JP2006245789A - Multicast output control method and system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a multicast output control method and system for preventing a high load of a network when a loop occurs, and also, for reducing effects due to erroneous connection or the like by some users on the whole network regarding the multicast output control method and system. <P>SOLUTION: When a plurality of terminals are grouped and connected with the network and transmission/reception of data are executed between a plurality of the terminals, the system is composed so as to detect the loop by comparing the predetermined threshold with the number of returning times of a packet by a comparison means, after providing a loop detection means for detecting that the packet returns to a prescribed port and the comparison means for comparing the output of the loop detection means with the predetermined threshold in the terminals. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a multicast output control method and system, and more particularly to a loop detection method and multicast output control method for a device mounted in a network using an Ethernet (registered trademark of Fuji Xerox Co., Ltd.) switch. The present invention relates to a multicast output control method and system.

  FIG. 4 is a block diagram showing a conventional network configuration example. In the figure, reference numeral 1 denotes a customer edge portion 1, 2 denotes a customer edge portion 2, which is divided into groups X, Y, and Z. The reason for being divided into groups is to prevent data from being erroneously transferred to other groups. 3 is a wide area network, and 4 is an edge switch device connected to the wide area network 3. The customer edge devices 1 and 2 are connected to the edge switch 4. In the wide area network 3, 3a is a core switch.

  In the figure, the group ID (X, Y, Z) of the customer edge unit 1 is forwarded by the edge switch device 4 to the end user side of the customer edge unit 2 by the forwarding process. A unique address (MAC, IP) is assigned to a terminal such as a personal computer (PC) belonging to the same group, and the edge switch device 4 determines an output route based on this address.

  FIG. 5 is a block diagram showing a configuration example of a conventional edge switch device. In the figure, 6 is a reception line interface unit for receiving input traffic from the opposite device, 7 is a transmission line interface unit for transmitting traffic to the opposite device, 8 is an L2SW (switch) engine for switching traffic at layer 2, 12 Is a maintenance operator operation terminal used for device control, 9 is connected to the maintenance operator operation terminal 12, interfaces the operation from the terminal 12 in the device, and outputs in-device information to the terminal 12, 10 is a device control monitoring unit that monitors all information inside the device, and 11 is a memory that has functions of controlling the L2SW engine 8 and storing the current L2SW engine information in response to a request from the device control monitoring unit 10.

  FIG. 6 is an explanatory diagram of a conventional learning method using a learning table. The same components as those in FIG. 5 are denoted by the same reference numerals. In the figure, 8a is a learning table provided in the memory of the L2SW engine 8. The learning table 8a includes an address, a group ID, and a port (PORT) number. The address (hereinafter abbreviated as SA (source address)) in the frame received from the reception line interface unit 6, the received port number, and the group ID are stored in the learning table 8a.

  In this state, when a frame is received from another port, the learning table 8a is searched from the received destination address (destination address: hereinafter abbreviated as DA) and group ID information, and a frame is found at the corresponding port hit. Forward. Here, “hit” means a case where the learning table 8a is searched for the destination address (DA) of the input packet and the DA exists in the learning table 8a.

  If there is no hit in the learning table 8a, it is assumed that a terminal is newly connected, and multicast transfer is performed in order to notify the presence of the terminal in the same group ID. If the learning table 8a is hit but the port numbers are different, it is assumed that the terminal has moved between the ports, and in this case also, multicast transfer is performed to update the learning table in another switch SW.

  Here, the port number is different from the port number corresponding to the learned DA when the port number into which the packet is actually input is different. For example, the port number is different when the SA of the terminal a is input from the port 1 in a state where learning of a: port0 (the terminal of DA = a is connected to the port 0) is performed. The term “terminal moves between ports” means, for example, a case where a terminal connected to port 0 is reconnected to port 1.

  As a conventional system of this type, a frame data relay device suitable for use in frame data communication between the first and second routers accommodating a narrow-band network using Ethernet is designed to improve line traffic. This technique is known (see, for example, Patent Document 1). In addition, a technique for restrictively connecting domains while ensuring security and maintenance independence is known (see, for example, Patent Document 2). Also, in a relay device that relays between multiple domains by address translation, even if inter-domain and intra-domain communication are mixed, communication that ensures security is performed without setting complicated filters between domains. In the domain, a technique for relaying packets at high speed is known (see, for example, Patent Document 3).

