JPH0998185A - Avoiding method for loop of data in network - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve network efficiency by resetting a TTL field in the IP header of a transmission packet at a transmission station into the irreducibly minimum value to arrive at a reception station. SOLUTION: When transferring a packet from a transmission station 6 connected to a certain network through gateways 4 and 5 to the reception station of the other network, while referring to a route control table at the transmission station 6, the number of hops to the network connecting the reception station is possessed. The TTL field in the IP header of a transmission packet is reset into a fixed value such as not '60' but a value adding '1' to the possessed number of hops. Based on this value, usual IP packet transfer is executed and each time the packet passes through the gateways 4 and 5, '1' is subtracted from the TTL. Even when the packet is looped in the network, that loop is found out by the minimum number of times of processing and immediately abandoned.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data loop avoiding method for preventing a loop of a packet flowing in a local area network or a wide area network.

[0002]

2. Description of the Related Art LAN (Local Area Network:
Local Area Network) and WAN (Wide Area Network:
When a packet is transferred from a source station to a destination station on a Wide Area Network), T developed for the purpose of internetworking is one of the means.
CP / IP (Transmission Control Protocol / InternetP
rotocol: a protocol corresponding to the transport layer / network layer).

In the transfer using TCP / IP, when the transfer packet cannot reach the destination station due to disconnection of the transmission line, erroneous static route change, etc., the transfer packet loops between the gateways for a long time. May cause Such a transfer packet increases unnecessary traffic in the network including the gateway and causes a decrease in network usage efficiency.

Therefore, in order to avoid an endless loop of data in the IP header of the IP datagram in the transfer packet,
A field called TTL (Time to Live: one field of the IP header, which indicates how many packets can exist on the Internet in seconds) is prepared. FIG. 4 is an explanatory diagram showing the structure of the IP header of the IP datagram including the TTL field.

First, a transfer packet is generated at the transmission source station. At this time, an IP header is attached and a TTL value is given. In normal TCP / IP communication, a TTL value of "60" is given. The TTL value of the IP packet that has left the source station decreases by "1" as the gateway performs the relay process. When this gateway receives a packet with TTL = 1, it is relayed to TTL = 0 and is removed before being sent to the network. In the above TCP / IP communication, the packet is destroyed before passing through the 60th gateway.

Each gateway determines a relay destination station by a route control table (a route control table is a database of route information acquired dynamically or statically, and is also called a routing table). This routing control table is held in the kernel of the host or gateway existing in the network, and is a routing information protocol (RIP: Routin) provided as a UNIX standard.
g Information Protcol). The above gateways perform RIP transmission between adjacent gateways at regular intervals to exchange route information and detect data collisions.

FIG. 5 refers to a network number i (DISTANCE OF NETWORK: a count value of the distance of a specific network) in a RIP frame to create a field called a metric of a routing table in the kernel.
The metric field of the route control table indicates the number of gateways through which a packet sent from a transmission source station reaches a reception destination station (the number of hops: packet is 1
When a plurality of transmission routes exist, the minimum number of hops of each route is shown.

[0008]

However, according to the conventional network system, an infinite loop of packets can be avoided in the network by using the TTL of the IP header. But normal TC
In communication using P (Transmission Control Protocol), the initial value of TTL is set to 60. Therefore, if a packet causes a loop, it cannot be discarded until it passes through the gateway 60 times. On the other hand, for example, it is conceivable to set the TTL to an appropriate small value, but this cannot be done because the packet may be discarded before reaching the receiving station.

Further, in the conventional system, the TT is set at the receiving station.
Since L was not confirmed, even if the transmitted packet reached the destination station via an unintended route, it could not be found. Therefore, it is an object of the present invention to provide a method for avoiding data loops in a network, which makes it possible to use an optimum TTL value and can significantly improve network usage efficiency.

[0010]

In order to achieve the above-mentioned object, the present invention is connected to one network and R
When a packet of information is transferred from a transmitting station supporting IP to a receiving station of a different network via a gateway, the receiving station is connected by referring to the routing control table existing in the kernel of the transmitting station. The number of hops to the different network is acquired, and the TTL field in the IP header of the transmission packet is set again with a value obtained by adding 1 to the acquired number of hops.

According to this method, the TTL in the IP header is
The numerical value set in the field is rewritten with a value obtained by adding 1 to the number of hops in the routing table existing in the kernel of the transmitting station. As a result, the data loop time can be reduced and the network efficiency can be significantly improved.

[0012]

1 is a flowchart showing a method for avoiding data loops in a network according to the present invention. FIG. 2 is a block diagram showing an example of a network (Ethernet: Ethernet) to which the data loop avoidance method in the network of the present invention is applied. Further, FIG. 3 is an explanatory diagram showing changes in TTL.

