JPH0346432A - Packet information transfer system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Industrial Field of Application J] The present invention relates to a packet information transfer method, and particularly to distributed communication with a large system scale that requires relaying through many nodes when transferring one packet to a destination. In the net,
This invention relates to a packet information transfer method that makes it possible to significantly reduce information transfer time (delay time).

[Conventional technology j In the conventional store-and-forward technology, in a distributed communication network where each node is connected to multiple input lines and multiple output lines, a terminal connected to one node is connected to another node. l\When transferring information.

The nodes in between operated as follows.

(1) Receive one entire packet.

(2) Check the routing information.

(3) Determine output links and new routing information.

(4) Transfer the packet to the transmission buffer of the determined output link.

Regarding this, see 1. For example, ``Packet Switching Technology and its Applications'' edited by the Institute of Electronics and Communication Engineers (published by Corona Publishing, first edition in 1972).
(Published in 2013) is helpful.

[Problems to be Solved by the Invention] In the above-mentioned conventional technology, packets from a plurality of input links are transferred to each output link, and stochastic forest waiting occurs. Therefore, in a communication network with a large number of nodes, there are many nodes that one packet must pass through while being transferred between departure and arrival, and each time a packet passes through a node, delay time is accumulated to satisfy quality requirements for delay time. The problem was that it was difficult to do so.

The present invention has been made in view of the above circumstances, and its purpose is to solve the above-mentioned problems in the conventional technology, and to improve system scale in a communication network configured by connecting nodes in an arbitrary shape. An object of the present invention is to provide a packet information transfer method that makes it possible to keep the delay time of information transfer within a network to a small value even when the number of nodes is large and the number of nodes is large.

[Means for Solving the Problems] The above-mentioned object of the present invention is to provide a communication network in which a plurality of nodes are connected to each other by links, and information is transferred in a time-sharing manner in units of information elements including routing information. By determining in advance the correspondence between one of the links input to the node and one of the links output from the node, multiple paths can be defined that connect some nodes in the network. At the same time, a routing table including whether or not the path is used is arranged for each output link of each node, and each node inspects the routing information in the information element received from the input link, and determines whether the routing information is the first one. If the routing information is of the first type, the information element is continued to be transferred along the path to which the input link belongs, and if the routing information is of the second type, the information element is extracted and the corresponding time interval is When a free state occurs in the path to which the link belongs, an output link to be sent next is determined for the extracted information element, and is accumulated in a buffer corresponding to the output link, and a time period in the free state occurs in the path to which the link belongs. This is achieved by a packet information transfer method characterized by transmitting the stored information.

[Effect]

In the packet information transfer method according to the present invention, a path consisting of consecutive links for non-storage transfer is determined in advance in the network, and this path is actively used to transfer packet information. 1 between departure and arrival by making it happen! This is designed to reduce the number of I-product relays.

That is, a feature of the present invention is that, unlike the conventional method, when an information packet is transferred, it is not configured to uniformly pass through a transmission buffer in all nodes.

[Example] Hereinafter, an example of the present invention will be described in detail based on the drawings.

FIG. 2 shows an example of the configuration of one communication network to which the packet information transfer system according to the present invention is applied. In the figure, AJ indicates a node, and Ll-L21 indicates a link. At each node, input links and output links have identifiers (a,
b, c, . . ), and the setting is such that a pair of an input link and an output link having the same identifier is regarded as a trunk line.

In Figure 1, examples of such settings are shown by thick lines within each node. For example, a path consisting of a sequence of nodes D-H-E-IJ is one main line.

On the trunk mentioned above, fixed length information packets are transferred in time slotted frames. The frame structure on the node jump link is shown in FIG. Each frame is 1. It consists of a time interval of two elements. In the first time interval 21, an empty status signal is transferred to indicate whether a valid information packet is present in the following time interval. In the second time interval 22, when the signal of the first time interval 21 indicates a "blocked" state, an information packet is transferred, and when the signal indicates an "empty" state, an invalid signal is transferred. be done.

In addition, in the following explanation, the link L in FIG.
l-L21 will be explained as a unidirectional transmission line in which information is transferred only in the direction of the arrow, but in the following explanation, even if each ring is configured as a bidirectional transmission line, it is assumed that there are individual transmission lines in each direction. It can be easily applied by assuming that there exists.

As shown in FIG. 4, the packet structure has a fixed length and is divided into a packet header section 31 and a packet information section 32. The packet header section 31 has routing information arranged therein. Here, it is assumed that the destination node address is used as the routing information.

FIG. 5 is a diagram showing an example of the configuration of each node in FIG. 2. This example shows a configuration example in which two trunk lines are set per node.

