardware address. The switch creates and maintains a route management table based on the MAC address, and transfers a frame.

[0048]

As described above, according to the present invention, when a switch is activated, information is obtained from an adjacent node to create a routing table, and the routing table stores the relay count for each port and each destination terminal address. Is held, and a route that minimizes the number of relays is selected at the time of frame transfer, so that the frame can be transferred without looping.

Further, a loop is detected by detecting a loop when a count value of a counter or the like exceeds a threshold value as in the prior art, or by adding identification information to a frame and recording the identification information during relaying. As described above, since the method does not generate a frame loop in the first place as described above, it is possible to prevent wasteful use of network resources until the loop is detected.

Further, a switch is configured in a mesh type,
Even if the network has multiple routes to the destination, instead of using the alternative path until the main path fails, unlike the spanning tree protocol of the prior art, these multiple routes are simultaneously used. Since it is used, it is possible to alleviate a bottleneck at a specific place, and it is possible to use network resources efficiently.

[Brief description of the drawings]

FIG. 1 is a flowchart showing a procedure for initializing a routing table according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a procedure for determining a frame transfer destination with reference to a route table according to the embodiment;

FIG. 3 is a configuration diagram showing a mesh network configuration using switches according to the embodiment;

FIG. 4 is a configuration diagram showing a network configuration having no closed circuit.

FIG. 5 is a diagram showing information contents held by a route table according to an embodiment of the present invention.

FIG. 6 is a diagram showing information contents held by a route table according to the embodiment;

FIG. 7 is a diagram showing information contents held by a route table according to the embodiment;

FIG. 8 is a configuration diagram showing an example of a connection configuration of switches according to the embodiment.

[Explanation of symbols]

N1 to N8, N11 Terminal P1 to P4 Port S1 to S6, S11 to S13, S21, S22 Switch

Claims (6)

[Claims] 1. A storage and switching switch for relaying frames in a data communication network, comprising: a plurality of ports for transmitting and receiving frames; a terminal address; and a port via which frames are transmitted to the terminal address. A route table that holds the number of times the frame is relayed until the terminal address is reached after sending the frame via the port, and referring to the route table, the port corresponding to the terminal address of the destination of the frame is assigned to the port. A storage and exchange switch, comprising: transfer means for transmitting a frame. 2. The transfer unit according to claim 1, wherein the transfer unit refers to the routing table and sends the frame to a port having the least number of frame relays with respect to the terminal address of the destination of the frame. Storage switch. 3. The transfer means refers to the route table, and when a port corresponding to the terminal address of the destination of the frame is in use, the transfer unit performs another processing according to the terminal address of the destination of the frame. 3. The storage and switching switch according to claim 1, wherein the frame is transmitted to another port when the port is available. 4. A routing table initializing means, which receives information of the routing table from another connected storage and exchange switch when the apparatus is started, and can be reached based on the received information. 4. The storage and exchange switch according to claim 1, wherein information relating to the terminal address is configured and written into its own route table. 5. A storage-and-forward switch comprising a path table for storing a terminal address, a port for transmitting a frame addressed to the terminal address, and the number of frame relays required to reach the terminal address via the port, The store-and-forward switch stores a terminal address, a port, and the number of relays in the routing table based on information held by the store-and-forward switch based on information held by the store-and-forward switch. With respect to the terminal connected via, the information of the routing table is received from the other storage and switching switch, and based on the received information, the terminal address, the port, and the number of relays are stored in the routing table of the terminal itself, and the frame transfer is performed. In this case, by referring to the routing table using the terminal address of the destination of the frame as a key, A store-and-forward method, wherein the frame is transmitted to a selected port. 6. The frame transfer device according to claim 5, wherein the frame table is referred to by using the terminal address of the destination of the frame as a key, and a port having the least number of relays for the terminal address is selected. The store-and-forward method described.