Also, as a prior art, the FCS learning processing unit stores the FCS generated based on the contents of the MAC frame in the FCS learning CAM 16, and when receiving the MAC frame, the FCS learning processing unit is based on the stored FCS information. And determining whether or not the communication path related to the MAC frame is in a loop, and when it is determined that the communication path related to the MAC frame is in a loop state, the filtering processing unit relays the MAC frame. There is a technology to limit (Japanese Patent Application No. 2004-057849).
JP 2003-324475 A (paragraphs 0021 to 0027, FIGS. 1 to 3) JP 2001-352337 A (paragraphs 0024 to 0032, FIGS. 1 to 3) JP 2001-292163 A (paragraphs 0035 to 0056, FIGS. 1 to 3)

  In the conventional method, if there is a movement between input ports, the source address is different, so flatting is performed, and multicast packets are generated simultaneously for the source address × the same domain, so the performance of the entire apparatus deteriorates. There is a problem. Here, the domain is a range in which a packet can be transmitted according to a common routing control rule in the network layer.

  The present invention has been made in view of such a problem, and can prevent a high network load when a loop occurs, and can reduce the influence on the entire network due to misconnection of some users, etc. An object is to provide an output control method and system.

  In the present invention, in order to suppress an increase in unnecessary traffic due to the occurrence of a loop, the learning information of the SA and the input port is compared to detect that the output packet has returned to the port, and the frequency is set to a certain threshold value. The loop is detected by monitoring the original. At the time of loop detection, an output threshold is provided to suppress packets output as multicast within the same group ID. Further, by excluding other multicasts from the main monitoring, communication using multicast such as broadcast communication is not hindered.

  (1) According to the first aspect of the present invention, when a plurality of terminals are grouped and connected to a network and data is transmitted / received between the plurality of terminals, an output packet is returned to a predetermined port. A loop is detected by detecting and comparing the predetermined threshold with the number of times of return.

  (2) The invention according to claim 2 is characterized in that the fact that the output packet has returned to a predetermined port is determined by comparing learning information of a source address (SA) and an input port (port). To do.

(3) The invention described in claim 3 is characterized in that when the loop is detected, a packet output as a multicast in the same group is suppressed.
(4) The invention described in claim 4 is characterized in that other multicasts are excluded from this monitoring.

  (5) In the invention according to claim 5, when a plurality of terminals are grouped and connected to a network and data is transmitted / received between the plurality of terminals, the packet returns to a predetermined port in the terminal. A loop detecting means for detecting the occurrence of the error and a comparing means for comparing the output of the loop detecting means with a predetermined threshold value, and comparing the predetermined threshold value with the number of times of return by comparing means. , And is configured to detect a loop.

  (1) According to the first aspect of the present invention, it is possible to detect that an output packet has returned to a predetermined port, and to detect a loop by comparing a predetermined threshold with the number of times of return. .

  (2) According to the invention described in claim 2, it is possible to determine that the packet has returned to the port by comparing the learning information of the source address (SA) and the input port (port).

(3) According to the invention described in claim 3, when a loop is detected, it is possible to prevent the network from becoming a heavy load by suppressing packets.
(4) According to the invention described in claim 4, by excluding other multicasts from the main monitoring, it is possible to prevent the communication using the multicasts from being hindered.

  (5) According to the invention described in claim 5, it is possible to detect that an output packet has returned to a predetermined port, and to detect a loop by comparing a predetermined threshold with the number of times of return. .

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a diagram showing a network image, and is a simplified illustration of FIG. The same components as those in FIG. 4 are denoted by the same reference numerals. The configuration of each edge switch 4 is the same as that shown in FIG. In the figure, reference numeral 20 denotes a terminal, which shows a terminal A and a terminal B. Reference numeral 4 denotes an edge switch device, which shows SW1 to SW4. Terminal A and SW1 are connected via port 0, SW1 and SW2 are connected via port 3, SW1 and SW3 are connected via port 1, and SW1 and SW4 are connected via port 2.

Consider a case in which a packet is transmitted from terminal A to terminal B in the system configured as described above. in this case,
1. Terminal A → SW1 → SW4 → Terminal B
2. Terminal A → SW1 → SW2 → SW4 → Terminal B
3. Terminal A → SW1 → SW3 → SW4 → Terminal B
Is selected and communication is performed.
When a loop occurs, for example, a packet input from terminal A to port SW1 via port 0 circulates on the network in the order of SW3->SW4->SW1->SW3->SW4->SW1->, causing an increase in traffic.

  FIG. 2 is a diagram showing a packet processing flow of the present invention. First, a packet is input from terminal A through port 0 (S1). The device control monitoring unit 10 searches the learning table 8a (S2). When a loop occurs, the same packet as the packet input from terminal A to port 0 of SW1 (when the loop shown in the previous example occurs) is input from port 2. This is the same as the state in which the terminal described in the related art moves between ports.

  Take the L2SW engine 8 of FIG. 6 as an example. When a packet having SA: A1 is input from the terminal A (S1), the learning table 8a is searched (S2). The device control monitoring unit 10 checks whether learning has been completed (S3). When learning is not performed, the learning table 8a learns SA: A1, port number 0 (S4: sequence 1).