In the network described below,
From the source terminal device H1 to the destination terminal device H5, T
A case where packet transfer is performed using CP / IP will be described. In addition, in the present embodiment, the terminal device H1 is assumed to support the above-mentioned route information protocol RIP. As shown in FIG. 2, the Ethernets 1, 2, and 3 are connected to each other by gateways 4 (G1) and 5 (G2). The terminal device 6 (H
1) and 7 (H2) are connected, Ethernet 2 is connected to terminal devices 8 (H3) and 9 (H4), and Ethernet 1 is connected to terminal devices 10 (H5) and 11 (H6). .

In the network having the above configuration,
When the application of the terminal device 6 (transmission station) needs to transmit a packet (step 101), the time to live TTL of this packet is given inside the terminal device 6 (step 102). At this time, the hop number "2" to the network in which the terminal device 10 exists is acquired by referring to the metric field of the route control table in the kernel of the terminal device 6 (step 103). The number of hops when the network where the destination terminal device exists can be reached by passing through one gateway is "1", but there is a restriction that a packet cannot be transmitted to that network unless the TTL is "2" or more. Therefore, the value obtained by adding 1 to the hop count is substituted into the TTL of the IP header (step 104). By this processing, the TTL given to the IP header of the packet transmitted from the terminal device 6
The value becomes "3".

After that, the gateway executes the packet relay until the terminal device 10 is reached by the transfer processing by the normal IP (step 105). At this time, the gateway decrements the TTL value by one each time it is passed (step 106). In the case where normal routing is performed, as shown in FIG. 3, when the processing in the gateway 5 is completed, TTL becomes 1 (step 107), and the packet is transferred to the terminal device 10 of the receiving destination (step 11).
3).

By the way, when an IP header of TTL = 1 is transferred, normal routing is not performed, and there is no receiving destination terminal device in the network connected to the gateway, or a packet of TTL = 1 is transmitted. When received by the gateway, when a loop occurs, or when there is an error in the route information (step 1
08), the gateway immediately discards the data (step 10) to reduce traffic in the network.
9) The RIP frame is transmitted to update the route information (step 110). In this way, when a defect is recognized at the time of TTL = 1, the packet is discarded, so that the time involved in the loop can be significantly shortened. Incidentally, conventionally, the TTL value is set to 60, and the packet is not discarded unless it passes through the gateway until it becomes 0. For this reason, a lot of time is involved in the loop.

When the transferred data has successfully arrived at the intended receiving station (terminal device 10) (step 113),
It is confirmed whether it is "1" by referring to the value of TTL (step 114). If it is not “1” (= 1
(Other than the above), since there is an error in the route control, the RIP frame is immediately transmitted in the network (step 11).
0). In this way, if the reason for immediately transmitting the RIP frame is that the data was discarded not because of collision between the data but because the TTL became 0, it is considered that there is an error in the routing control table. . Therefore, the RIP frame is immediately transmitted without waiting for the transmission of the next RIP frame so that the route can be repaired more quickly. After that, new route information is obtained (step 111), and the route control table is rewritten based on this (step 112).

[0018]

As is apparent from the above, according to the present invention, when transmitting a packet of information from a transmitting station connected to a certain network and supporting RIP to a receiving station of a different network via a gateway. , The hop number to the network to which the receiving station is connected is acquired by referring to the routing control table existing inside the kernel of the transmitting station, and the TTL field in the IP header of the transmission packet is set to 1 for the acquired hop number. Since the setting is performed again with the value added with, even if a packet in the network causes a loop, it can be immediately detected, and the network efficiency can be greatly improved.

Further, by checking the TTL value of the received data, it is possible to confirm whether or not there is an error in the path control table (error if other than TTL = 1).

[Brief description of drawings]

FIG. 1 is a flowchart showing a method for avoiding a data loop in a network according to the present invention.

FIG. 2 is a configuration diagram showing an example of a network to which the data loop avoidance method in the network of the present invention is applied.

FIG. 3 is an explanatory diagram showing changes in TTL.

FIG. 4I of an IP datagram containing a TTL field
It is explanatory drawing which shows the structure of P header.

FIG. 5 is an explanatory diagram showing a configuration of a RIP frame.

[Explanation of symbols]

 1,2,3 Ethernet 4,5 Gateway 6,7,8,9,10,11 Terminal device

Claims (1)

[Claims] 1. RIP connected to one network
In a network system which transfers packets of information from a transmitting station supporting (Routing Information Protocol) to a receiving station of a different network, by referring to a routing control table existing inside the kernel of the transmitting station, Data in a network characterized in that the number of hops to the different network to which the receiving station is connected is obtained, and the TTL field in the IP header of the transmission packet is reset by a value obtained by adding 1 to the obtained number of hops. Loop avoidance method.