JP3030831B2 - Interrupt reset type packet communication method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is capable of sharing a transmission line with a plurality of communication nodes and performing multimedia communication of data, voice, images, etc., and a LAN / MAN.
The present invention relates to a communication method that can be used as a WAN, and particularly to a packet communication method that can reliably transmit a priority transmission packet within an allowable delay time.

[0002]

2. Description of the Related Art Conventionally, in a communication network in which communication nodes are connected via one or a plurality of transmission lines and packet communication is performed between arbitrary terminals accommodated in these communication nodes,
The following method has been used as a method of transmitting a packet to a transmission path (access control method). That is,
(A) A communication node transmitting a packet confirms that the transmission path is unused, and then transmits the packet, and the transmitted packet collides with a packet transmitted from another communication node. In this case, the carrier detection multiple access / collision detection (CSMA /
CD: Carrier Sense Multiple Access with
(B) a packet called a token representing a transmission right is circulated on a transmission line, and a communication node receiving the token transmits the packet. When the packet transmission is completed, the token is transmitted. The next communication node
A token system for delivering the information to a computer has been known. Also, depending on the difference in the distance between the communication nodes, LAN (Local A)
rea Network is located on the premises, and MAN (Metropolitan Area)
Network (WAN) is used in the city, WAN (Wide Area Network) is used domestically (wide area). In each of these systems, (a) transmission efficiency does not increase when there is a lot of traffic. (B) It is difficult to perform high-quality voice and image communication in terms of delay time. And so on. On the other hand, in terms of transmission efficiency,
Because of the advantage that a good delay characteristic can be obtained, a slotted ring system for performing packet communication in a fixed-length slot is being studied. As an example, a cyclic monitoring type packet switching system (the 63-503
No. 10 and drawings).

[0003]

By the way, in order for a plurality of media having different allowable delay qualities to share a transmission path and perform communication, high-priority communication with strict delay is required.
A mechanism is required that temporarily suspends lower-priority communication with a slower delay and preferentially communicates with higher-priority communication. In the above-mentioned cyclic monitoring type packet switching system, a priority control mechanism, that is, (a) a mechanism for transmitting a priority transmission packet with priority over a non-priority transmission packet within one communication node, and (b) a whole communication system A mechanism for transmitting a priority transmission packet with priority over a non-priority transmission packet is provided as a protocol. However, in the cyclic monitoring type packet switching system, a communication mode T corresponding to the priority is provided, and in the communication mode T, only the information of the priority T can be transmitted. In this system, there is a problem that information of high priority that arrives in the low communication mode cannot be transmitted until the corresponding communication mode is reached. Further, when the communication mode changes, overhead occurs because the communication of all communication nodes is temporarily stopped, but when the number of transitions is large, the overhead of the communication system as a whole is increased. There was a risk of lowering the throughput.

As a system for solving this problem, an "interrupt reset type communication system using a timer" described in Japanese Patent Application No. Hei 1-208575 and the drawings has been proposed prior to the present application. In this communication method, normally, information of all priorities can be communicated, but if there is a possibility that the priority transmission packet cannot satisfy the allowable delay time,
By generating a high class interrupt reset, it is possible to shift to a priority class communication mode in which only high priority information can be transmitted. That is, as shown in FIG.
Only the high-priority information transmission mode and the information transmission modes of all priorities 1 and 2 have a higher probability of transmitting the high-priority information as compared with the traveling monitoring type packet switching system. However, in the interrupt-reset communication method using a timer, the conditions for determining that the priority transmission packet cannot satisfy the allowable delay time are also taken into consideration. Was ambiguous. That is, in the conventional method, each communication node uses a value obtained by measuring the time from the transition to the both-class communication mode by the timer, a value of the priority transmission packet number counter of the own communication node, and a value from the terminal. Upon arriving at the communication node, the number of priority transmission packets staying in the transmission queue is compared with the number of priority transmission packets to determine whether to send a high-class interrupt reset. However, since this determination criterion determines the transmission only from the transmission status of the own communication node, it is guaranteed that the priority transmission packet of the assigned priority transmission packet number limit value can always be transmitted within the allowable delay time. However, a safety factor for increasing the probability was included in the judgment formula. However, it was not clear how to determine the value of the safety factor. In addition, besides this, Japanese Patent Application No. 3-59.
No. 326, "Circuit Monitoring Packet Communication System" and Japanese Patent Application No. 3-90005, "Packet Communication Method with Limited Number of Transmission Packets", have been proposed in the respective specifications and drawings. SUMMARY OF THE INVENTION An object of the present invention is to solve the conventional problems described above, share a transmission path with a plurality of communication nodes, and perform multimedia communication such as data, voice, and image. It is an object of the present invention to provide a communication method that can be used as a MAN / WAN and can always transmit a priority transmission packet within an allowable delay time.