In the figure, 100 indicates an input link from another node, and 104 indicates an output link going to another node. Also, 1
Reference numerals 13 and 114 respectively indicate an input link through which packets from a terminal device (not shown) housed in the node are input, and an output link through which packets are output to the terminal device.

Further, lot is an address analysis section provided corresponding to each input link, and has a corresponding routing table +20. Taking the routing table for the input link a of node B in FIG. 2 as an example, its configuration is shown in FIG. The three items in each row of the table shown in Figure 6 are "destination node address", "direct",
"Output link" is treated as a unit. This set of items determines, when an information packet with a header matching the destination node address is received, whether or not to pass it directly through that node (corresponding values of Y and H are set). (2) Indicates from which output link the transmission is to be made (the identifier of the corresponding output link is set).

The operation of the address analysis unit 101 described above is as follows. When a frame starts to be input from input link 100, the empty block state display 2 at the beginning of the frame is first displayed! Check for "empty" frames. If the frame is "blocked", the routing table described above is searched using the subsequently received destination node address value as a key, and a set of items with matching address values is extracted. As a result, if the value of the direct field is "YJ", the idle state indication and information packet of the received frame are transferred to the output terminal as is, and if the value of the direct field is rNJ, the received information packet is After adding the output link identifier, it is transferred to the output terminal ■.The structure of the information packet transferred to the output terminal ■ at this time is as follows.

It is shown in FIG.

In FIG. 7, the first element 61 is an output link identifier;
Second 1! ! The element 62 is the human power information packet shown in FIG. Furthermore, when the frame is r'''!J and when the item of the r direct J in the routing table is rNJ, the frame with the empty state display 2I set to the empty state is output to the output terminal. Output Terminal ■ is connected to multiplex #103 via direct path 1.02, and output terminal ■
The first distribution unit 1 connected to the distribution unit 1 via the transfer path 110 receives the information packet from the transfer path +1 (l), and based on the output link information 61 in the packet, The information packet is transferred to the output buffer 112 corresponding to the specified output link 104.

The output buffer 112 is connected to each distribution section! The information packets input from 11 are accumulated in the order of arrival, and according to the read operation from the multiplexing IWI03, the oldest information is taken out one by one and transferred to the multiplexing unit +03.

Multiplexing 1171+ 103 is the address analysis section! Monitors the empty blockage status display at the beginning of the frame input from 01,
Unless there is a notification of a vacant state, the input frame from the direct path +02 is transferred to the output link 104 as is. At this time, a "blocked" state is displayed in the output frame. When the empty/busy state indication of the input frame indicates a free state, the output buffer 112 is inspected, and if an information packet exists there, it is extracted and the frame is sent to the output link as a "busy" state indication, and the information packet is If it does not exist, it outputs an invalid signal to the output link with the frame as an "empty" status indication.

The details of the operation of this embodiment will be described below, taking as an example the case where the information packet shown in FIG. 4 is transferred from node A to node j in FIG. 2. Here, the value "J' is set in the packet header, and the routing table of each node is
It is assumed that data is set to be transferred through a route between nodes A-BE-I-J. Specifically, in the routing table of each node, the destination node address is node J.
This means that the entry corresponding to is set as shown in FIG.

Here, it is assumed that an information packet is input from a terminal housed in node A along input link +13. The address analysis unit 101c determines that the destination node address is J
”, the routing table 120c is referred to and the data in the first row of FIG. Then, in the distribution department 1c, based on the output link identifier being a,
The received information packet is transferred to the output buffer l12a corresponding to the output link 104a. Multiplexing unit 103
In a, information packets are sequentially extracted from the output buffer 112a and output from the output link 104a every time there is a notification of a vacant state from the address analysis unit 101c. As a result, the packet is guided to the a input of node B through link L4 shown in FIG.

At node B, assume that the input link is 100a. At this time, the information packet is address analysis unit +01
a, refer to the routing table 120a, and
Get row data. As a result, the information packet will be forwarded along the direct path 102a. in this case,
Since the idle state signal of the frame on the inter-node link is "blocked" and there is direct communication, no idle state notification is performed. Therefore, in the multiplexing unit 103a, the direct path 102a
The input from the output link 104a is transferred to the output link 104a. In this case, since the packet is transferred without storage along the main line, the series of operations from the time the head of the packet arrives at the node until the head of the packet is output from the node is the sum of the header reception time and analysis time. Completed in time. However, the latter process refers to the process required to refer to the routing table and determine its contents, and can be executed in a sufficiently shorter time than the former. As a result, the packet will be guided to the a input of node E through link L8.