  If it has been learned, it is checked whether it is the same as the learned input port information (S5). If the input port information is the same as the learned input port information, normal communication processing is performed (S8). That is, after that, when the SA: A1 packet is input from the terminal A, the process of sequence 3 is performed.

  If it is not the same as the learned input port information in step S5, when this packet is looped and inputted from port 2, sequence 2 is performed to learn SA: A1 and port number 2. Thus, when a loop occurs, not only the packet circulation is continued, but also the terminal movement is detected (the same SA packet is input from port 2 next to port 0. The same applies to port 0 next to port 2). 2 is implemented. That is, in sequence 2, port information is relearned (S6), and multicast output is performed as unlearned (S7). In this case, a multicast packet is output for learning, resulting in an increase in traffic.

  Based on the above, a monitoring cycle of terminal movement between ports is provided, and the number of movements between terminals of the same SA in the same SA is monitored and counted, and a predetermined threshold is compared to detect loop occurrence. To do. That is, when the number of movements between ports becomes larger than a predetermined threshold, it is determined that a loop has occurred. In this case, when a loop is detected, the multicast packet output is suppressed.

  Thus, according to the present invention, it is possible to detect a loop by detecting that an output packet has returned to a predetermined port and comparing a predetermined threshold with the number of times of return. Further, according to the present invention, it is possible to determine that the packet has returned to the port by comparing the source address (SA) with the learning information of the input port. Further, when a loop is detected, it is possible to prevent the network from becoming a heavy load by suppressing packets.

In this case, regarding other multicasts, by excluding them from the main monitoring, communication using multicasts can be prevented from being hindered.
Next, a specific embodiment of the multicast suppression method using the loop detection function will be described. FIG. 3 is a diagram showing a time chart embodiment showing the same SA detection port. 3 shows an example in which a packet having the same SA is input to the described port at the timing shown in FIG. 3 based on the packet processing flow shown in FIG. 2 and the network image shown in FIG.

  In this example, packets having the same SA (A1 as an example) are input from port 0/2. For example, one monitoring cycle is 1 second. In the period A, since the packet having SA: A1 is input only to the port 0, it is determined that no loop has occurred (the operation of the sequence 3 in FIG. 2 is performed).

  In the next cycle B, an example in which packets having the same SA are input in the same monitoring cycle is shown. For example, if it is determined that a loop is detected when movement between ports is detected twice in the same monitoring period, unnecessary traffic increase is suppressed by stopping multicast transmission by the same SA during the remaining relevant monitoring period (for example, the remaining time of period B). To do.

  In the embodiment diagram (FIG. 2) of the packet processing flow, this is realized by counting the number of times the processing of the sequence 2 is performed in a certain period. The multicast packet output stop when the loop is detected is implemented by skipping “unlearned multi-packet output”.

  As described above in detail, according to the present invention, it is possible to prevent a high network load at the time of a loop occurrence, reduce the influence on the entire network due to erroneous connection of some users, It is possible to avoid the worst situation such as the above, and contribute to the high reliability of the network. Further, by performing loop detection at each port, the network can be repaired at an early stage.

It is a figure which shows a network image. It is a figure which shows the packet processing flow of this invention. It is a figure which shows the time chart Example which shows the same SA detection port. It is a figure which shows the example of a conventional network structure. It is a block diagram which shows the structural example of the conventional edge switch apparatus. It is explanatory drawing of the conventional learning method using a learning table.

Explanation of symbols

6 receiving line interface unit 7 transmitting line interface unit 8 L2SW engine 8a learning table 9 user interface unit 10 device control monitoring unit 11 memory 12 maintenance operator operation terminal

Claims (5)

In the case where a plurality of terminals are grouped and connected to a network and data is transmitted / received between the plurality of terminals,
Detect that the output packet has returned to a given port,
A multicast output control method, wherein a loop is detected by comparing a predetermined threshold with the number of times of return.   2. The multicast output control method according to claim 1, wherein the return of the output packet to the predetermined port is determined by comparing learning information of the source address (SA) and the input port (port).   2. The multicast output control method according to claim 1, wherein when the loop is detected, a packet output as a multicast in the same group is suppressed.   4. The multicast output control method according to claim 3, wherein other multicasts are excluded from the main monitoring. In the case where a plurality of terminals are grouped and connected to a network and data is transmitted / received between the plurality of terminals,
Loop detecting means for detecting that a packet has returned to a predetermined port in the terminal;
Comparison means for comparing the output of the loop detection means with a predetermined threshold;
A multicast output control system configured to detect a loop by comparing a predetermined threshold value with the number of times of return by a comparison means.

Images