[0005]

To achieve the above object, an interrupt reset type packet communication method according to the present invention comprises:
The communication node counts the number of remaining packets that can be transmitted until reaching the limit values of the number of priority transmission packets and the number of non-priority transmission packets, respectively. It has two types of communication modes, a two-class communication mode in which transmission of both transmission packets is permitted, and all communication nodes in the priority class communication mode.
When the mode detects a transmission stop state in which there is no priority transmission packet to be transmitted and there is no non-priority transmission packet to transmit, the mode is switched to the both-class communication mode. When the communication node detects a transmission stop state in which there is no priority transmission packet to be transmitted and transmission of a non-priority transmission packet is prohibited, both classes of communication mode are used.
If it is determined in both communication modes that the staying time of the priority transmission packet in the communication node is long and the allowable delay time cannot be guaranteed, a high-class interrupt reset is performed. The other communication node is shifted to a high-class communication node by transmitting the packet, and the transmission of the non-priority transmission packet of the other communication node is stopped so that the priority transmission packet of the communication node can be transmitted. Therefore, in a packet communication method capable of keeping the allowable fluctuation delay time of the priority transmission packet, each communication node determines a criterion for sending a high-class interrupt reset, and all other communication nodes determine their own communication. Assuming that the same number of priority transmission packet number counters as the node remain, a high class interrupt reset is transmitted, and after transmission, all other communication nodes are reset.
Even if it is assumed that the communication node sends out all priority transmission packets of the priority transmission packet number counter value that has not been transmitted until that point and that the communication node finally transmits the priority transmission packet, the communication node It is characterized in that the residence time in the node is kept within the allowable fluctuation delay time.

[0006]

According to the present invention, based on the priority control method of the interrupt reset type communication method using a timer, a predictive control method which protects an allowable delay time of a priority transmission packet and does not include an ambiguous value such as a safety coefficient is used. Specifically, the conditions regarding the timing of sending the high class interrupt reset were set. That is, in the method of the present invention, the criterion is set on the assumption that all other communication nodes have the same number of priority transmission packet number counters as the own communication node. After transmitting the high-class interrupt reset, the same number of priority transmission packets as the value of the priority transmission packet number counter which the communication node has not been able to transmit by that time is transmitted to all other communication nodes. Are transmitted, and even if the communication node finally sends a priority transmission packet, it is determined that the staying time in the communication node can be suppressed within the allowable fluctuation delay time. The reason that another communication node determines that a priority transmission packet having the same value as that of the communication node cannot be transmitted and remains is that this algorithm is used in the same manner as all other communication nodes. In this case, the fact that no high-class interrupt reset has occurred at that time means that the number of priority transmission packets that cannot be transmitted by other communication nodes and remains is at most This is because it can be determined that the number is the same as the number of priority transmission packets remaining because a node cannot be transmitted. According to this determination method, the priority transmission packet of the priority transmission packet number limit value allocated within the allowable delay time can always be transmitted.

[0007]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of the present invention; In the communication system to which the present invention is applied, delay quality is guaranteed by generating a high-class interrupt reset. The state to be resolved by the high-class interrupt reset is as follows. That is, there is a danger that the transmission of the non-priority transmission packet of the other communication node does not keep the transmission of the priority transmission packet of the own communication node, and the priority transmission packet of the priority transmission packet limit value cannot be transmitted within the allowable delay time. There is a certain situation ”. The decision to generate a high class interrupt reset is made as follows. That is, the value of the priority transmission packet number counter with respect to the elapsed time of the reset cycle is sequentially compared using the timer that restarts when both classes are reset. As a result of the comparison, a high-class interrupt reset is generated, and after that point, the high-class communication mode is reset.
Only when the communication node transitions to the communication mode, when it is determined that the transmission of the number of remaining transmission priority packets is completed within the allowable delay time, the communication node generates a high-class interrupt reset.