In node E, the human link is assumed to be 100a,
This will be explained below using FIG. First, the information packet is sent to the routing table 120 by the address analysis unit 101a.
With reference to a, the data shown in the third row of FIG. 8 is obtained. As a result, the information packet is determined to be transferred along the transfer path 102a, and after adding the output link identifier b, the information packet is transferred to the transfer path Oa, as shown by ■ in FIG. . At this time, address analysis section 1
0Ia notifies the multiplexing unit 103a of the vacant state. When the multiplexer +03a receives this notification, it inspects the output buffer 112a, and if an information packet exists, takes it out and transfers it to the output link 104a. On the other hand, the information packet transferred to the transfer path 110a is transferred to the distribution unit 111
Since the output link identifier in a is b, the data is accumulated in the corresponding output buffer 112a, as shown by ■ in FIG. Then, the multiplex @ 103a extracts an information packet from the output buffer 1I2a and outputs it from the output link 104a every time there is a free status notification from the address analysis unit 101a (■■ in FIG. 1). As a result, the packet is guided to the b input of node I through link L15 in FIG.

At node ■, the information packet from input link +00b is processed by address analysis unit 1, similar to the operation at node B.
01b, the routing table 120b is referred to, and the eighth
Obtain the data in the fourth row of the figure. As a result, it was determined that it was a direct line,
Transferred along direct path +02b. Multiplexer 10
3b, the human power from the direct path is output link 104
Transfer to b. In this case, non-storage transfer is performed along the main line. As a result, in FIG. 2, the packet is guided to the b input of node J through link L20.

It is assumed that the input link at node J is 1ooa. At this time, the information packet is
With reference to the routing table 120b,
As a result, it is determined that it is a transfer. Then, an output link identifier C is added to the information packet, and the information packet is transferred to the distribution unit Illb. At the same time, the multiplexing unit 103
Notify b of the free status. When the multiplexer 103b receives this notification, it inspects the output buffer 112b, and if an information packet exists, takes it out and transfers it to the output link 104b.

On the other hand, the information packet transferred to transfer path 1lOb is
In the distribution unit Illb, the data is distributed to the output buffer 112c corresponding to the output link C. The information packet is then outputted to a terminal (not shown) accommodated at the end of the link via the output link +04.

According to the above embodiment, information packets are accumulated in the output buffer only twice during a series of transfer operations, at the originating node A and at the transfer node E, and at other nodes, the information packets are stored in the output buffer only twice. Since the transfer is completed with a delay time equal to the time required to receive the data, the overall delay time is significantly reduced.

It goes without saying that the above embodiment is shown as an example of the present invention, and that the present invention is not limited thereto.

〔Effect of the invention〕

As described in detail above, according to the present invention, in a communication network in which a plurality of nodes are connected to each other by links and information is transferred in a time-sharing manner in units of information elements including routing information, an input signal is input to each node. By determining the correspondence between one of the links output from the node and one of the links output from the node, multiple paths can be defined that connect some nodes in the network. Then, I actively used this path to cause the transfer to take place.
Even when the system scale is large and the number of nodes involved is large, it has the remarkable effect of realizing the Bakke-Soto information transfer method, which makes it possible to keep the delay time of information transfer within the network to a small value. be.

[Brief explanation of drawings]

FIG. 1 is a diagram for explaining an example of the transfer operation of information packets input to a specific node in an embodiment of the present invention, and FIG. 2 is an example of the configuration of a communication network to which the packet information transfer method according to the present invention is applied. 3 is a diagram showing an example of a frame configuration on an inter-node link, FIG. 4 is a diagram showing an example of a packet configuration, FIG. 5 is a diagram showing an example of a node hardware configuration, and FIG. 6 is a diagram showing an example of a node hardware configuration. FIG. 7 is a diagram showing an example of the configuration of information packets and sotos passing on a transfer path. FIG. 8 is a diagram showing an example of a routing table when the destination node address is J. be. A to j: Node, 1 to L21: Link, 100: Input link from other node, IO4: Output link to other node, 113: Input link from terminal, 114 Output link to two terminal devices, lot ni Address analysis department, 120
IO2: direct path, 103: multiplexing section, 110: transfer path, II distribution section, 112: output buffer. To representative patent attorney Wakatoshi Isomura 1 (Figure 6 Figure 1 2)

Claims (1)

[Claims] (1) In a communication network in which a plurality of nodes are connected to each other by links and information is transferred in a time-sharing manner in units of information elements including routing information, one of the links input to each node and the node By predetermining the correspondence between one of the links output from the network, multiple paths can be defined that connect some nodes in the network, and the use of the path can be defined for each output link of each node. each node checks the routing information in the information element received from the input link, and if the routing information is of the first type, the route to which the input link belongs. If the routing information is of the second type, the information element is extracted and the corresponding time interval is made vacant, and the next step is performed for the extracted information element. It is characterized by determining an output link to be sent out, storing it in a buffer corresponding to the output link, and sending out the stored information when an idle time period occurs in the path to which the link belongs. A packet information transfer method that uses