[0008] Whether or not a high-class interrupt reset is generated at a certain time can be determined based on the following conditions. (Maximum time (TR) required to finish sending priority transmission packets of the value of the remaining priority transmission packet number counter of the communication node) = (Stop sending non-priority packets of all communication nodes. And the time required for all non-priority transmission packets to be removed from the ring (L)) + (required until transmission of priority transmission packets of the value of the remaining priority transmission packet number counter of all communication nodes is completed). Time (L +
(ΣWSR)) ·········· (1) The first term on the right-hand side of the above equation (1) “Stop sending non-priority packets of all communication nodes, and Time required for all non-priority transmission packets to disappear from the ring "
The worst case is the time required for the high-class interrupt reset generated at the communication node to make a circuit on the ring and to change the communication levels of all the communication nodes to the high-class communication mode. This corresponds to the ring round time (L). 12 and 13 are diagrams for explaining the state of the above equation. A to F are communication nodes, C is a non-priority packet by another communication node, High is a high-class interrupt reset packet, and a frame within a broken line indicates a packet within the transmission packet number limit value. The black rectangles inside are priority transmission packets, and the white rectangles are non-priority packets. FIG.
2 shows a state in which the transmission of the non-priority packet of the upstream communication node C prevents the transmission of the priority transmission packet of the communication node D. Here, the communication node D is
This indicates that a high class interrupt reset has been sent. In FIG. 13, one round after the communication node D sends out the high class interrupt reset, the communication node C stops sending the non-priority transmission packet.

Next, the second term on the right-hand side of the above equation (1), "the time required to finish transmitting the priority transmission packet of the value of the remaining priority transmission packet number counter of all communication nodes", is the worst. In this case, the situation is as shown in FIG. In other words, the time from when all communication nodes enter the high-class communication mode and start sending priority transmission packets until the high-class communication mode ends and transitions to the both-class communication mode is as follows. , L + ΣWSR. Where ΣWSR
Uses the value obtained by multiplying the value of the priority transmission packet number counter of the communication node that has generated the high-class interrupt reset by the number of communication nodes. This is because the communication node that first generated the high-class interrupt reset has the worst transmission condition of the priority transmission packet among all the communication nodes, and the other communication nodes have the worst condition. Is assumed, the same number of priority transmission packet number counter values as the communication node are considered to be left. FIG. 14 shows the time from the occurrence of a high-class interrupt reset until all communication nodes transmit priority transmission packets. That is, after the communication node A generates the high class interrupt reset packet (A → A), the high access communication mode is set.
(B → A) (A → B) (C → A) (A → C) (D →
A) (A → D) (E → A) (A → E) (F → A) (A →
After each communication of F), the communication node F generates both class reset packets, and A receives both class reset packets.

The conditions for sending a high-class interrupt reset at the communication node are described. The value of the timer is set to TR, and the value is reduced as time passes. The initial value of the timer sets the allowable delay time of the priority transmission packet. The value of the priority transmission packet number counter at each point in the communication node is WSR. The high-class interrupt reset is transmitted when the following condition is satisfied. TR ≦ (L) + (L + ΣWS) ≦ (L) + (L + WSR × 6) ≦ (10) + (10 + WSR × 6) (2) , The number of communication nodes is 6, and the number of packets on the ring is 10. FIG. 15 shows conditions for generating a high-class interrupt reset. In FIG.
The vertical axis shows the value of the remaining priority transmission packet counter, and the horizontal axis shows the timer value. Current,
When the value of the priority transmission packet number counter is WSR, a value equal to or less than 2L + ΣWSR is set in the timer. If the value is larger than this, the transmission of the priority transmission packet is completed, so that the high class interrupt reset is not generated. Therefore, the range on the straight line from the intersection of the remaining time 2L + ΔWSR and the counter value WSR to the point 2L on the horizontal axis is the condition for the occurrence of the high-class interrupt reset.

An example will be described below using specific numerical values with reference to FIGS. It is assumed that the number of communication nodes = 6 nodes and the number of packets on the ring = 10 packets. In addition, the priority transmission packet number limit value and the non-priority transmission packet number limit value are
The same value for all communication nodes, and the priority transmission packet number limit value = 3 and the non-priority transmission packet number limit value = 3
7 is assumed. The value of the timer is set to TR, and the value is reduced as time passes. The initial value of the timer is set to the allowable delay time of the priority transmission packet, TR = 45 in this example.
The value of the priority transmission packet number counter at each point in the communication node is WSR. All communications are broadcast communications, and all packets are addressed to the own communication node. The destination communication node is described in the control information area of the header of the packet.

FIG. 1 is a diagram showing a state of a priority transmission packet and a non-priority transmission packet staying in a transmission queue of each communication node at a certain point in time. A to F are communication nodes,
The dotted line indicates that the packet is within the transmission packet number limit value. Black rectangles indicate priority transmission packets, and white rectangles indicate non-priority transmission packets. The packet inside the dotted frame is a packet that can be transmitted in the next communication period. FIG. 2 shows a state in which the communication node A transmits both class reset packets (Both) and then transmits priority transmission packets (3) and non-priority transmission packets (5). Has returned after one lap. The fact that the packet is a reset packet is described in the control information area of the header of the packet. The reset packet is returned to a normal packet at the communication node that sent the reset. Here, the packets marked with * in FIG. 2 represent the same packets in the subsequent figures. FIG. 3 is a diagram showing a state where the communication node B has transmitted a priority transmission packet (three) and a non-priority transmission packet (seven). At this point, the communication node D has transmitted the high-class reset packet (High) because the determination formula for the generation of the high-class reset packet (formula 3 below) was satisfied. TR ≦ (10) + (10 + WSR × 6) 32 ≦ (10) + (10 + 2 × 6) (3)

FIG. 4 shows that the communication node C sends out a priority transmission packet (one packet) and a non-priority transmission packet (five packets). Therefore, the transmission of the non-priority transmission packet is stopped. After communication node D sends a high-class interrupt reset (see FIG. 3), ring 1 is used to stop sending non-priority transmission packets of communication node C, which is an upstream adjacent communication node. Time for a lap (L = 10)
Cost. FIG. 5 is a diagram showing a state where the communication nodes D and E in the high-class communication mode have transmitted priority transmission packets. A non-priority transmission packet transmitted by the communication node C on the upstream side of the communication node D that has transmitted the high-class reset packet remains on the ring. 2)
And two priority transmission packets of the communication node E are transmitted. FIG. 6 is a diagram showing a state in which the communication node F in the high-class communication mode has transmitted a priority transmission packet. That is, the priority transmission packets (two) are transmitted from the communication node F after the two priority transmission packets of the communication nodes D and E transmitted in FIG. FIG. 7 is a diagram showing a state in which the communication node F has detected that the other communication node has stopped transmitting, and has transmitted both class interrupt reset packets (Both). In this example,
When the communication node D sends the high-class interrupt reset, the communication nodes A, B, and C have no priority transmission packets to be transmitted. The transmission is completed with seven packet times left before the delay time.

As a method for detecting that all other communication nodes have stopped transmitting, a busy address packet switching system (Japanese Patent Application No. 63-50310, specification and drawings) is used. Is assumed. According to this method, the priority transmission packet counter has not reached 0, and the communication node holding the priority transmission packet to be transmitted transmits the image of the own communication node to the control information area of the header of the received packet. By continuing to overwrite the address, a packet describing the busy address of the own communication node is received. As a result, it is possible to detect the stoppage of transmission of all other communication nodes. Each communication node returns to the both-class communication mode by receiving the both-class reset packet, and transmits both the priority transmission packet and the non-priority transmission packet.

FIGS. 8 and 9 are sequence charts showing the state of packet transmission of each communication node shown in FIGS. First, after the communication node A generates the both class reset packets, three priority transmission packets and five non-priority transmission packets from A, two priority transmission packets and five non-priority transmission packets from B are respectively obtained. Has been sent. Immediately after the two non-priority transmission packets, the communication packet D generated a high-class interrupt reset packet. However, one priority transmission packet and five non-priority transmission packets from communication node C to B on the upstream side of D, and communication node A
To B and one priority transmission packet and non-priority transmission packet 2
, And five non-priority transmission packets are transmitted from A to C. D that generated a high-class interrupt reset packet
Two priority transmission packets from the downstream communication node D,
2 priority transmission packets from communication node E, communication node F
Has transmitted two priority transmission packets. Also, A
To F, two priority transmission packets are transmitted. At that time, the communication node F has generated both class reset packets.

FIG. 10 is a diagram showing the relationship between the number of priority transmission packets staying at each communication node and time. The vertical axis shows the value WSR of the priority transmission packet counter, and the horizontal axis shows the remaining time until the allowable delay time of the priority transmission packet. That is, the conditions for the occurrence of the high-class interrupt reset will be clear from now on. The black squares indicate the occurrence of a high-class interrupt reset. When the values of the priority transmission packet number counters are 3, 2, and 1 and the timer values are 38, 32, and 26, respectively,
It indicates that a high class interrupt reset can be sent. That is, from the above equation (3), when the value of the priority transmission packet number counter WSR is 2, 20 + 2 ×
Since 6 = 32, when the remaining time until the allowable delay time of the priority transmission packet is 32, D, E, and F satisfy the condition, so that a high class interrupt reset can be generated. Similarly, when WSR = 3, 20 + 3 × 6 = 38, so that it can be generated if the value of the packet number counter of B is high, and WS
When R = 1, 20 + 1 × 6 = 26, so C
Can occur if the remaining time up to the permissible delay time is small. As described above, in the present invention, after transmitting the high-class interrupt reset, all the communication nodes transmit all the priority transmission packets of the priority transmission packet number counter value which have not been transmitted until that point and remain. Finally, even when the communication node finally sends the priority transmission packet, the staying time in the communication node can be suppressed within the allowable fluctuation delay time. The fluctuation delay time is
This is a delay time that changes, such as a time for transmission access or a wait. That is, the time required to arrive at the receiving terminal via the network includes a fixed delay time, which requires a certain time elapse, such as a transfer time on a transmission path and a processing time in a communication node, and a transmission path access time. And a delay time that varies depending on the state of the communication system at that time and the congestion of the network, such as a time for waiting.
The fluctuation delay time is closely related to the window size of each communication node and the specified time of a communication section commonly set among all communication nodes. For example, as the sum of the window sizes of all the communication nodes is larger and the specified time of the communication section is shorter, an access wait occurs between the communication nodes, and the fluctuation delay time increases.

[0017]

As described above, according to the present invention,
Since the timing condition for sending the high-class interrupt reset has been clarified, it becomes possible to always transmit the priority transmission packet of the priority transmission packet number limit value allocated within the allowable delay time.

[0018]

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a first operation of interrupt reset type packet communication according to an embodiment of the present invention.

FIG. 2 is also a diagram illustrating a second operation of the interrupt reset type packet communication.

FIG. 3 is a third operation explanatory diagram of the interrupt reset type packet communication.

FIG. 4 is a fourth operation explanatory diagram of the interrupt reset type packet communication.

FIG. 5 is also a diagram explaining a fifth operation of the interrupt reset type packet communication.

FIG. 6 is a view for explaining a sixth operation of the interrupt reset type packet communication.

FIG. 7 is a diagram illustrating a seventh operation of the interrupt reset type packet communication.

FIG. 8 is a sequence chart showing a state of packet transmission of a communication node in FIGS. 1 to 7;

FIG. 9 is a sequence chart showing another part of the state of the packet transmission of the communication node.

FIG. 10 is a diagram showing conditions for occurrence of a high-class interrupt reset in the present invention.

FIG. 11 is a diagram illustrating a time division of a communication mode and a transmittable class in an interrupt reset type and a cyclic monitoring type using a timer.

FIG. 12 is a diagram illustrating a state in which transmission of a high-class interrupt reset packet is obstructed.

FIG. 13 is a diagram showing a state after a high-class interrupt reset is generated.

FIG. 14 is a time chart showing the time from when a high-class interrupt reset occurs until all communication nodes transmit a priority transmission packet.

FIG. 15 is a diagram showing conditions for occurrence of a high-class interrupt reset according to the present invention.

[Explanation of symbols]

 A to F Communication node Both class reset packet High High class interrupt reset packet WSR packet number counter value

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) H04L 12/28

Claims (1)

(57) [Claims] A plurality of communication nodes are connected by a ring-shaped or bus-shaped transmission link, and two types of a priority transmission packet and a non-priority transmission packet are provided between arbitrary communication nodes via the transmission link. A packet communication system for transmitting and receiving packets, wherein the communication node counts the number of remaining packets that can be transmitted until the respective limit values of the number of priority transmission packets and the number of non-priority transmission packets are reached; It has two types of communication modes, a priority class communication mode in which transmission is permitted only, and a two-class communication mode in which transmission of both transmission packets is permitted. There is no priority transmission packet to be transmitted by the communication node
When a transmission stop state in which there is no non-priority transmission packet to be transmitted is detected, a transition is made to both classes of communication mode, and in both classes of communication mode, priority transmission packets to be transmitted by all communication nodes are to be transmitted. When a transmission stop state is detected in which no transmission of the non-priority transmission packet is prohibited and the transmission of the non-priority transmission packet is prohibited, a transition is made to the both-class communication mode. If it is determined that the dwell time in the network becomes longer and the allowable delay time cannot be guaranteed, a high-class interrupt reset is sent out, the other communication node is transited to the high-class communication node, and the other communication node is changed. By stopping the transmission of the non-priority transmission packet of the node,
In a packet communication method in which a priority transmission packet of the communication node can be transmitted and an allowable fluctuation delay time of the priority transmission packet can be maintained, each communication node can be transmitted.
The node considers the criterion for transmitting the high-class interrupt reset assuming that all other communication nodes have the same number of priority transmission packet number counters as the own communication node.
After transmitting the high-class interrupt reset, after transmitting the high-class interrupt reset, all other communication nodes
Even if it is assumed that the communication node sends out all priority transmission packets of the priority transmission packet number counter value that has not been transmitted before that time and that the communication node finally transmits the priority transmission packet, the communication node An interrupt reset type packet communication method, characterized in that the staying time in the network is kept within an allowable fluctuation delay time.